Flask Documentation

Flask Documentation
Flask Documentation
Release 0.13.dev
May 30, 2017
User’s Guide
Foreword for Experienced Programmers
Testing Flask Applications
Application Errors
Debugging Application Errors
10 Configuration Handling
11 Signals
12 Pluggable Views
13 The Application Context
14 The Request Context
15 Modular Applications with Blueprints
16 Flask Extensions
17 Command Line Interface
18 Development Server
19 Working with the Shell
20 Patterns for Flask
21 Deployment Options
22 Becoming Big
API Reference
23 API
Additional Notes
24 Design Decisions in Flask
26 Security Considerations
27 Unicode in Flask
28 Flask Extension Development
29 Pocoo Styleguide
30 Python 3 Support
31 Upgrading to Newer Releases
32 Flask Changelog
33 License
Part I
This part of the documentation, which is mostly prose, begins with some background
information about Flask, then focuses on step-by-step instructions for web development with Flask.
Read this before you get started with Flask. This hopefully answers some questions
about the purpose and goals of the project, and when you should or should not be
using it.
What does “micro” mean?
“Micro” does not mean that your whole web application has to fit into a single Python
file (although it certainly can), nor does it mean that Flask is lacking in functionality.
The “micro” in microframework means Flask aims to keep the core simple but extensible. Flask won’t make many decisions for you, such as what database to use. Those
decisions that it does make, such as what templating engine to use, are easy to change.
Everything else is up to you, so that Flask can be everything you need and nothing
you don’t.
By default, Flask does not include a database abstraction layer, form validation or anything else where different libraries already exist that can handle that. Instead, Flask
supports extensions to add such functionality to your application as if it was implemented in Flask itself. Numerous extensions provide database integration, form validation, upload handling, various open authentication technologies, and more. Flask
may be “micro”, but it’s ready for production use on a variety of needs.
Configuration and Conventions
Flask has many configuration values, with sensible defaults, and a few conventions
when getting started. By convention, templates and static files are stored in subdi3
rectories within the application’s Python source tree, with the names templates and
static respectively. While this can be changed, you usually don’t have to, especially
when getting started.
Growing with Flask
Once you have Flask up and running, you’ll find a variety of extensions available in
the community to integrate your project for production. The Flask core team reviews
extensions and ensures approved extensions do not break with future releases.
As your codebase grows, you are free to make the design decisions appropriate for
your project. Flask will continue to provide a very simple glue layer to the best that
Python has to offer. You can implement advanced patterns in SQLAlchemy or another database tool, introduce non-relational data persistence as appropriate, and take
advantage of framework-agnostic tools built for WSGI, the Python web interface.
Flask includes many hooks to customize its behavior. Should you need more customization, the Flask class is built for subclassing. If you are interested in that, check
out the Becoming Big chapter. If you are curious about the Flask design principles, head
over to the section about Design Decisions in Flask.
Continue to Installation, the Quickstart, or the Foreword for Experienced Programmers.
Foreword for Experienced Programmers
Thread-Locals in Flask
One of the design decisions in Flask was that simple tasks should be simple; they
should not take a lot of code and yet they should not limit you. Because of that, Flask
has a few design choices that some people might find surprising or unorthodox. For
example, Flask uses thread-local objects internally so that you don’t have to pass objects around from function to function within a request in order to stay threadsafe.
This approach is convenient, but requires a valid request context for dependency injection or when attempting to reuse code which uses a value pegged to the request.
The Flask project is honest about thread-locals, does not hide them, and calls out in
the code and documentation where they are used.
Develop for the Web with Caution
Always keep security in mind when building web applications.
If you write a web application, you are probably allowing users to register and leave
their data on your server. The users are entrusting you with data. And even if you are
the only user that might leave data in your application, you still want that data to be
stored securely.
Unfortunately, there are many ways the security of a web application can be compromised. Flask protects you against one of the most common security problems of
modern web applications: cross-site scripting (XSS). Unless you deliberately mark insecure HTML as secure, Flask and the underlying Jinja2 template engine have you
covered. But there are many more ways to cause security problems.
The documentation will warn you about aspects of web development that require attention to security. Some of these security concerns are far more complex than one
might think, and we all sometimes underestimate the likelihood that a vulnerability
will be exploited - until a clever attacker figures out a way to exploit our applications.
And don’t think that your application is not important enough to attract an attacker.
Depending on the kind of attack, chances are that automated bots are probing for ways
to fill your database with spam, links to malicious software, and the like.
Flask is no different from any other framework in that you the developer must build
with caution, watching for exploits when building to your requirements.
Python 3 Support in Flask
Flask, its dependencies, and most Flask extensions all support Python 3. If you want
to use Flask with Python 3 have a look at the Python 3 Support page.
Continue to Installation or the Quickstart.
Python Version
We recommend using the latest version of Python 3. Flask supports Python 3.3 and
newer, Python 2.6 and newer, and PyPy.
These distributions will be installed automatically when installing Flask.
• Werkzeug implements WSGI, the standard Python interface between applications and servers.
• Jinja is a template language that renders the pages your application serves.
• MarkupSafe comes with Jinja. It escapes untrusted input when rendering templates to avoid injection attacks.
• ItsDangerous securely signs data to ensure its integrity. This is used to protect
Flask’s session cookie.
• Click is a framework for writing command line applications. It provides the
flask command and allows adding custom management commands.
Optional dependencies
These distributions will not be installed automatically. Flask will detect and use them
if you install them.
• Blinker provides support for Signals.
• SimpleJSON is a fast JSON implementation that is compatible with Python’s json
module. It is preferred for JSON operations if it is installed.
Virtual environments
Use a virtual environment to manage the dependencies for your project, both in development and in production.
What problem does a virtual environment solve? The more Python projects you have,
the more likely it is that you need to work with different versions of Python libraries, or
even Python itself. Newer versions of libraries for one project can break compatibility
in another project.
Virtual environments are independent groups of Python libraries, one for each project.
Packages installed for one project will not affect other projects or the operating system’s packages.
Python 3 comes bundled with the venv module to create virtual environments. If
you’re using a modern version of Python, you can continue on to the next section.
If you’re using Python 2, see Install virtualenv first.
Create an environment
Create a project folder and a venv folder within:
mkdir myproject
cd myproject
python3 -m venv venv
On Windows:
py -3 -m venv venv
If you needed to install virtualenv because you are on an older version of Python, use
the following command instead:
virtualenv venv
On Windows:
\Python27\Scripts\virtualenv.exe venv
Activate the environment
Before you work on your project, activate the corresponding environment:
. venv/bin/activate
On Windows:
Your shell prompt will change to show the name of the activated environment.
Install Flask
Within the activated environment, use the following command to install Flask:
pip install Flask
Living on the edge
If you want to work with the latest Flask code before it’s released, install or update the
code from the master branch:
pip install -U https://github.com/pallets/flask/archive/master.tar.gz
Install virtualenv
If you are using Python 2, the venv module is not available. Instead, install virtualenv.
On Linux, virtualenv is provided by your package manager:
# Debian, Ubuntu
sudo apt-get install python-virtualenv
# CentOS, Fedora
sudo yum install python-virtualenv
# Arch
sudo pacman -S python-virtualenv
If you are on Mac OS X or Windows, download get-pip.py, then:
sudo python2 Downloads/get-pip.py
sudo python2 -m pip install virtualenv
On Windows, as an administrator:
\Python27\python.exe Downloads\get-pip.py
\Python27\python.exe -m pip install virtualenv
Now you can continue to Create an environment.
Eager to get started? This page gives a good introduction to Flask. It assumes you
already have Flask installed. If you do not, head over to the Installation section.
A Minimal Application
A minimal Flask application looks something like this:
from flask import Flask
app = Flask(__name__)
def hello_world():
return 'Hello, World!'
So what did that code do?
1. First we imported the Flask class. An instance of this class will be our WSGI
2. Next we create an instance of this class. The first argument is the name of the
application’s module or package. If you are using a single module (as in this
example), you should use __name__ because depending on if it’s started as application or imported as module the name will be different ('__main__' versus
the actual import name). This is needed so that Flask knows where to look for
templates, static files, and so on. For more information have a look at the Flask
3. We then use the route() decorator to tell Flask what URL should trigger our
4. The function is given a name which is also used to generate URLs for that particular function, and returns the message we want to display in the user’s browser.
Just save it as hello.py or something similar. Make sure to not call your application
flask.py because this would conflict with Flask itself.
To run the application you can either use the flask command or python’s -m switch
with Flask. Before you can do that you need to tell your terminal the application to
work with by exporting the FLASK_APP environment variable:
$ export FLASK_APP=hello.py
$ flask run
* Running on
If you are on Windows, the environment variable syntax depends on command line
interpreter. On Command Prompt:
C:\path\to\app>set FLASK_APP=hello.py
And on PowerShell:
PS C:\path\to\app> $env:FLASK_APP = "hello.py"
Alternatively you can use python -m flask:
$ export FLASK_APP=hello.py
$ python -m flask run
* Running on
This launches a very simple builtin server, which is good enough for testing but probably not what you want to use in production. For deployment options see Deployment
Now head over to, and you should see your hello world greeting.
Externally Visible Server
If you run the server you will notice that the server is only accessible from your own
computer, not from any other in the network. This is the default because in debugging
mode a user of the application can execute arbitrary Python code on your computer.
If you have the debugger disabled or trust the users on your network, you can make
the server publicly available simply by adding --host= to the command line:
flask run --host=
This tells your operating system to listen on all public IPs.
What to do if the Server does not Start
In case the python -m flask fails or flask does not exist, there are multiple reasons
this might be the case. First of all you need to look at the error message.
Old Version of Flask
Versions of Flask older than 0.11 use to have different ways to start the application. In
short, the flask command did not exist, and neither did python -m flask. In that case
you have two options: either upgrade to newer Flask versions or have a look at the
Development Server docs to see the alternative method for running a server.
Invalid Import Name
The FLASK_APP environment variable is the name of the module to import at flask run.
In case that module is incorrectly named you will get an import error upon start (or if
debug is enabled when you navigate to the application). It will tell you what it tried
to import and why it failed.
The most common reason is a typo or because you did not actually create an app object.
Debug Mode
(Want to just log errors and stack traces? See Application Errors)
The flask script is nice to start a local development server, but you would have to
restart it manually after each change to your code. That is not very nice and Flask can
do better. If you enable debug support the server will reload itself on code changes,
and it will also provide you with a helpful debugger if things go wrong.
To enable debug mode you can export the FLASK_DEBUG environment variable before
running the server:
$ export FLASK_DEBUG=1
$ flask run
(On Windows you need to use set instead of export).
This does the following things:
1. it activates the debugger
2. it activates the automatic reloader
3. it enables the debug mode on the Flask application.
There are more parameters that are explained in the Development Server docs.
Even though the interactive debugger does not work in forking environments (which
makes it nearly impossible to use on production servers), it still allows the execution
of arbitrary code. This makes it a major security risk and therefore it must never be
used on production machines.
Screenshot of the debugger in action:
More information on using the debugger can be found in the Werkzeug documentation.
Have another debugger in mind? See Working with Debuggers.
Modern web applications use meaningful URLs to help users. Users are more likely
to like a page and come back if the page uses a meaningful URL they can remember
and use to directly visit a page.
Use the route() decorator to bind a function to a URL.
def index():
return 'Index Page'
def hello():
return 'Hello, World'
You can do more! You can make parts of the URL dynamic and attach multiple rules
to a function.
Variable Rules
You can add variable sections to a URL by marking sections with <variable_name>.
Your function then receives the <variable_name> as a keyword argument. Optionally, you can use a converter to specify the type of the argument like
def show_user_profile(username):
# show the user profile for that user
return 'User %s' % username
def show_post(post_id):
# show the post with the given id, the id is an integer
return 'Post %d' % post_id
def show_subpath(subpath):
# show the subpath after /path/
return 'Subpath %s' % subpath
Converter types:
(default) accepts any text without a slash
accepts positive integers
accepts positive floating point values
like string but also accepts slashes
accepts UUID strings
Unique URLs / Redirection Behavior
Take these two rules:
def projects():
return 'The project page'
def about():
return 'The about page'
Though they look similar, they differ in their use of the trailing slash in the URL. In
the first case, the canonical URL for the projects endpoint uses a trailing slash. It’s
similar to a folder in a file system; if you access the URL without a trailing slash, Flask
redirects you to the canonical URL with the trailing slash.
In the second case, however, the URL definition lacks a trailing slash, like the pathname of a file on UNIX-like systems. Accessing the URL with a trailing slash produces
a 404 “Not Found” error.
This behavior allows relative URLs to continue working even if the trailing slash is
omitted, consistent with how Apache and other servers work. Also, the URLs will
stay unique, which helps search engines avoid indexing the same page twice.
URL Building
To build a URL to a specific function, use the url_for() function. It accepts the name
of the function as its first argument and any number of keyword arguments, each corresponding to a variable part of the URL rule. Unknown variable parts are appended
to the URL as query parameters.
Why would you want to build URLs using the URL reversing function url_for() instead of hard-coding them into your templates?
1. Reversing is often more descriptive than hard-coding the URLs.
2. You can change your URLs in one go instead of needing to remember to
manually change hard-coded URLs.
3. URL building handles escaping of special characters and Unicode data
4. If your application is placed outside the URL root, for example, in /
myapplication instead of /, url_for() properly handles that for you.
For example, here we use the test_request_context() method to try out url_for().
test_request_context() tells Flask to behave as though it’s handling a request even
while we use a Python shell. See Context Locals.
from flask import Flask, url_for
app = Flask(__name__)
def index():
return 'index'
def login():
return 'login'
def profile(username):
return '{}'s profile'.format(username)
with app.test_request_context():
print(url_for('login', next='/'))
print(url_for('profile', username='John Doe'))
HTTP Methods
Web applications use different HTTP methods when accessing URLs. You should familiarize yourself with the HTTP methods as you work with Flask. By default, a route
only answers to GET requests. You can use the methods argument of the route() decorator to handle different HTTP methods.
@app.route('/login', methods=['GET', 'POST'])
def login():
if request.method == 'POST':
If GET is present, Flask automatically adds support for the HEAD method and handles
HEAD requests according to the the HTTP RFC. Likewise, OPTIONS is automatically implemented for you.
Static Files
Dynamic web applications also need static files. That’s usually where the CSS and
JavaScript files are coming from. Ideally your web server is configured to serve them
for you, but during development Flask can do that as well. Just create a folder called
static in your package or next to your module and it will be available at /static on
the application.
To generate URLs for static files, use the special 'static' endpoint name:
url_for('static', filename='style.css')
The file has to be stored on the filesystem as static/style.css.
Rendering Templates
Generating HTML from within Python is not fun, and actually pretty cumbersome because you have to do the HTML escaping on your own to keep the application secure.
Because of that Flask configures the Jinja2 template engine for you automatically.
To render a template you can use the render_template() method. All you have to do
is provide the name of the template and the variables you want to pass to the template
engine as keyword arguments. Here’s a simple example of how to render a template:
from flask import render_template
def hello(name=None):
return render_template('hello.html', name=name)
Flask will look for templates in the templates folder. So if your application is a module,
this folder is next to that module, if it’s a package it’s actually inside your package:
Case 1: a module:
Case 2: a package:
For templates you can use the full power of Jinja2 templates. Head over to the official
Jinja2 Template Documentation for more information.
Here is an example template:
<!doctype html>
<title>Hello from Flask</title>
{% if name %}
<h1>Hello {{ name }}!</h1>
{% else %}
<h1>Hello, World!</h1>
{% endif %}
Inside templates you also have access to the request, session and g1 objects as well as
the get_flashed_messages() function.
Templates are especially useful if inheritance is used. If you want to know how that
works, head over to the Template Inheritance pattern documentation. Basically template inheritance makes it possible to keep certain elements on each page (like header,
navigation and footer).
Automatic escaping is enabled, so if name contains HTML it will be escaped automatically. If you can trust a variable and you know that it will be safe HTML (for example
because it came from a module that converts wiki markup to HTML) you can mark
it as safe by using the Markup class or by using the |safe filter in the template. Head
over to the Jinja 2 documentation for more examples.
Here is a basic introduction to how the Markup class works:
>>> from flask import Markup
>>> Markup('<strong>Hello %s!</strong>') % '<blink>hacker</blink>'
Markup(u'<strong>Hello &lt;blink&gt;hacker&lt;/blink&gt;!</strong>')
>>> Markup.escape('<blink>hacker</blink>')
>>> Markup('<em>Marked up</em> &raquo; HTML').striptags()
u'Marked up \xbb HTML'
Changed in version 0.5: Autoescaping is no longer enabled for all templates. The
following extensions for templates trigger autoescaping: .html, .htm, .xml, .xhtml.
Templates loaded from a string will have autoescaping disabled.
Accessing Request Data
For web applications it’s crucial to react to the data a client sends to the server. In
Flask this information is provided by the global request object. If you have some
experience with Python you might be wondering how that object can be global and
how Flask manages to still be threadsafe. The answer is context locals:
Context Locals
Insider Information
Unsure what that g object is? It’s something in which you can store information for your own
needs, check the documentation of that object (g) and the Using SQLite 3 with Flask for more information.
If you want to understand how that works and how you can implement tests with
context locals, read this section, otherwise just skip it.
Certain objects in Flask are global objects, but not of the usual kind. These objects are
actually proxies to objects that are local to a specific context. What a mouthful. But
that is actually quite easy to understand.
Imagine the context being the handling thread. A request comes in and the web server
decides to spawn a new thread (or something else, the underlying object is capable
of dealing with concurrency systems other than threads). When Flask starts its internal request handling it figures out that the current thread is the active context and
binds the current application and the WSGI environments to that context (thread). It
does that in an intelligent way so that one application can invoke another application
without breaking.
So what does this mean to you? Basically you can completely ignore that this is the
case unless you are doing something like unit testing. You will notice that code which
depends on a request object will suddenly break because there is no request object. The
solution is creating a request object yourself and binding it to the context. The easiest
solution for unit testing is to use the test_request_context() context manager. In
combination with the with statement it will bind a test request so that you can interact
with it. Here is an example:
from flask import request
with app.test_request_context('/hello', method='POST'):
# now you can do something with the request until the
# end of the with block, such as basic assertions:
assert request.path == '/hello'
assert request.method == 'POST'
The other possibility is passing a whole WSGI environment to the request_context()
from flask import request
with app.request_context(environ):
assert request.method == 'POST'
The Request Object
The request object is documented in the API section and we will not cover it here in
detail (see Request). Here is a broad overview of some of the most common operations.
First of all you have to import it from the flask module:
from flask import request
The current request method is available by using the method attribute. To access form
data (data transmitted in a POST or PUT request) you can use the form attribute. Here is
a full example of the two attributes mentioned above:
@app.route('/login', methods=['POST', 'GET'])
def login():
error = None
if request.method == 'POST':
if valid_login(request.form['username'],
return log_the_user_in(request.form['username'])
error = 'Invalid username/password'
# the code below is executed if the request method
# was GET or the credentials were invalid
return render_template('login.html', error=error)
What happens if the key does not exist in the form attribute? In that case a special
KeyError is raised. You can catch it like a standard KeyError but if you don’t do that, a
HTTP 400 Bad Request error page is shown instead. So for many situations you don’t
have to deal with that problem.
To access parameters submitted in the URL (?key=value) you can use the args attribute:
searchword = request.args.get('key', '')
We recommend accessing URL parameters with get or by catching the KeyError because users might change the URL and presenting them a 400 bad request page in that
case is not user friendly.
For a full list of methods and attributes of the request object, head over to the Request
File Uploads
You can handle uploaded files with Flask easily. Just make sure not to forget to
set the enctype="multipart/form-data" attribute on your HTML form, otherwise the
browser will not transmit your files at all.
Uploaded files are stored in memory or at a temporary location on the filesystem. You
can access those files by looking at the files attribute on the request object. Each
uploaded file is stored in that dictionary. It behaves just like a standard Python file
object, but it also has a save() method that allows you to store that file on the filesystem of the server. Here is a simple example showing how that works:
from flask import request
@app.route('/upload', methods=['GET', 'POST'])
def upload_file():
if request.method == 'POST':
f = request.files['the_file']
If you want to know how the file was named on the client before it was uploaded to
your application, you can access the filename attribute. However please keep in mind
that this value can be forged so never ever trust that value. If you want to use the filename of the client to store the file on the server, pass it through the secure_filename()
function that Werkzeug provides for you:
from flask import request
from werkzeug.utils import secure_filename
@app.route('/upload', methods=['GET', 'POST'])
def upload_file():
if request.method == 'POST':
f = request.files['the_file']
f.save('/var/www/uploads/' + secure_filename(f.filename))
For some better examples, checkout the Uploading Files pattern.
To access cookies you can use the cookies attribute. To set cookies you can use the
set_cookie method of response objects. The cookies attribute of request objects is a
dictionary with all the cookies the client transmits. If you want to use sessions, do not
use the cookies directly but instead use the Sessions in Flask that add some security on
top of cookies for you.
Reading cookies:
from flask import request
def index():
username = request.cookies.get('username')
# use cookies.get(key) instead of cookies[key] to not get a
# KeyError if the cookie is missing.
Storing cookies:
from flask import make_response
def index():
resp = make_response(render_template(...))
resp.set_cookie('username', 'the username')
return resp
Note that cookies are set on response objects. Since you normally just return strings
from the view functions Flask will convert them into response objects for you. If you
explicitly want to do that you can use the make_response() function and then modify
Sometimes you might want to set a cookie at a point where the response object does
not exist yet. This is possible by utilizing the Deferred Request Callbacks pattern.
For this also see About Responses.
Redirects and Errors
To redirect a user to another endpoint, use the redirect() function; to abort a request
early with an error code, use the abort() function:
from flask import abort, redirect, url_for
def index():
return redirect(url_for('login'))
def login():
This is a rather pointless example because a user will be redirected from the index to
a page they cannot access (401 means access denied) but it shows how that works.
By default a black and white error page is shown for each error code. If you want to
customize the error page, you can use the errorhandler() decorator:
from flask import render_template
def page_not_found(error):
return render_template('page_not_found.html'), 404
Note the 404 after the render_template() call. This tells Flask that the status code of
that page should be 404 which means not found. By default 200 is assumed which
translates to: all went well.
See Error handlers for more details.
About Responses
The return value from a view function is automatically converted into a response object for you. If the return value is a string it’s converted into a response object with the
string as response body, a 200 OK status code and a text/html mimetype. The logic
that Flask applies to converting return values into response objects is as follows:
1. If a response object of the correct type is returned it’s directly returned from the
2. If it’s a string, a response object is created with that data and the default parameters.
3. If a tuple is returned the items in the tuple can provide extra information. Such
tuples have to be in the form (response, status, headers) or (response,
headers) where at least one item has to be in the tuple. The status value will
override the status code and headers can be a list or dictionary of additional
header values.
4. If none of that works, Flask will assume the return value is a valid WSGI application and convert that into a response object.
If you want to get hold of the resulting response object inside the view you can use the
make_response() function.
Imagine you have a view like this:
def not_found(error):
return render_template('error.html'), 404
You just need to wrap the return expression with make_response() and get the response object to modify it, then return it:
def not_found(error):
resp = make_response(render_template('error.html'), 404)
resp.headers['X-Something'] = 'A value'
return resp
In addition to the request object there is also a second object called session which
allows you to store information specific to a user from one request to the next. This is
implemented on top of cookies for you and signs the cookies cryptographically. What
this means is that the user could look at the contents of your cookie but not modify it,
unless they know the secret key used for signing.
In order to use sessions you have to set a secret key. Here is how sessions work:
from flask import Flask, session, redirect, url_for, escape, request
app = Flask(__name__)
def index():
if 'username' in session:
return 'Logged in as %s' % escape(session['username'])
return 'You are not logged in'
@app.route('/login', methods=['GET', 'POST'])
def login():
if request.method == 'POST':
session['username'] = request.form['username']
return redirect(url_for('index'))
return '''
<form method="post">
<p><input type=text name=username>
<p><input type=submit value=Login>
def logout():
# remove the username from the session if it's there
session.pop('username', None)
return redirect(url_for('index'))
# set the secret key. keep this really secret:
app.secret_key = 'A0Zr98j/3yX R~XHH!jmN]LWX/,?RT'
The escape() mentioned here does escaping for you if you are not using the template
engine (as in this example).
How to generate good secret keys
The problem with random is that it’s hard to judge what is truly random. And a secret
key should be as random as possible. Your operating system has ways to generate
pretty random stuff based on a cryptographic random generator which can be used to
get such a key:
>>> import os
>>> os.urandom(24)
Just take that thing and copy/paste it into your code and you're done.
A note on cookie-based sessions: Flask will take the values you put into the session
object and serialize them into a cookie. If you are finding some values do not persist across requests, cookies are indeed enabled, and you are not getting a clear error
message, check the size of the cookie in your page responses compared to the size
supported by web browsers.
Besides the default client-side based sessions, if you want to handle sessions on the
server-side instead, there are several Flask extensions that support this.
Message Flashing
Good applications and user interfaces are all about feedback. If the user does not get
enough feedback they will probably end up hating the application. Flask provides a
really simple way to give feedback to a user with the flashing system. The flashing
system basically makes it possible to record a message at the end of a request and
access it on the next (and only the next) request. This is usually combined with a
layout template to expose the message.
To flash a message use the flash() method, to get hold of the messages you can use
get_flashed_messages() which is also available in the templates. Check out the Message Flashing for a full example.
New in version 0.3.
Sometimes you might be in a situation where you deal with data that should be correct,
but actually is not. For example you may have some client-side code that sends an
HTTP request to the server but it’s obviously malformed. This might be caused by a
user tampering with the data, or the client code failing. Most of the time it’s okay to
reply with 400 Bad Request in that situation, but sometimes that won’t do and the
code has to continue working.
You may still want to log that something fishy happened. This is where loggers come
in handy. As of Flask 0.3 a logger is preconfigured for you to use.
Here are some example log calls:
app.logger.debug('A value for debugging')
app.logger.warning('A warning occurred (%d apples)', 42)
app.logger.error('An error occurred')
The attached logger is a standard logging Logger, so head over to the official logging
documentation for more information.
Read more on Application Errors.
Hooking in WSGI Middlewares
If you want to add a WSGI middleware to your application you can wrap the internal
WSGI application. For example if you want to use one of the middlewares from the
Werkzeug package to work around bugs in lighttpd, you can do it like this:
from werkzeug.contrib.fixers import LighttpdCGIRootFix
app.wsgi_app = LighttpdCGIRootFix(app.wsgi_app)
Using Flask Extensions
Extensions are packages that help you accomplish common tasks. For example, FlaskSQLAlchemy provides SQLAlchemy support that makes it simple and easy to use
with Flask.
For more on Flask extensions, have a look at Flask Extensions.
Deploying to a Web Server
Ready to deploy your new Flask app? Go to Deployment Options.
Learn by example to develop an application with Python and Flask.
In this tutorial, we will create a simple blogging application. It only supports one user,
only allows text entries, and has no feeds or comments.
While very simple, this example still features everything you need to get started. In
addition to Flask, we will use SQLite for the database, which is built-in to Python, so
there is nothing else you need.
If you want the full source code in advance or for comparison, check out the example
Introducing Flaskr
This tutorial will demonstrate a blogging application named Flaskr, but feel free to
choose your own less Web-2.0-ish name ;) Essentially, it will do the following things:
1. Let the user sign in and out with credentials specified in the configuration. Only
one user is supported.
2. When the user is logged in, they can add new entries to the page consisting of a
text-only title and some HTML for the text. This HTML is not sanitized because
we trust the user here.
3. The index page shows all entries so far in reverse chronological order (newest on
top) and the user can add new ones from there if logged in.
SQLite3 will be used directly for this application because it’s good enough for an application of this size. For larger applications, however, it makes a lot of sense to use
SQLAlchemy, as it handles database connections in a more intelligent way, allowing
you to target different relational databases at once and more. You might also want to
consider one of the popular NoSQL databases if your data is more suited for those.
Here a screenshot of the final application:
Continue with Step 0: Creating The Folders.
Step 0: Creating The Folders
It is recommended to install your Flask application within a virtualenv. Please read
the Installation section to set up your environment.
Now that you have installed Flask, you will need to create the folders required for this
tutorial. Your directory structure will look like this:
The application will be installed and run as Python package. This is the recommended
way to install and run Flask applications. You will see exactly how to run flaskr later
on in this tutorial.
For now go ahead and create the applications directory structure. In the next few steps
you will be creating the database schema as well as the main module.
As a quick side note, the files inside of the static folder are available to users of the
application via HTTP. This is the place where CSS and JavaScript files go. Inside the
templates folder, Flask will look for Jinja2 templates. You will see examples of this
later on.
For now you should continue with Step 1: Database Schema.
Step 1: Database Schema
In this step, you will create the database schema. Only a single table is needed for this
application and it will only support SQLite. All you need to do is put the following
contents into a file named schema.sql in the flaskr/flaskr folder:
drop table if exists entries;
create table entries (
id integer primary key autoincrement,
title text not null,
'text' text not null
This schema consists of a single table called entries. Each row in this table has an id,
a title, and a text. The id is an automatically incrementing integer and a primary
key, the other two are strings that must not be null.
Continue with Step 2: Application Setup Code.
Step 2: Application Setup Code
Next, we will create the application module, flaskr.py. Just like the schema.sql file
you created in the previous step, this file should be placed inside of the flaskr/flaskr
For this tutorial, all the Python code we use will be put into this file (except for one
line in __init__.py, and any testing or optional files you decide to create).
The first several lines of code in the application module are the needed import statements. After that there will be a few lines of configuration code.
For small applications like flaskr, it is possible to drop the configuration directly into
the module. However, a cleaner solution is to create a separate .py file, load that, and
import the values from there.
Here are the import statements (in flaskr.py):
import os
import sqlite3
from flask import (Flask, request, session, g, redirect, url_for, abort,
render_template, flash)
The next couple lines will create the actual application instance and initialize it with
the config from the same file in flaskr.py:
app = Flask(__name__) # create the application instance :)
app.config.from_object(__name__) # load config from this file , flaskr.py
# Load default config and override config from an environment variable
DATABASE=os.path.join(app.root_path, 'flaskr.db'),
SECRET_KEY='development key',
app.config.from_envvar('FLASKR_SETTINGS', silent=True)
In the above code, the Config object works similarly to a dictionary, so it can be updated with new values.
Database Path
Operating systems know the concept of a current working directory for each process.
Unfortunately, you cannot depend on this in web applications because you might have
more than one application in the same process.
For this reason the app.root_path attribute can be used to get the path to the application. Together with the os.path module, files can then easily be found. In this example,
we place the database right next to it.
For a real-world application, it’s recommended to use Instance Folders instead.
Usually, it is a good idea to load a separate, environment-specific configuration file.
Flask allows you to import multiple configurations and it will use the setting defined
in the last import. This enables robust configuration setups. from_envvar() can help
achieve this.
app.config.from_envvar('FLASKR_SETTINGS', silent=True)
If you want to do this (not required for this tutorial) simply define the environment
variable FLASKR_SETTINGS that points to a config file to be loaded. The silent switch
just tells Flask to not complain if no such environment key is set.
In addition to that, you can use the from_object() method on the config object and
provide it with an import name of a module. Flask will then initialize the variable
from that module. Note that in all cases, only variable names that are uppercase are
The SECRET_KEY is needed to keep the client-side sessions secure. Choose that key
wisely and as hard to guess and complex as possible.
Lastly, add a method that allows for easy connections to the specified database.
def connect_db():
"""Connects to the specific database."""
rv = sqlite3.connect(app.config['DATABASE'])
rv.row_factory = sqlite3.Row
return rv
This can be used to open a connection on request and also from the interactive Python
shell or a script. This will come in handy later. You can create a simple database
connection through SQLite and then tell it to use the sqlite3.Row object to represent
rows. This allows the rows to be treated as if they were dictionaries instead of tuples.
In the next section you will see how to run the application.
Continue with Step 3: Installing flaskr as a Package.
Step 3: Installing flaskr as a Package
Flask is now shipped with built-in support for Click. Click provides Flask with enhanced and extensible command line utilities. Later in this tutorial you will see exactly
how to extend the flask command line interface (CLI).
A useful pattern to manage a Flask application is to install your app following the
Python Packaging Guide. Presently this involves creating two new files; setup.py and
MANIFEST.in in the projects root directory. You also need to add an __init__.py file to
make the flaskr/flaskr directory a package. After these changes, your code structure
should be:
Create the setup.py file for flaskr with the following content:
from setuptools import setup
When using setuptools, it is also necessary to specify any special files that should be
included in your package (in the MANIFEST.in). In this case, the static and templates
directories need to be included, as well as the schema.
Create the MANIFEST.in and add the following lines:
graft flaskr/templates
graft flaskr/static
include flaskr/schema.sql
Next, to simplify locating the application, create the file, flaskr/__init__.py containing only the following import statement:
from .flaskr import app
This import statement brings the application instance into the top-level of the application package. When it is time to run the application, the Flask development server
needs the location of the app instance. This import statement simplifies the location
process. Without the above import statement, the export statement a few steps below
would need to be export FLASK_APP=flaskr.flaskr.
At this point you should be able to install the application. As usual, it is recommended
to install your Flask application within a virtualenv. With that said, from the flaskr/
directory, go ahead and install the application with:
pip install --editable .
The above installation command assumes that it is run within the projects root directory, flaskr/. The editable flag allows editing source code without having to reinstall
the Flask app each time you make changes. The flaskr app is now installed in your virtualenv (see output of pip freeze).
With that out of the way, you should be able to start up the application. Do this on
Mac or Linux with the following commands in flaskr/:
export FLASK_APP=flaskr
export FLASK_DEBUG=true
flask run
(In case you are on Windows you need to use set instead of export). The FLASK_DEBUG
flag enables or disables the interactive debugger. Never leave debug mode activated in a
production system, because it will allow users to execute code on the server!
You will see a message telling you that server has started along with the address at
which you can access it in a browser.
When you head over to the server in your browser, you will get a 404 error because
we don’t have any views yet. That will be addressed a little later, but first, you should
get the database working.
Externally Visible Server
Want your server to be publicly available? Check out the externally visible server section
for more information.
Continue with Step 4: Database Connections.
Step 4: Database Connections
Let’s continue building our code in the flaskr.py file. (Scroll to the end of the page
for more about project layout.)
You currently have a function for establishing a database connection with connect_db,
but by itself, it is not particularly useful. Creating and closing database connections
all the time is very inefficient, so you will need to keep it around for longer. Because
database connections encapsulate a transaction, you will need to make sure that only
one request at a time uses the connection. An elegant way to do this is by utilizing the
application context.
Flask provides two contexts: the application context and the request context. For the time
being, all you have to know is that there are special variables that use these. For instance, the request variable is the request object associated with the current request,
whereas g is a general purpose variable associated with the current application context. The tutorial will cover some more details of this later on.
For the time being, all you have to know is that you can store information safely on
the g object.
So when do you put it on there? To do that you can make a helper function. The first
time the function is called, it will create a database connection for the current context,
and successive calls will return the already established connection:
def get_db():
"""Opens a new database connection if there is none yet for the
current application context.
if not hasattr(g, 'sqlite_db'):
g.sqlite_db = connect_db()
return g.sqlite_db
Now you know how to connect, but how can you properly disconnect? For that, Flask
provides us with the teardown_appcontext() decorator. It’s executed every time the
application context tears down:
def close_db(error):
"""Closes the database again at the end of the request."""
if hasattr(g, 'sqlite_db'):
Functions marked with teardown_appcontext() are called every time the app context
tears down. What does this mean? Essentially, the app context is created before the
request comes in and is destroyed (torn down) whenever the request finishes. A teardown can happen because of two reasons: either everything went well (the error parameter will be None) or an exception happened, in which case the error is passed to
the teardown function.
Curious about what these contexts mean? Have a look at the The Application Context
documentation to learn more.
Continue to Step 5: Creating The Database.
Hint: Where do I put this code?
If you’ve been following along in this tutorial, you might be wondering where to put
the code from this step and the next. A logical place is to group these module-level
functions together, and put your new get_db and close_db functions below your existing connect_db function (following the tutorial line-by-line).
If you need a moment to find your bearings, take a look at how the example source
is organized. In Flask, you can put all of your application code into a single Python
module. You don’t have to, and if your app grows larger, it’s a good idea not to.
Step 5: Creating The Database
As outlined earlier, Flaskr is a database powered application, and more precisely, it is
an application powered by a relational database system. Such systems need a schema
that tells them how to store that information. Before starting the server for the first
time, it’s important to create that schema.
Such a schema could be created by piping the schema.sql file into the sqlite3 command as follows:
sqlite3 /tmp/flaskr.db < schema.sql
However, the downside of this is that it requires the sqlite3 command to be installed,
which is not necessarily the case on every system. This also requires that you provide
the path to the database, which can introduce errors.
Instead of the sqlite3 command above, it’s a good idea to add a function to our application that initializes the database for you. To do this, you can create a function and
hook it into a flask command that initializes the database.
Take a look at the code segment below. A good place to add this function, and command, is just below the connect_db function in flaskr.py:
def init_db():
db = get_db()
with app.open_resource('schema.sql', mode='r') as f:
def initdb_command():
"""Initializes the database."""
print('Initialized the database.')
The app.cli.command() decorator registers a new command with the flask script.
When the command executes, Flask will automatically create an application context
which is bound to the right application. Within the function, you can then access
flask.g and other things as you might expect. When the script ends, the application
context tears down and the database connection is released.
You will want to keep an actual function around that initializes the database, though,
so that we can easily create databases in unit tests later on. (For more information see
Testing Flask Applications.)
The open_resource() method of the application object is a convenient helper function
that will open a resource that the application provides. This function opens a file from
the resource location (the flaskr/flaskr folder) and allows you to read from it. It is
used in this example to execute a script on the database connection.
The connection object provided by SQLite can give you a cursor object. On that cursor, there is a method to execute a complete script. Finally, you only have to commit
the changes. SQLite3 and other transactional databases will not commit unless you
explicitly tell it to.
Now, in a terminal, from the application root directory flaskr/ it is possible to create
a database with the flask script:
flask initdb
Initialized the database.
If you get an exception later on stating that a table cannot be found, check that you
did execute the initdb command and that your table names are correct (singular vs.
plural, for example).
Continue with Step 6: The View Functions
Step 6: The View Functions
Now that the database connections are working, you can start writing the view functions. You will need four of them; Show Entries, Add New Entry, Login and Logout.
Add the following code snipets to flaskr.py.
Show Entries
This view shows all the entries stored in the database. It listens on the root of the
application and will select title and text from the database. The one with the highest
id (the newest entry) will be on top. The rows returned from the cursor look a bit like
dictionaries because we are using the sqlite3.Row row factory.
The view function will pass the entries to the show_entries.html template and return
the rendered one:
def show_entries():
db = get_db()
cur = db.execute('select title, text from entries order by id desc')
entries = cur.fetchall()
return render_template('show_entries.html', entries=entries)
Add New Entry
This view lets the user add new entries if they are logged in. This only responds to
POST requests; the actual form is shown on the show_entries page. If everything worked
out well, it will flash() an information message to the next request and redirect back
to the show_entries page:
@app.route('/add', methods=['POST'])
def add_entry():
if not session.get('logged_in'):
db = get_db()
db.execute('insert into entries (title, text) values (?, ?)',
[request.form['title'], request.form['text']])
flash('New entry was successfully posted')
return redirect(url_for('show_entries'))
Note that this view checks that the user is logged in (that is, if the logged_in key is
present in the session and True).
Security Note
Be sure to use question marks when building SQL statements, as done in the example
above. Otherwise, your app will be vulnerable to SQL injection when you use string
formatting to build SQL statements. See Using SQLite 3 with Flask for more.
Login and Logout
These functions are used to sign the user in and out. Login checks the username and
password against the ones from the configuration and sets the logged_in key for the
session. If the user logged in successfully, that key is set to True, and the user is redirected back to the show_entries page. In addition, a message is flashed that informs the
user that he or she was logged in successfully. If an error occurred, the template is
notified about that, and the user is asked again:
@app.route('/login', methods=['GET', 'POST'])
def login():
error = None
if request.method == 'POST':
if request.form['username'] != app.config['USERNAME']:
error = 'Invalid username'
elif request.form['password'] != app.config['PASSWORD']:
error = 'Invalid password'
session['logged_in'] = True
flash('You were logged in')
return redirect(url_for('show_entries'))
return render_template('login.html', error=error)
The logout function, on the other hand, removes that key from the session again. There
is a neat trick here: if you use the pop() method of the dict and pass a second parameter
to it (the default), the method will delete the key from the dictionary if present or do
nothing when that key is not in there. This is helpful because now it is not necessary
to check if the user was logged in.
def logout():
session.pop('logged_in', None)
flash('You were logged out')
return redirect(url_for('show_entries'))
Security Note
Passwords should never be stored in plain text in a production system. This tutorial
uses plain text passwords for simplicity. If you plan to release a project based off this
tutorial out into the world, passwords should be both hashed and salted before being
stored in a database or file.
Fortunately, there are Flask extensions for the purpose of hashing passwords and verifying passwords against hashes, so adding this functionality is fairly straight forward.
There are also many general python libraries that can be used for hashing.
You can find a list of recommended Flask extensions here
Continue with Step 7: The Templates.
Step 7: The Templates
Now it is time to start working on the templates. As you may have noticed, if you
make requests with the app running, you will get an exception that Flask cannot find
the templates. The templates are using Jinja2 syntax and have autoescaping enabled
by default. This means that unless you mark a value in the code with Markup or with
the |safe filter in the template, Jinja2 will ensure that special characters such as < or >
are escaped with their XML equivalents.
We are also using template inheritance which makes it possible to reuse the layout of
the website in all pages.
Create the follwing three HTML files and place them in the templates folder:
This template contains the HTML skeleton, the header and a link to log in (or log out
if the user was already logged in). It also displays the flashed messages if there are
any. The {% block body %} block can be replaced by a block of the same name (body)
in a child template.
The session dict is available in the template as well and you can use that to check
if the user is logged in or not. Note that in Jinja you can access missing attributes
and items of objects / dicts which makes the following code work, even if there is no
'logged_in' key in the session:
<!doctype html>
<link rel=stylesheet type=text/css href="{{ url_for('static', filename='style.css
,→') }}">
<div class=page>
<div class=metanav>
{% if not session.logged_in %}
<a href="{{ url_for('login') }}">log in</a>
{% else %}
<a href="{{ url_for('logout') }}">log out</a>
{% endif %}
{% for message in get_flashed_messages() %}
<div class=flash>{{ message }}</div>
{% endfor %}
{% block body %}{% endblock %}
This template extends the layout.html template from above to display the messages. Note that the for loop iterates over the messages we passed in with the
render_template() function. Notice that the form is configured to submit to the
add_entry view function and use POST as HTTP method:
{% extends "layout.html" %}
{% block body %}
{% if session.logged_in %}
<form action="{{ url_for('add_entry') }}" method=post class=add-entry>
<dd><input type=text size=30 name=title>
<dd><textarea name=text rows=5 cols=40></textarea>
<dd><input type=submit value=Share>
{% endif %}
<ul class=entries>
{% for entry in entries %}
<li><h2>{{ entry.title }}</h2>{{ entry.text|safe }}</li>
{% else %}
<li><em>Unbelievable. No entries here so far</em></li>
{% endfor %}
{% endblock %}
This is the login template, which basically just displays a form to allow the user to
{% extends "layout.html" %}
{% block body %}
{% if error %}<p class=error><strong>Error:</strong> {{ error }}{% endif %}
<form action="{{ url_for('login') }}" method=post>
<dd><input type=text name=username>
<dd><input type=password name=password>
<dd><input type=submit value=Login>
{% endblock %}
Continue with Step 8: Adding Style.
Step 8: Adding Style
Now that everything else works, it’s time to add some style to the application. Just
create a stylesheet called style.css in the static folder:
a, h1, h2
h1, h2
font-family: sans-serif; background: #eee; }
color: #377ba8; }
font-family: 'Georgia', serif; margin: 0; }
border-bottom: 2px solid #eee; }
font-size: 1.2em; }
{ margin: 2em auto; width: 35em; border: 5px solid #ccc;
padding: 0.8em; background: white; }
{ list-style: none; margin: 0; padding: 0; }
.entries li
{ margin: 0.8em 1.2em; }
.entries li h2 { margin-left: -1em; }
{ font-size: 0.9em; border-bottom: 1px solid #ccc; }
.add-entry dl
{ font-weight: bold; }
{ text-align: right; font-size: 0.8em; padding: 0.3em;
margin-bottom: 1em; background: #fafafa; }
{ background: #cee5F5; padding: 0.5em;
border: 1px solid #aacbe2; }
{ background: #f0d6d6; padding: 0.5em; }
Continue with Bonus: Testing the Application.
Bonus: Testing the Application
Now that you have finished the application and everything works as expected, it’s
probably not a bad idea to add automated tests to simplify modifications in the future.
The application above is used as a basic example of how to perform unit testing in the
Testing Flask Applications section of the documentation. Go there to see how easy it is
to test Flask applications.
Adding tests to flaskr
Assuming you have seen the Testing Flask Applications section and have either written
your own tests for flaskr or have followed along with the examples provided, you
might be wondering about ways to organize the project.
One possible and recommended project structure is:
For now go ahead a create the tests/ directory as well as the test_flaskr.py file.
Running the tests
At this point you can run the tests. Here pytest will be used.
Note: Make sure that pytest is installed in the same virtualenv as flaskr. Otherwise
pytest test will not be able to import the required components to test the application:
pip install -e .
pip install pytest
Run and watch the tests pass, within the top-level flaskr/ directory as:
Testing + setuptools
One way to handle testing is to integrate it with setuptools. Here that requires adding
a couple of lines to the setup.py file and creating a new file setup.cfg. One benefit
of running the tests this way is that you do not have to install pytest. Go ahead and
update the setup.py file to contain:
from setuptools import setup
Now create setup.cfg in the project root (alongside setup.py):
Now you can run:
python setup.py test
This calls on the alias created in setup.cfg which in turn runs pytest via
pytest-runner, as the setup.py script has been called. (Recall the setup_requires argument in setup.py) Following the standard rules of test-discovery your tests will be
found, run, and hopefully pass.
This is one possible way to run and manage testing. Here pytest is used, but there are
other options such as nose. Integrating testing with setuptools is convenient because
it is not necessary to actually download pytest or any other testing framework one
might use.
Flask leverages Jinja2 as template engine. You are obviously free to use a different template engine, but you still have to install Jinja2 to run Flask itself. This requirement is
necessary to enable rich extensions. An extension can depend on Jinja2 being present.
This section only gives a very quick introduction into how Jinja2 is integrated into
Flask. If you want information on the template engine’s syntax itself, head over to the
official Jinja2 Template Documentation for more information.
Jinja Setup
Unless customized, Jinja2 is configured by Flask as follows:
• autoescaping is enabled for all templates ending in .html, .htm, .xml as well as
.xhtml when using render_template().
• autoescaping is enabled for all strings when using render_template_string().
• a template has the ability to opt in/out autoescaping with the {% autoescape %}
• Flask inserts a couple of global functions and helpers into the Jinja2 context, additionally to the values that are present by default.
Standard Context
The following global variables are available within Jinja2 templates by default:
The current configuration object (flask.config)
New in version 0.6.
Changed in version 0.10: This is now always available, even in imported templates.
The current request object (flask.request). This variable is unavailable if the
template was rendered without an active request context.
The current session object (flask.session). This variable is unavailable if the
template was rendered without an active request context.
The request-bound object for global variables (flask.g). This variable is unavailable if the template was rendered without an active request context.
The flask.url_for() function.
The flask.get_flashed_messages() function.
The Jinja Context Behavior
These variables are added to the context of variables, they are not global variables.
The difference is that by default these will not show up in the context of imported
templates. This is partially caused by performance considerations, partially to keep
things explicit.
What does this mean for you? If you have a macro you want to import, that needs to
access the request object you have two possibilities:
1. you explicitly pass the request to the macro as parameter, or the attribute of the
request object you are interested in.
2. you import the macro “with context”.
Importing with context looks like this:
{% from '_helpers.html' import my_macro with context %}
Standard Filters
These filters are available in Jinja2 additionally to the filters provided by Jinja2 itself:
This function converts the given object into JSON representation. This is for
example very helpful if you try to generate JavaScript on the fly.
Note that inside script tags no escaping must take place, so make sure to disable
escaping with |safe before Flask 0.10 if you intend to use it inside script tags:
<script type=text/javascript>
doSomethingWith({{ user.username|tojson|safe }});
Controlling Autoescaping
Autoescaping is the concept of automatically escaping special characters for you. Special characters in the sense of HTML (or XML, and thus XHTML) are &, >, <, " as well
as '. Because these characters carry specific meanings in documents on their own you
have to replace them by so called “entities” if you want to use them for text. Not doing
so would not only cause user frustration by the inability to use these characters in text,
but can also lead to security problems. (see Cross-Site Scripting (XSS))
Sometimes however you will need to disable autoescaping in templates. This can be
the case if you want to explicitly inject HTML into pages, for example if they come
from a system that generates secure HTML like a markdown to HTML converter.
There are three ways to accomplish that:
• In the Python code, wrap the HTML string in a Markup object before passing it to
the template. This is in general the recommended way.
• Inside the template, use the |safe filter to explicitly mark a string as safe HTML
({{ myvariable|safe }})
• Temporarily disable the autoescape system altogether.
To disable the autoescape system in templates, you can use the {% autoescape %}
{% autoescape false %}
<p>autoescaping is disabled here
<p>{{ will_not_be_escaped }}
{% endautoescape %}
Whenever you do this, please be very cautious about the variables you are using in
this block.
Registering Filters
If you want to register your own filters in Jinja2 you have two ways to do that.
You can either put them by hand into the jinja_env of the application or use the
template_filter() decorator.
The two following examples work the same and both reverse an object:
def reverse_filter(s):
return s[::-1]
def reverse_filter(s):
return s[::-1]
app.jinja_env.filters['reverse'] = reverse_filter
In case of the decorator the argument is optional if you want to use the function name
as name of the filter. Once registered, you can use the filter in your templates in the
same way as Jinja2’s builtin filters, for example if you have a Python list in context
called mylist:
{% for x in mylist | reverse %}
{% endfor %}
Context Processors
To inject new variables automatically into the context of a template, context processors
exist in Flask. Context processors run before the template is rendered and have the
ability to inject new values into the template context. A context processor is a function
that returns a dictionary. The keys and values of this dictionary are then merged with
the template context, for all templates in the app:
def inject_user():
return dict(user=g.user)
The context processor above makes a variable called user available in the template
with the value of g.user. This example is not very interesting because g is available in
templates anyways, but it gives an idea how this works.
Variables are not limited to values; a context processor can also make functions available to templates (since Python allows passing around functions):
def utility_processor():
def format_price(amount, currency=u'€'):
return u'{0:.2f}{1}'.format(amount, currency)
return dict(format_price=format_price)
The context processor above makes the format_price function available to all templates:
{{ format_price(0.33) }}
You could also build format_price as a template filter (see Registering Filters), but this
demonstrates how to pass functions in a context processor.
Testing Flask Applications
Something that is untested is broken.
The origin of this quote is unknown and while it is not entirely correct, it is also not
far from the truth. Untested applications make it hard to improve existing code and
developers of untested applications tend to become pretty paranoid. If an application
has automated tests, you can safely make changes and instantly know if anything
Flask provides a way to test your application by exposing the Werkzeug test Client
and handling the context locals for you. You can then use that with your favourite
testing solution.
In this documentation we will use the pytest package as the base framework for our
tests. You can install it with pip, like so:
pip install pytest
The Application
First, we need an application to test; we will use the application from the Tutorial. If
you don’t have that application yet, get the source code from the examples.
The Testing Skeleton
We begin by adding a tests directory under the application root. Then create a Python
file to store our tests (test_flaskr.py). When we format the filename like test_*.py,
it will be auto-discoverable by pytest.
Next, we create a pytest fixture called client() that configures the application for
testing and initializes a new database.:
import os
import tempfile
import pytest
from flaskr import flaskr
def client():
db_fd, flaskr.app.config['DATABASE'] = tempfile.mkstemp()
flaskr.app.config['TESTING'] = True
client = flaskr.app.test_client()
with flaskr.app.app_context():
yield client
This client fixture will be called by each individual test. It gives us a simple interface
to the application, where we can trigger test requests to the application. The client will
also keep track of cookies for us.
During setup, the TESTING config flag is activated. What this does is disable error catching during request handling, so that you get better error reports when performing test
requests against the application.
Because SQLite3 is filesystem-based, we can easily use the tempfile module to create
a temporary database and initialize it. The mkstemp() function does two things for
us: it returns a low-level file handle and a random file name, the latter we use as
database name. We just have to keep the db_fd around so that we can use the os.
close() function to close the file.
To delete the database after the test, the fixture closes the file and removes it from the
If we now run the test suite, we should see the following output:
$ pytest
================ test session starts ================
rootdir: ./flask/examples/flaskr, inifile: setup.cfg
collected 0 items
=========== no tests ran in 0.07 seconds ============
Even though it did not run any actual tests, we already know that our flaskr application is syntactically valid, otherwise the import would have died with an exception.
The First Test
Now it’s time to start testing the functionality of the application. Let’s check that the
application shows “No entries here so far” if we access the root of the application (/).
To do this, we add a new test function to test_flaskr.py, like this:
def test_empty_db(client):
"""Start with a blank database."""
rv = client.get('/')
assert b'No entries here so far' in rv.data
Notice that our test functions begin with the word test; this allows pytest to automatically identify the function as a test to run.
By using client.get we can send an HTTP GET request to the application with the
given path. The return value will be a response_class object. We can now use the
data attribute to inspect the return value (as string) from the application. In this case,
we ensure that 'No entries here so far' is part of the output.
Run it again and you should see one passing test:
$ pytest -v
================ test session starts ================
rootdir: ./flask/examples/flaskr, inifile: setup.cfg
collected 1 items
tests/test_flaskr.py::test_empty_db PASSED
============= 1 passed in 0.10 seconds ==============
Logging In and Out
The majority of the functionality of our application is only available for the administrative user, so we need a way to log our test client in and out of the application. To do
this, we fire some requests to the login and logout pages with the required form data
(username and password). And because the login and logout pages redirect, we tell
the client to follow_redirects.
Add the following two functions to your test_flaskr.py file:
def login(client, username, password):
return client.post('/login', data=dict(
), follow_redirects=True)
def logout(client):
return client.get('/logout', follow_redirects=True)
Now we can easily test that logging in and out works and that it fails with invalid
credentials. Add this new test function:
def test_login_logout(client):
"""Make sure login and logout works."""
rv = login(client, flaskr.app.config['USERNAME'], flaskr.app.config['PASSWORD
assert b'You were logged in' in rv.data
rv = logout(client)
assert b'You were logged out' in rv.data
rv = login(client, flaskr.app.config['USERNAME'] + 'x', flaskr.app.config[
assert b'Invalid username' in rv.data
rv = login(client, flaskr.app.config['USERNAME'], flaskr.app.config['PASSWORD
'] + 'x')
assert b'Invalid password' in rv.data
Test Adding Messages
We should also test that adding messages works. Add a new test function like this:
def test_messages(client):
"""Test that messages work."""
login(client, flaskr.app.config['USERNAME'], flaskr.app.config['PASSWORD'])
rv = client.post('/add', data=dict(
text='<strong>HTML</strong> allowed here'
), follow_redirects=True)
assert b'No entries here so far' not in rv.data
assert b'&lt;Hello&gt;' in rv.data
assert b'<strong>HTML</strong> allowed here' in rv.data
Here we check that HTML is allowed in the text but not in the title, which is the in54
tended behavior.
Running that should now give us three passing tests:
$ pytest -v
================ test session starts ================
rootdir: ./flask/examples/flaskr, inifile: setup.cfg
collected 3 items
tests/test_flaskr.py::test_empty_db PASSED
tests/test_flaskr.py::test_login_logout PASSED
tests/test_flaskr.py::test_messages PASSED
============= 3 passed in 0.23 seconds ==============
For more complex tests with headers and status codes, check out the MiniTwit Example from the sources which contains a larger test suite.
Other Testing Tricks
Besides using the test client as shown above, there is also the test_request_context()
method that can be used in combination with the with statement to activate a request
context temporarily. With this you can access the request, g and session objects like
in view functions. Here is a full example that demonstrates this approach:
import flask
app = flask.Flask(__name__)
with app.test_request_context('/?name=Peter'):
assert flask.request.path == '/'
assert flask.request.args['name'] == 'Peter'
All the other objects that are context bound can be used in the same way.
If you want to test your application with different configurations and there does not
seem to be a good way to do that, consider switching to application factories (see
Application Factories).
Note however that if you are using a test request context, the before_request()
and after_request() functions are not called automatically.
teardown_request() functions are indeed executed when the test request context leaves the with block. If you do want the before_request() functions to be called
as well, you need to call preprocess_request() yourself:
app = flask.Flask(__name__)
with app.test_request_context('/?name=Peter'):
This can be necessary to open database connections or something similar depending
on how your application was designed.
If you want to call the after_request() functions you need to call into
process_response() which however requires that you pass it a response object:
app = flask.Flask(__name__)
with app.test_request_context('/?name=Peter'):
resp = Response('...')
resp = app.process_response(resp)
This in general is less useful because at that point you can directly start using the test
Faking Resources and Context
New in version 0.10.
A very common pattern is to store user authorization information and database connections on the application context or the flask.g object. The general pattern for this is
to put the object on there on first usage and then to remove it on a teardown. Imagine
for instance this code to get the current user:
def get_user():
user = getattr(g, 'user', None)
if user is None:
user = fetch_current_user_from_database()
g.user = user
return user
For a test it would be nice to override this user from the outside without having to change some code. This can be accomplished with hooking the flask.
appcontext_pushed signal:
from contextlib import contextmanager
from flask import appcontext_pushed, g
def user_set(app, user):
def handler(sender, **kwargs):
g.user = user
with appcontext_pushed.connected_to(handler, app):
And then to use it:
from flask import json, jsonify
def users_me():
return jsonify(username=g.user.username)
with user_set(app, my_user):
with app.test_client() as c:
resp = c.get('/users/me')
data = json.loads(resp.data)
self.assert_equal(data['username'], my_user.username)
Keeping the Context Around
New in version 0.4.
Sometimes it is helpful to trigger a regular request but still keep the context around
for a little longer so that additional introspection can happen. With Flask 0.4 this is
possible by using the test_client() with a with block:
app = flask.Flask(__name__)
with app.test_client() as c:
rv = c.get('/?tequila=42')
assert request.args['tequila'] == '42'
If you were to use just the test_client() without the with block, the assert would
fail with an error because request is no longer available (because you are trying to use
it outside of the actual request).
Accessing and Modifying Sessions
New in version 0.8.
Sometimes it can be very helpful to access or modify the sessions from the test client.
Generally there are two ways for this. If you just want to ensure that a session has
certain keys set to certain values you can just keep the context around and access
with app.test_client() as c:
rv = c.get('/')
assert flask.session['foo'] == 42
This however does not make it possible to also modify the session or to access the session before a request was fired. Starting with Flask 0.8 we provide a so called “session
transaction” which simulates the appropriate calls to open a session in the context of
the test client and to modify it. At the end of the transaction the session is stored. This
works independently of the session backend used:
with app.test_client() as c:
with c.session_transaction() as sess:
sess['a_key'] = 'a value'
# once this is reached the session was stored
Note that in this case you have to use the sess object instead of the flask.session
proxy. The object however itself will provide the same interface.
Application Errors
New in version 0.3.
Applications fail, servers fail. Sooner or later you will see an exception in production.
Even if your code is 100% correct, you will still see exceptions from time to time. Why?
Because everything else involved will fail. Here are some situations where perfectly
fine code can lead to server errors:
• the client terminated the request early and the application was still reading from
the incoming data
• the database server was overloaded and could not handle the query
• a filesystem is full
• a harddrive crashed
• a backend server overloaded
• a programming error in a library you are using
• network connection of the server to another system failed
And that’s just a small sample of issues you could be facing. So how do we deal with
that sort of problem? By default if your application runs in production mode, Flask
will display a very simple page for you and log the exception to the logger.
But there is more you can do, and we will cover some better setups to deal with errors.
Error Logging Tools
Sending error mails, even if just for critical ones, can become overwhelming if enough
users are hitting the error and log files are typically never looked at. This is why
we recommend using Sentry for dealing with application errors. It’s available as an
Open Source project on GitHub and is also available as a hosted version which you
can try for free. Sentry aggregates duplicate errors, captures the full stack trace and
local variables for debugging, and sends you mails based on new errors or frequency
To use Sentry you need to install the raven client:
$ pip install raven
And then add this to your Flask app:
from raven.contrib.flask import Sentry
sentry = Sentry(app, dsn='YOUR_DSN_HERE')
Or if you are using factories you can also init it later:
from raven.contrib.flask import Sentry
sentry = Sentry(dsn='YOUR_DSN_HERE')
def create_app():
app = Flask(__name__)
return app
The YOUR_DSN_HERE value needs to be replaced with the DSN value you get from
your Sentry installation.
Afterwards failures are automatically reported to Sentry and from there you can receive error notifications.
Error handlers
You might want to show custom error pages to the user when an error occurs. This
can be done by registering error handlers.
Error handlers are normal Pluggable Views but instead of being registered for routes,
they are registered for exceptions that are raised while trying to do something else.
Register error handlers using errorhandler() or register_error_handler():
def handle_bad_request(e):
return 'bad request!'
app.register_error_handler(400, lambda e: 'bad request!')
Those two ways are equivalent, but the first one is more clear and leaves you with
a function to call on your whim (and in tests). Note that werkzeug.exceptions.
HTTPException subclasses like BadRequest from the example and their HTTP codes are
interchangeable when handed to the registration methods or decorator (BadRequest.
code == 400).
You are however not limited to HTTPException or HTTP status codes but can register
a handler for every exception class you like.
Changed in version 0.11: Errorhandlers are now prioritized by specificity of the exception classes they are registered for instead of the order they are registered in.
Once an exception instance is raised, its class hierarchy is traversed, and searched for
in the exception classes for which handlers are registered. The most specific handler is
if an instance of ConnectionRefusedError is raised, and a handler is
registered for ConnectionError and ConnectionRefusedError, the more specific
ConnectionRefusedError handler is called on the exception instance, and its response
is shown to the user.
Error Mails
If the application runs in production mode (which it will do on your server) you might
not see any log messages. The reason for that is that Flask by default will just report
to the WSGI error stream or stderr (depending on what’s available). Where this ends
up is sometimes hard to find. Often it’s in your webserver’s log files.
I can pretty much promise you however that if you only use a logfile for the application
errors you will never look at it except for debugging an issue when a user reported it
for you. What you probably want instead is a mail the second the exception happened.
Then you get an alert and you can do something about it.
Flask uses the Python builtin logging system, and it can actually send you mails for
errors which is probably what you want. Here is how you can configure the Flask
logger to send you mails for exceptions:
ADMINS = ['[email protected]']
if not app.debug:
import logging
from logging.handlers import SMTPHandler
mail_handler = SMTPHandler('',
'[email protected]',
ADMINS, 'YourApplication Failed')
So what just happened? We created a new SMTPHandler that will send mails with
the mail server listening on to all the ADMINS from the address [email protected] with the subject “YourApplication Failed”. If your mail server requires credentials, these can also be provided. For that check out the documentation
for the SMTPHandler.
We also tell the handler to only send errors and more critical messages. Because we
certainly don’t want to get a mail for warnings or other useless logs that might happen
during request handling.
Before you run that in production, please also look at Controlling the Log Format to put
more information into that error mail. That will save you from a lot of frustration.
Logging to a File
Even if you get mails, you probably also want to log warnings. It’s a good idea to keep
as much information around that might be required to debug a problem. By default
as of Flask 0.11, errors are logged to your webserver’s log automatically. Warnings
however are not. Please note that Flask itself will not issue any warnings in the core
system, so it’s your responsibility to warn in the code if something seems odd.
There are a couple of handlers provided by the logging system out of the box but not
all of them are useful for basic error logging. The most interesting are probably the
• FileHandler - logs messages to a file on the filesystem.
• RotatingFileHandler - logs messages to a file on the filesystem and will rotate
after a certain number of messages.
• NTEventLogHandler - will log to the system event log of a Windows system. If
you are deploying on a Windows box, this is what you want to use.
• SysLogHandler - sends logs to a UNIX syslog.
Once you picked your log handler, do like you did with the SMTP handler above, just
make sure to use a lower setting (I would recommend WARNING):
if not app.debug:
import logging
from themodule import TheHandlerYouWant
file_handler = TheHandlerYouWant(...)
Controlling the Log Format
By default a handler will only write the message string into a file or send you that
message as mail. A log record stores more information, and it makes a lot of sense to
configure your logger to also contain that information so that you have a better idea
of why that error happened, and more importantly, where it did.
A formatter can be instantiated with a format string. Note that tracebacks are appended to the log entry automatically. You don’t have to do that in the log formatter
format string.
Here are some example setups:
from logging import Formatter
Message type:
File logging
from logging import Formatter
'%(asctime)s %(levelname)s: %(message)s '
'[in %(pathname)s:%(lineno)d]'
Complex Log Formatting
Here is a list of useful formatting variables for the format string. Note that this list is
not complete, consult the official documentation of the logging package for a full list.
Text logging level for the message ('DEBUG', 'INFO', 'WARNING',
Full pathname of the source file where the logging call was issued
(if available).
Filename portion of pathname.
Module (name portion of filename).
Name of function containing the logging call.
Source line number where the logging call was issued (if available).
Human-readable time when the LogRecord was created. By default this is of the form "2003-07-08 16:49:45,896" (the numbers after the comma are millisecond portion of the time). This
can be changed by subclassing the formatter and overriding the
formatTime() method.
The logged message, computed as msg % args
If you want to further customize the formatting, you can subclass the formatter. The
formatter has three interesting methods:
format(): handles the actual formatting. It is passed a LogRecord object and has to
return the formatted string.
formatTime(): called for asctime formatting. If you want a different time format you
can override this method.
formatException() called for exception formatting. It is passed an exc_info tuple and
has to return a string. The default is usually fine, you don’t have to override it.
For more information, head over to the official documentation.
Other Libraries
So far we only configured the logger your application created itself. Other libraries
might log themselves as well. For example, SQLAlchemy uses logging heavily in its
core. While there is a method to configure all loggers at once in the logging package, I
would not recommend using it. There might be a situation in which you want to have
multiple separate applications running side by side in the same Python interpreter and
then it becomes impossible to have different logging setups for those.
Instead, I would recommend figuring out which loggers you are interested in, getting
the loggers with the getLogger() function and iterating over them to attach handlers:
from logging import getLogger
loggers = [app.logger, getLogger('sqlalchemy'),
for logger in loggers:
Debugging Application Errors
For production applications, configure your application with logging and notifications
as described in Application Errors. This section provides pointers when debugging
deployment configuration and digging deeper with a full-featured Python debugger.
When in Doubt, Run Manually
Having problems getting your application configured for production? If you have
shell access to your host, verify that you can run your application manually from the
shell in the deployment environment. Be sure to run under the same user account
as the configured deployment to troubleshoot permission issues. You can use Flask’s
builtin development server with debug=True on your production host, which is helpful
in catching configuration issues, but be sure to do this temporarily in a controlled
environment. Do not run in production with debug=True.
Working with Debuggers
To dig deeper, possibly to trace code execution, Flask provides a debugger out of the
box (see Debug Mode). If you would like to use another Python debugger, note that
debuggers interfere with each other. You have to set some options in order to use your
favorite debugger:
• debug - whether to enable debug mode and catch exceptions
• use_debugger - whether to use the internal Flask debugger
• use_reloader - whether to reload and fork the process on exception
debug must be True (i.e., exceptions must be caught) in order for the other two options
to have any value.
If you’re using Aptana/Eclipse for debugging you’ll need to set both use_debugger
and use_reloader to False.
A possible useful pattern for configuration is to set the following in your config.yaml
(change the block as appropriate for your application, of course):
Then in your application’s entry-point (main.py), you could have something like:
if __name__ == "__main__":
# To allow aptana to receive errors, set use_debugger=False
app = create_app(config="config.yaml")
if app.debug: use_debugger = True
# Disable Flask's debugger if external debugger is requested
use_debugger = not(app.config.get('DEBUG_WITH_APTANA'))
app.run(use_debugger=use_debugger, debug=app.debug,
use_reloader=use_debugger, host='')
Configuration Handling
New in version 0.3.
Applications need some kind of configuration. There are different settings you might
want to change depending on the application environment like toggling the debug
mode, setting the secret key, and other such environment-specific things.
The way Flask is designed usually requires the configuration to be available when the
application starts up. You can hardcode the configuration in the code, which for many
small applications is not actually that bad, but there are better ways.
Independent of how you load your config, there is a config object available which
holds the loaded configuration values: The config attribute of the Flask object. This
is the place where Flask itself puts certain configuration values and also where extensions can put their configuration values. But this is also where you can have your own
Configuration Basics
The config is actually a subclass of a dictionary and can be modified just like any
app = Flask(__name__)
app.config['DEBUG'] = True
Certain configuration values are also forwarded to the Flask object so you can read
and write them from there:
app.debug = True
To update multiple keys at once you can use the dict.update() method:
Debug Mode with the flask Script
If you use the flask script to start a local development server, to enable the debug
mode, you need to export the FLASK_DEBUG environment variable before running the
$ export FLASK_DEBUG=1
$ flask run
(On Windows you need to use set instead of export).
app.debug and app.config['DEBUG'] are not compatible with the flask script. They
only worked when using Flask.run() method.
Builtin Configuration Values
The following configuration values are used internally by Flask:
enable/disable debug mode when using
Flask.run() method to start server
enable/disable testing mode
explicitly enable or disable the propagation of
exceptions. If not set or explicitly set to None
this is implicitly true if either TESTING or DEBUG
is true.
By default if the application is in debug mode
the request context is not popped on exceptions to enable debuggers to introspect the
data. This can be disabled by this key. You
can also use this setting to force-enable it for
non debug execution which might be useful to
debug production applications (but also very
the secret key
the name of the session cookie
the domain for the session cookie. If this is not
set, the cookie will be valid for all subdomains
the path for the session cookie. If this is
not set the cookie will be valid for all of
APPLICATION_ROOT or if that is not set for '/'.
controls if the cookie should be set with the
httponly flag. Defaults to True.
controls if the cookie should be set with the
secure flag. Defaults to False.
the lifetime of a permanent session as
datetime.timedelta object. Starting with
Flask 0.8 this can also be an integer representing seconds.
this flag controls how permanent sessions are
refreshed. If set to True (which is the default) then the cookie is refreshed each request
which automatically bumps the lifetime. If set
to False a set-cookie header is only sent if the
session is modified. Non permanent sessions
are not affected by this.
enable/disable x-sendfile
the name of the logger
the policy of the default logging handler.
The default is 'always' which means that
the default logging handler is always active.
'debug' will only activate logging in debug
mode, 'production' will only log in production and 'never' disables it entirely.
the name and port number of the server. Required for subdomain support (e.g.: 'myapp.
dev:5000') Note that localhost does not support subdomains so setting this to “localhost”69
does not help. Setting a SERVER_NAME also by
default enables URL generation without a request context but with an application context.
The SERVER_NAME key is used for the subdomain support. Because Flask cannot guess
the subdomain part without the knowledge of the actual server name, this is required
if you want to work with subdomains. This is also used for the session cookie.
Please keep in mind that not only Flask has the problem of not knowing what subdomains are, your web browser does as well. Most modern web browsers will not
allow cross-subdomain cookies to be set on a server name without dots in it. So if
your server name is 'localhost' you will not be able to set a cookie for 'localhost'
and every subdomain of it. Please choose a different server name in that case, like
'myapplication.local' and add this name + the subdomains you want to use into
your host config or setup a local bind.
New in version 0.4: LOGGER_NAME
New in version 0.5: SERVER_NAME
New in version 0.6: MAX_CONTENT_LENGTH
New in version 0.8:
New in version 0.9: PREFERRED_URL_SCHEME
Configuring from Files
Configuration becomes more useful if you can store it in a separate file, ideally located
outside the actual application package. This makes packaging and distributing your
application possible via various package handling tools (Deploying with Setuptools) and
finally modifying the configuration file afterwards.
So a common pattern is this:
app = Flask(__name__)
This first loads the configuration from the yourapplication.default_settings module and
then overrides the values with the contents of the file the YOURAPPLICATION_SETTINGS
environment variable points to. This environment variable can be set on Linux or OS
X with the export command in the shell before starting the server:
$ export YOURAPPLICATION_SETTINGS=/path/to/settings.cfg
$ python run-app.py
* Running on
* Restarting with reloader...
On Windows systems use the set builtin instead:
>set YOURAPPLICATION_SETTINGS=\path\to\settings.cfg
The configuration files themselves are actual Python files. Only values in uppercase
are actually stored in the config object later on. So make sure to use uppercase letters
for your config keys.
Here is an example of a configuration file:
# Example configuration
DEBUG = False
SECRET_KEY = '?\xbf,\xb4\x8d\xa3"<\x9c\[email protected]\x0f5\xab,w\xee\x8d$0\x13\x8b83'
Make sure to load the configuration very early on, so that extensions have the ability
to access the configuration when starting up. There are other methods on the config
object as well to load from individual files. For a complete reference, read the Config
object’s documentation.
Configuration Best Practices
The downside with the approach mentioned earlier is that it makes testing a little
harder. There is no single 100% solution for this problem in general, but there are a
couple of things you can keep in mind to improve that experience:
1. Create your application in a function and register blueprints on it. That way
you can create multiple instances of your application with different configurations attached which makes unittesting a lot easier. You can use this to pass in
configuration as needed.
2. Do not write code that needs the configuration at import time. If you limit yourself to request-only accesses to the configuration you can reconfigure the object
later on as needed.
Development / Production
Most applications need more than one configuration. There should be at least separate
configurations for the production server and the one used during development. The
easiest way to handle this is to use a default configuration that is always loaded and
part of the version control, and a separate configuration that overrides the values as
necessary as mentioned in the example above:
app = Flask(__name__)
Then you just have to add a separate config.py file and export
YOURAPPLICATION_SETTINGS=/path/to/config.py and you are done.
there are alternative ways as well. For example you could use imports or subclassing.
What is very popular in the Django world is to make the import explicit in the config
file by adding from yourapplication.default_settings import * to the top of the file
and then overriding the changes by hand. You could also inspect an environment variable like YOURAPPLICATION_MODE and set that to production, development etc and import
different hardcoded files based on that.
An interesting pattern is also to use classes and inheritance for configuration:
class Config(object):
DEBUG = False
DATABASE_URI = 'sqlite://:memory:'
class ProductionConfig(Config):
DATABASE_URI = 'mysql://[email protected]/foo'
class DevelopmentConfig(Config):
DEBUG = True
class TestingConfig(Config):
To enable such a config you just have to call into from_object():
There are many different ways and it’s up to you how you want to manage your configuration files. However here a list of good recommendations:
• Keep a default configuration in version control. Either populate the config with
this default configuration or import it in your own configuration files before
overriding values.
• Use an environment variable to switch between the configurations. This can be
done from outside the Python interpreter and makes development and deployment much easier because you can quickly and easily switch between different
configs without having to touch the code at all. If you are working often on different projects you can even create your own script for sourcing that activates a
virtualenv and exports the development configuration for you.
• Use a tool like fabric in production to push code and configurations separately
to the production server(s). For some details about how to do that, head over to
the Deploying with Fabric pattern.
Instance Folders
New in version 0.8.
Flask 0.8 introduces instance folders. Flask for a long time made it possible to refer to
paths relative to the application’s folder directly (via Flask.root_path). This was also
how many developers loaded configurations stored next to the application. Unfortunately however this only works well if applications are not packages in which case the
root path refers to the contents of the package.
With Flask 0.8 a new attribute was introduced: Flask.instance_path. It refers to a
new concept called the “instance folder”. The instance folder is designed to not be
under version control and be deployment specific. It’s the perfect place to drop things
that either change at runtime or configuration files.
You can either explicitly provide the path of the instance folder when creating the
Flask application or you can let Flask autodetect the instance folder. For explicit configuration use the instance_path parameter:
app = Flask(__name__, instance_path='/path/to/instance/folder')
Please keep in mind that this path must be absolute when provided.
If the instance_path parameter is not provided the following default locations are used:
• Uninstalled module:
• Uninstalled package:
• Installed module or package:
$PREFIX is the prefix of your Python installation. This can be /usr or the path to
your virtualenv. You can print the value of sys.prefix to see what the prefix is
set to.
Since the config object provided loading of configuration files from relative filenames
we made it possible to change the loading via filenames to be relative to the instance
path if wanted. The behavior of relative paths in config files can be flipped between
“relative to the application root” (the default) to “relative to instance folder” via the
instance_relative_config switch to the application constructor:
app = Flask(__name__, instance_relative_config=True)
Here is a full example of how to configure Flask to preload the config from a module
and then override the config from a file in the config folder if it exists:
app = Flask(__name__, instance_relative_config=True)
app.config.from_pyfile('application.cfg', silent=True)
The path to the instance folder can be found via the Flask.instance_path. Flask
also provides a shortcut to open a file from the instance folder with Flask.
Example usage for both:
filename = os.path.join(app.instance_path, 'application.cfg')
with open(filename) as f:
config = f.read()
# or via open_instance_resource:
with app.open_instance_resource('application.cfg') as f:
config = f.read()
New in version 0.6.
Starting with Flask 0.6, there is integrated support for signalling in Flask. This support
is provided by the excellent blinker library and will gracefully fall back if it is not
What are signals? Signals help you decouple applications by sending notifications
when actions occur elsewhere in the core framework or another Flask extensions. In
short, signals allow certain senders to notify subscribers that something happened.
Flask comes with a couple of signals and other extensions might provide more. Also
keep in mind that signals are intended to notify subscribers and should not encourage
subscribers to modify data. You will notice that there are signals that appear to do
the same thing like some of the builtin decorators do (eg: request_started is very
similar to before_request()). However, there are differences in how they work. The
core before_request() handler, for example, is executed in a specific order and is able
to abort the request early by returning a response. In contrast all signal handlers are
executed in undefined order and do not modify any data.
The big advantage of signals over handlers is that you can safely subscribe to them
for just a split second. These temporary subscriptions are helpful for unit testing for
example. Say you want to know what templates were rendered as part of a request:
signals allow you to do exactly that.
Subscribing to Signals
To subscribe to a signal, you can use the connect() method of a signal. The first argument is the function that should be called when the signal is emitted, the optional
second argument specifies a sender. To unsubscribe from a signal, you can use the
disconnect() method.
For all core Flask signals, the sender is the application that issued the signal. When
you subscribe to a signal, be sure to also provide a sender unless you really want to
listen for signals from all applications. This is especially true if you are developing an
For example, here is a helper context manager that can be used in a unit test to determine which templates were rendered and what variables were passed to the template:
from flask import template_rendered
from contextlib import contextmanager
def captured_templates(app):
recorded = []
def record(sender, template, context, **extra):
recorded.append((template, context))
template_rendered.connect(record, app)
yield recorded
template_rendered.disconnect(record, app)
This can now easily be paired with a test client:
with captured_templates(app) as templates:
rv = app.test_client().get('/')
assert rv.status_code == 200
assert len(templates) == 1
template, context = templates[0]
assert template.name == 'index.html'
assert len(context['items']) == 10
Make sure to subscribe with an extra **extra argument so that your calls don’t fail if
Flask introduces new arguments to the signals.
All the template rendering in the code issued by the application app in the body of
the with block will now be recorded in the templates variable. Whenever a template is
rendered, the template object as well as context are appended to it.
Additionally there is a convenient helper method (connected_to()) that allows you
to temporarily subscribe a function to a signal with a context manager on its own.
Because the return value of the context manager cannot be specified that way, you
have to pass the list in as an argument:
from flask import template_rendered
def captured_templates(app, recorded, **extra):
def record(sender, template, context):
recorded.append((template, context))
return template_rendered.connected_to(record, app)
The example above would then look like this:
templates = []
with captured_templates(app, templates, **extra):
template, context = templates[0]
Blinker API Changes
The connected_to() method arrived in Blinker with version 1.1.
Creating Signals
If you want to use signals in your own application, you can use the blinker library
directly. The most common use case are named signals in a custom Namespace.. This is
what is recommended most of the time:
from blinker import Namespace
my_signals = Namespace()
Now you can create new signals like this:
model_saved = my_signals.signal('model-saved')
The name for the signal here makes it unique and also simplifies debugging. You can
access the name of the signal with the name attribute.
For Extension Developers
If you are writing a Flask extension and you want to gracefully degrade for missing
blinker installations, you can do so by using the flask.signals.Namespace class.
Sending Signals
If you want to emit a signal, you can do so by calling the send() method. It accepts a
sender as first argument and optionally some keyword arguments that are forwarded
to the signal subscribers:
class Model(object):
def save(self):
Try to always pick a good sender. If you have a class that is emitting a signal, pass
self as sender. If you are emitting a signal from a random function, you can pass
current_app._get_current_object() as sender.
Passing Proxies as Senders
Never pass current_app as sender to a signal.
Use current_app.
_get_current_object() instead. The reason for this is that current_app is a
proxy and not the real application object.
Signals and Flask’s Request Context
Signals fully support The Request Context when receiving signals. Context-local variables are consistently available between request_started and request_finished, so
you can rely on flask.g and others as needed. Note the limitations described in Sending Signals and the request_tearing_down signal.
Decorator Based Signal Subscriptions
With Blinker 1.1 you can also easily subscribe to signals by using the new
connect_via() decorator:
from flask import template_rendered
def when_template_rendered(sender, template, context, **extra):
print 'Template %s is rendered with %s' % (template.name, context)
Core Signals
Take a look at Signals for a list of all builtin signals.
Pluggable Views
New in version 0.7.
Flask 0.7 introduces pluggable views inspired by the generic views from Django which
are based on classes instead of functions. The main intention is that you can replace
parts of the implementations and this way have customizable pluggable views.
Basic Principle
Consider you have a function that loads a list of objects from the database and renders
into a template:
def show_users(page):
users = User.query.all()
return render_template('users.html', users=users)
This is simple and flexible, but if you want to provide this view in a generic fashion
that can be adapted to other models and templates as well you might want more flexibility. This is where pluggable class-based views come into place. As the first step to
convert this into a class based view you would do this:
from flask.views import View
class ShowUsers(View):
def dispatch_request(self):
users = User.query.all()
return render_template('users.html', objects=users)
app.add_url_rule('/users/', view_func=ShowUsers.as_view('show_users'))
As you can see what you have to do is to create a subclass of flask.views.View and implement dispatch_request(). Then we have to convert that class into an actual view
function by using the as_view() class method. The string you pass to that function is
the name of the endpoint that view will then have. But this by itself is not helpful, so
let’s refactor the code a bit:
from flask.views import View
class ListView(View):
def get_template_name(self):
raise NotImplementedError()
def render_template(self, context):
return render_template(self.get_template_name(), **context)
def dispatch_request(self):
context = {'objects': self.get_objects()}
return self.render_template(context)
class UserView(ListView):
def get_template_name(self):
return 'users.html'
def get_objects(self):
return User.query.all()
This of course is not that helpful for such a small example, but it’s good enough to
explain the basic principle. When you have a class-based view the question comes up
what self points to. The way this works is that whenever the request is dispatched a
new instance of the class is created and the dispatch_request() method is called with
the parameters from the URL rule. The class itself is instantiated with the parameters
passed to the as_view() function. For instance you can write a class like this:
class RenderTemplateView(View):
def __init__(self, template_name):
self.template_name = template_name
def dispatch_request(self):
return render_template(self.template_name)
And then you can register it like this:
app.add_url_rule('/about', view_func=RenderTemplateView.as_view(
'about_page', template_name='about.html'))
Method Hints
Pluggable views are attached to the application like a regular function by either using
route() or better add_url_rule(). That however also means that you would have to
provide the names of the HTTP methods the view supports when you attach this. In
order to move that information to the class you can provide a methods attribute that
has this information:
class MyView(View):
methods = ['GET', 'POST']
def dispatch_request(self):
if request.method == 'POST':
app.add_url_rule('/myview', view_func=MyView.as_view('myview'))
Method Based Dispatching
For RESTful APIs it’s especially helpful to execute a different function for each HTTP
method. With the flask.views.MethodView you can easily do that. Each HTTP method
maps to a function with the same name (just in lowercase):
from flask.views import MethodView
class UserAPI(MethodView):
def get(self):
users = User.query.all()
def post(self):
user = User.from_form_data(request.form)
app.add_url_rule('/users/', view_func=UserAPI.as_view('users'))
That way you also don’t have to provide the methods attribute. It’s automatically set
based on the methods defined in the class.
Decorating Views
Since the view class itself is not the view function that is added to the routing system
it does not make much sense to decorate the class itself. Instead you either have to
decorate the return value of as_view() by hand:
def user_required(f):
"""Checks whether user is logged in or raises error 401."""
def decorator(*args, **kwargs):
if not g.user:
return f(*args, **kwargs)
return decorator
view = user_required(UserAPI.as_view('users'))
app.add_url_rule('/users/', view_func=view)
Starting with Flask 0.8 there is also an alternative way where you can specify a list of
decorators to apply in the class declaration:
class UserAPI(MethodView):
decorators = [user_required]
Due to the implicit self from the caller’s perspective you cannot use regular view decorators on the individual methods of the view however, keep this in mind.
Method Views for APIs
Web APIs are often working very closely with HTTP verbs so it makes a lot of sense
to implement such an API based on the MethodView. That said, you will notice that the
API will require different URL rules that go to the same method view most of the time.
For instance consider that you are exposing a user object on the web:
Gives a list of all users
Creates a new user
Shows a single user
Updates a single user
Deletes a single user
So how would you go about doing that with the MethodView? The trick is to take
advantage of the fact that you can provide multiple rules to the same view.
Let’s assume for the moment the view would look like this:
class UserAPI(MethodView):
def get(self, user_id):
if user_id is None:
# return a list of users
# expose a single user
def post(self):
# create a new user
def delete(self, user_id):
# delete a single user
def put(self, user_id):
# update a single user
So how do we hook this up with the routing system? By adding two rules and explicitly mentioning the methods for each:
user_view = UserAPI.as_view('user_api')
app.add_url_rule('/users/', defaults={'user_id': None},
view_func=user_view, methods=['GET',])
app.add_url_rule('/users/', view_func=user_view, methods=['POST',])
app.add_url_rule('/users/<int:user_id>', view_func=user_view,
methods=['GET', 'PUT', 'DELETE'])
If you have a lot of APIs that look similar you can refactor that registration code:
def register_api(view, endpoint, url, pk='id', pk_type='int'):
view_func = view.as_view(endpoint)
app.add_url_rule(url, defaults={pk: None},
view_func=view_func, methods=['GET',])
app.add_url_rule(url, view_func=view_func, methods=['POST',])
app.add_url_rule('%s<%s:%s>' % (url, pk_type, pk), view_func=view_func,
methods=['GET', 'PUT', 'DELETE'])
register_api(UserAPI, 'user_api', '/users/', pk='user_id')
The Application Context
New in version 0.9.
One of the design ideas behind Flask is that there are two different “states” in which
code is executed. The application setup state in which the application implicitly is on
the module level. It starts when the Flask object is instantiated, and it implicitly ends
when the first request comes in. While the application is in this state a few assumptions
are true:
• the programmer can modify the application object safely.
• no request handling happened so far
• you have to have a reference to the application object in order to modify it, there
is no magic proxy that can give you a reference to the application object you’re
currently creating or modifying.
In contrast, during request handling, a couple of other rules exist:
• while a request is active, the context local objects (flask.request and others)
point to the current request.
• any code can get hold of these objects at any time.
There is a third state which is sitting in between a little bit. Sometimes you are dealing
with an application in a way that is similar to how you interact with applications
during request handling; just that there is no request active. Consider, for instance,
that you’re sitting in an interactive Python shell and interacting with the application,
or a command line application.
The application context is what powers the current_app context local.
Purpose of the Application Context
The main reason for the application’s context existence is that in the past a bunch of
functionality was attached to the request context for lack of a better solution. Since
one of the pillars of Flask’s design is that you can have more than one application in
the same Python process.
So how does the code find the “right” application? In the past we recommended passing applications around explicitly, but that caused issues with libraries that were not
designed with that in mind.
A common workaround for that problem was to use the current_app proxy later on,
which was bound to the current request’s application reference. Since creating such
a request context is an unnecessarily expensive operation in case there is no request
around, the application context was introduced.
Creating an Application Context
There are two ways to make an application context. The first one is implicit: whenever
a request context is pushed, an application context will be created alongside if this is
necessary. As a result, you can ignore the existence of the application context unless
you need it.
The second way is the explicit way using the app_context() method:
from flask import Flask, current_app
app = Flask(__name__)
with app.app_context():
# within this block, current_app points to app.
print current_app.name
The application context is also used by the url_for() function in case a SERVER_NAME
was configured. This allows you to generate URLs even in the absence of a request.
If no request context has been pushed and an application context has not been explicitly set, a RuntimeError will be raised.
RuntimeError: Working outside of application context.
Locality of the Context
The application context is created and destroyed as necessary. It never moves between
threads and it will not be shared between requests. As such it is the perfect place to
store database connection information and other things. The internal stack object is
called flask._app_ctx_stack. Extensions are free to store additional information on
the topmost level, assuming they pick a sufficiently unique name and should put their
information there, instead of on the flask.g object which is reserved for user code.
For more information about that, see Flask Extension Development.
Context Usage
The context is typically used to cache resources that need to be created on a per-request
or usage case. For instance, database connections are destined to go there. When
storing things on the application context unique names should be chosen as this is a
place that is shared between Flask applications and extensions.
The most common usage is to split resource management into two parts:
1. an implicit resource caching on the context.
2. a context teardown based resource deallocation.
Generally there would be a get_X() function that creates resource X if it does not exist yet and otherwise returns the same resource, and a teardown_X() function that is
registered as teardown handler.
This is an example that connects to a database:
import sqlite3
from flask import g
def get_db():
db = getattr(g, '_database', None)
if db is None:
db = g._database = connect_to_database()
return db
def teardown_db(exception):
db = getattr(g, '_database', None)
if db is not None:
The first time get_db() is called the connection will be established. To make this implicit a LocalProxy can be used:
from werkzeug.local import LocalProxy
db = LocalProxy(get_db)
That way a user can directly access db which internally calls get_db().
The Request Context
This document describes the behavior in Flask 0.7 which is mostly in line with the old
behavior but has some small, subtle differences.
It is recommended that you read the The Application Context chapter first.
Diving into Context Locals
Say you have a utility function that returns the URL the user should be redirected to.
Imagine it would always redirect to the URL’s next parameter or the HTTP referrer or
the index page:
from flask import request, url_for
def redirect_url():
return request.args.get('next') or \
request.referrer or \
As you can see, it accesses the request object. If you try to run this from a plain Python
shell, this is the exception you will see:
>>> redirect_url()
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
AttributeError: 'NoneType' object has no attribute 'request'
That makes a lot of sense because we currently do not have a request we could
access. So we have to make a request and bind it to the current context.
test_request_context method can create us a RequestContext:
>>> ctx = app.test_request_context('/?next=http://example.com/')
This context can be used in two ways. Either with the with statement or by calling the
push() and pop() methods:
>>> ctx.push()
From that point onwards you can work with the request object:
>>> redirect_url()
Until you call pop:
>>> ctx.pop()
Because the request context is internally maintained as a stack you can push and pop
multiple times. This is very handy to implement things like internal redirects.
For more information of how to utilize the request context from the interactive Python
shell, head over to the Working with the Shell chapter.
How the Context Works
If you look into how the Flask WSGI application internally works, you will find a piece
of code that looks very much like this:
def wsgi_app(self, environ):
with self.request_context(environ):
response = self.full_dispatch_request()
except Exception as e:
response = self.make_response(self.handle_exception(e))
return response(environ, start_response)
The method request_context() returns a new RequestContext object and uses it in
combination with the with statement to bind the context. Everything that is called
from the same thread from this point onwards until the end of the with statement will
have access to the request globals (flask.request and others).
The request context internally works like a stack: The topmost level on the stack is
the current active request. push() adds the context to the stack on the very top, pop()
removes it from the stack again. On popping the application’s teardown_request()
functions are also executed.
Another thing of note is that the request context will automatically also create an application context when it’s pushed and there is no application context for that application
so far.
Callbacks and Errors
What happens if an error occurs in Flask during request processing? This particular
behavior changed in 0.7 because we wanted to make it easier to understand what is
actually happening. The new behavior is quite simple:
1. Before each request, before_request() functions are executed. If one of these
functions return a response, the other functions are no longer called. In any case
however the return value is treated as a replacement for the view’s return value.
2. If the before_request() functions did not return a response, the regular request
handling kicks in and the view function that was matched has the chance to
return a response.
3. The return value of the view is then converted into an actual response object and
handed over to the after_request() functions which have the chance to replace
it or modify it in place.
4. At the end of the request the teardown_request() functions are executed. This
always happens, even in case of an unhandled exception down the road or if a
before-request handler was not executed yet or at all (for example in test environments sometimes you might want to not execute before-request callbacks).
Now what happens on errors? If you are not in debug mode and an exception is not
caught, the 500 internal server handler is called. In debug mode however the exception
is not further processed and bubbles up to the WSGI server. That way things like the
interactive debugger can provide helpful debug information.
An important change in 0.7 is that the internal server error is now no longer post
processed by the after request callbacks and after request callbacks are no longer guaranteed to be executed. This way the internal dispatching code looks cleaner and is
easier to customize and understand.
The new teardown functions are supposed to be used as a replacement for things that
absolutely need to happen at the end of request.
Teardown Callbacks
The teardown callbacks are special callbacks in that they are executed at a different
point. Strictly speaking they are independent of the actual request handling as they
are bound to the lifecycle of the RequestContext object. When the request context is
popped, the teardown_request() functions are called.
This is important to know if the life of the request context is prolonged by using the
test client in a with statement or when using the request context from the command
with app.test_client() as client:
resp = client.get('/foo')
# the teardown functions are still not called at that point
# even though the response ended and you have the response
# object in your hand
# only when the code reaches this point the teardown functions
# are called. Alternatively the same thing happens if another
# request was triggered from the test client
It’s easy to see the behavior from the command line:
>>> app = Flask(__name__)
>>> @app.teardown_request
... def teardown_request(exception=None):
print 'this runs after request'
>>> ctx = app.test_request_context()
>>> ctx.push()
>>> ctx.pop()
this runs after request
Keep in mind that teardown callbacks are always executed, even if before-request callbacks were not executed yet but an exception happened. Certain parts of the test system might also temporarily create a request context without calling the before-request
handlers. Make sure to write your teardown-request handlers in a way that they will
never fail.
Notes On Proxies
Some of the objects provided by Flask are proxies to other objects. The reason behind
this is that these proxies are shared between threads and they have to dispatch to the
actual object bound to a thread behind the scenes as necessary.
Most of the time you don’t have to care about that, but there are some exceptions
where it is good to know that this object is an actual proxy:
• The proxy objects do not fake their inherited types, so if you want to perform
actual instance checks, you have to do that on the instance that is being proxied
(see _get_current_object below).
• if the object reference is important (so for example for sending Signals)
If you need to get access to the underlying object that is proxied, you can use the
_get_current_object() method:
app = current_app._get_current_object()
Context Preservation on Error
If an error occurs or not, at the end of the request the request context is popped and
all data associated with it is destroyed. During development however that can be
problematic as you might want to have the information around for a longer time in
case an exception occurred. In Flask 0.6 and earlier in debug mode, if an exception
occurred, the request context was not popped so that the interactive debugger can still
provide you with important information.
Starting with Flask 0.7 you have finer control over that behavior by setting the
PRESERVE_CONTEXT_ON_EXCEPTION configuration variable. By default it’s linked to the
setting of DEBUG. If the application is in debug mode the context is preserved. If debug
mode is set to off, the context is not preserved.
Do not force activate PRESERVE_CONTEXT_ON_EXCEPTION if debug mode is set to off as it
will cause your application to leak memory on exceptions. However, it can be useful
during development to get the same error preserving behavior as debug mode when
attempting to debug an error that only occurs under production settings.
Modular Applications with Blueprints
New in version 0.7.
Flask uses a concept of blueprints for making application components and supporting
common patterns within an application or across applications. Blueprints can greatly
simplify how large applications work and provide a central means for Flask extensions
to register operations on applications. A Blueprint object works similarly to a Flask
application object, but it is not actually an application. Rather it is a blueprint of how
to construct or extend an application.
Why Blueprints?
Blueprints in Flask are intended for these cases:
• Factor an application into a set of blueprints. This is ideal for larger applications;
a project could instantiate an application object, initialize several extensions, and
register a collection of blueprints.
• Register a blueprint on an application at a URL prefix and/or subdomain. Parameters in the URL prefix/subdomain become common view arguments (with
defaults) across all view functions in the blueprint.
• Register a blueprint multiple times on an application with different URL rules.
• Provide template filters, static files, templates, and other utilities through
blueprints. A blueprint does not have to implement applications or view functions.
• Register a blueprint on an application for any of these cases when initializing a
Flask extension.
A blueprint in Flask is not a pluggable app because it is not actually an application
– it’s a set of operations which can be registered on an application, even multiple
times. Why not have multiple application objects? You can do that (see Application
Dispatching), but your applications will have separate configs and will be managed at
the WSGI layer.
Blueprints instead provide separation at the Flask level, share application config, and
can change an application object as necessary with being registered. The downside is
that you cannot unregister a blueprint once an application was created without having
to destroy the whole application object.
The Concept of Blueprints
The basic concept of blueprints is that they record operations to execute when registered on an application. Flask associates view functions with blueprints when dispatching requests and generating URLs from one endpoint to another.
My First Blueprint
This is what a very basic blueprint looks like. In this case we want to implement a
blueprint that does simple rendering of static templates:
from flask import Blueprint, render_template, abort
from jinja2 import TemplateNotFound
simple_page = Blueprint('simple_page', __name__,
@simple_page.route('/', defaults={'page': 'index'})
def show(page):
return render_template('pages/%s.html' % page)
except TemplateNotFound:
When you bind a function with the help of the @simple_page.route decorator the
blueprint will record the intention of registering the function show on the application
when it’s later registered. Additionally it will prefix the endpoint of the function with
the name of the blueprint which was given to the Blueprint constructor (in this case
also simple_page).
Registering Blueprints
So how do you register that blueprint? Like this:
from flask import Flask
from yourapplication.simple_page import simple_page
app = Flask(__name__)
If you check the rules registered on the application, you will find these:
[<Rule '/static/<filename>' (HEAD, OPTIONS, GET) -> static>,
<Rule '/<page>' (HEAD, OPTIONS, GET) -> simple_page.show>,
<Rule '/' (HEAD, OPTIONS, GET) -> simple_page.show>]
The first one is obviously from the application itself for the static files. The other two
are for the show function of the simple_page blueprint. As you can see, they are also
prefixed with the name of the blueprint and separated by a dot (.).
Blueprints however can also be mounted at different locations:
app.register_blueprint(simple_page, url_prefix='/pages')
And sure enough, these are the generated rules:
[<Rule '/static/<filename>' (HEAD, OPTIONS, GET) -> static>,
<Rule '/pages/<page>' (HEAD, OPTIONS, GET) -> simple_page.show>,
<Rule '/pages/' (HEAD, OPTIONS, GET) -> simple_page.show>]
On top of that you can register blueprints multiple times though not every blueprint
might respond properly to that. In fact it depends on how the blueprint is implemented if it can be mounted more than once.
Blueprint Resources
Blueprints can provide resources as well. Sometimes you might want to introduce a
blueprint only for the resources it provides.
Blueprint Resource Folder
Like for regular applications, blueprints are considered to be contained in a folder.
While multiple blueprints can originate from the same folder, it does not have to be
the case and it’s usually not recommended.
The folder is inferred from the second argument to Blueprint which is usually
__name__. This argument specifies what logical Python module or package corresponds to the blueprint. If it points to an actual Python package that package (which is
a folder on the filesystem) is the resource folder. If it’s a module, the package the module is contained in will be the resource folder. You can access the Blueprint.root_path
property to see what the resource folder is:
>>> simple_page.root_path
To quickly open sources from this folder you can use the open_resource() function:
with simple_page.open_resource('static/style.css') as f:
code = f.read()
Static Files
A blueprint can expose a folder with static files by providing a path to a folder on the
filesystem via the static_folder keyword argument. It can either be an absolute path or
one relative to the folder of the blueprint:
admin = Blueprint('admin', __name__, static_folder='static')
By default the rightmost part of the path is where it is exposed on the web. Because
the folder is called static here it will be available at the location of the blueprint +
/static. Say the blueprint is registered for /admin the static folder will be at /admin/
The endpoint is named blueprint_name.static so you can generate URLs to it like you
would do to the static folder of the application:
url_for('admin.static', filename='style.css')
If you want the blueprint to expose templates you can do that by providing the template_folder parameter to the Blueprint constructor:
admin = Blueprint('admin', __name__, template_folder='templates')
For static files, the path can be absolute or relative to the blueprint resource folder.
The template folder is added to the search path of templates but with a lower priority than the actual application’s template folder. That way you can easily override
templates that a blueprint provides in the actual application. This also means that if
you don’t want a blueprint template to be accidentally overridden, make sure that no
other blueprint or actual application template has the same relative path. When multiple blueprints provide the same relative template path the first blueprint registered
takes precedence over the others.
So if you have a blueprint in the folder yourapplication/admin and you want to render the template 'admin/index.html' and you have provided templates as a template_folder you will have to create a file like this: yourapplication/admin/templates/
admin/index.html. The reason for the extra admin folder is to avoid getting our tem-
plate overridden by a template named index.html in the actual application template
To further reiterate this: if you have a blueprint named admin and you want to render
a template called index.html which is specific to this blueprint, the best idea is to lay
out your templates like this:
And then when you want to render the template, use admin/index.html as the name
to look up the template by. If you encounter problems loading the correct templates
enable the EXPLAIN_TEMPLATE_LOADING config variable which will instruct Flask to print
out the steps it goes through to locate templates on every render_template call.
Building URLs
If you want to link from one page to another you can use the url_for() function just
like you normally would do just that you prefix the URL endpoint with the name of
the blueprint and a dot (.):
Additionally if you are in a view function of a blueprint or a rendered template and
you want to link to another endpoint of the same blueprint, you can use relative redirects by prefixing the endpoint with a dot only:
This will link to admin.index for instance in case the current request was dispatched
to any other admin blueprint endpoint.
Error Handlers
Blueprints support the errorhandler decorator just like the Flask application object, so
it is easy to make Blueprint-specific custom error pages.
Here is an example for a “404 Page Not Found” exception:
def page_not_found(e):
return render_template('pages/404.html')
Most errorhandlers will simply work as expected; however, there is a caveat concerning handlers for 404 and 405 exceptions. These errorhandlers are only invoked from
an appropriate raise statement or a call to abort in another of the blueprint’s view
functions; they are not invoked by, e.g., an invalid URL access. This is because the
blueprint does not “own” a certain URL space, so the application instance has no way
of knowing which blueprint errorhandler it should run if given an invalid URL. If
you would like to execute different handling strategies for these errors based on URL
prefixes, they may be defined at the application level using the request proxy object:
def _handle_api_error(ex):
if request.path.startswith('/api/'):
return jsonify_error(ex)
return ex
More information on error handling see Custom Error Pages.
Flask Extensions
Flask extensions extend the functionality of Flask in various different ways. For instance they add support for databases and other common tasks.
Finding Extensions
Flask extensions are listed on the Flask Extension Registry and can be downloaded
with easy_install or pip. If you add a Flask extension as dependency to your
requirements.txt or setup.py file they are usually installed with a simple command
or when your application installs.
Using Extensions
Extensions typically have documentation that goes along that shows how to use it.
There are no general rules in how extensions are supposed to behave but they are
imported from common locations. If you have an extension called Flask-Foo or
Foo-Flask it should be always importable from flask_foo:
import flask_foo
Building Extensions
While Flask Extension Registry contains many Flask extensions, you may not find an
extension that fits your need. If this is the case, you can always create your own.
Consider reading Flask Extension Development to develop your own Flask extension.
Flask Before 0.8
If you are using Flask 0.7 or earlier the flask.ext package will not exist, instead you
have to import from flaskext.foo or flask_foo depending on how the extension is
distributed. If you want to develop an application that supports Flask 0.7 or earlier you
should still import from the flask.ext package. We provide you with a compatibility
module that provides this package for older versions of Flask. You can download it
from GitHub: flaskext_compat.py
And here is how you can use it:
import flaskext_compat
from flask.ext import foo
Once the flaskext_compat module is activated the flask.ext will exist and you can
start importing from there.
Command Line Interface
New in version 0.11.
One of the nice new features in Flask 0.11 is the built-in integration of the click command line interface. This enables a wide range of new features for the Flask ecosystem
and your own applications.
Basic Usage
After installation of Flask you will now find a flask script installed into your virtualenv. If you don’t want to install Flask or you have a special use-case you can also
use python -m flask to accomplish exactly the same.
The way this script works is by providing access to all the commands on your Flask
application’s Flask.cli instance as well as some built-in commands that are always
there. Flask extensions can also register more commands there if they desire so.
For the flask script to work, an application needs to be discovered. This is achieved
by exporting the FLASK_APP environment variable. It can be either set to an import
path or to a filename of a Python module that contains a Flask application.
In that imported file the name of the app needs to be called app or optionally be specified after a colon. For instance mymodule:application would tell it to use the application
object in the mymodule.py file.
Given a hello.py file with the application in it named app this is how it can be run.
Environment variables (On Windows use set instead of export):
export FLASK_APP=hello
flask run
Or with a filename:
export FLASK_APP=/path/to/hello.py
flask run
Virtualenv Integration
If you are constantly working with a virtualenv you can also put the export FLASK_APP
into your activate script by adding it to the bottom of the file. That way every time
you activate your virtualenv you automatically also activate the correct application
Edit the activate script for the shell you use. For example:
Unix Bash: venv/bin/activate:
export FLASK_APP
Windows CMD.exe: venv\Scripts\activate.bat:
set "FLASK_APP=hello"
Debug Flag
The flask script can also be instructed to enable the debug mode of the application
automatically by exporting FLASK_DEBUG. If set to 1 debug is enabled or 0 disables it:
export FLASK_DEBUG=1
Running a Shell
To run an interactive Python shell you can use the shell command:
flask shell
This will start up an interactive Python shell, setup the correct application context
and setup the local variables in the shell. This is done by invoking the Flask.
make_shell_context() method of the application. By default you have access to your
app and g.
Custom Commands
If you want to add more commands to the shell script you can do this easily. Flask
uses click for the command interface which makes creating custom commands very
easy. For instance if you want a shell command to initialize the database you can do
import click
from flask import Flask
app = Flask(__name__)
def initdb():
"""Initialize the database."""
click.echo('Init the db')
The command will then show up on the command line:
$ flask initdb
Init the db
Application Context
Most commands operate on the application so it makes a lot of sense if they have
the application context setup. Because of this, if you register a callback on app.
cli with the command() the callback will automatically be wrapped through cli.
with_appcontext() which informs the cli system to ensure that an application context is set up. This behavior is not available if a command is added later with
add_command() or through other means.
It can also be disabled by passing with_appcontext=False to the decorator:
def example():
Factory Functions
In case you are using factory functions to create your application (see Application Factories) you will discover that the flask command cannot work with them directly. Flask
won’t be able to figure out how to instantiate your application properly by itself. Because of this reason the recommendation is to create a separate file that instantiates
applications. This is not the only way to make this work. Another is the Custom Scripts
For instance if you have a factory function that creates an application from a filename
you could make a separate file that creates such an application from an environment
This could be a file named autoapp.py with these contents:
import os
from yourapplication import create_app
app = create_app(os.environ['YOURAPPLICATION_CONFIG'])
Once this has happened you can make the flask command automatically pick it up:
export YOURAPPLICATION_CONFIG=/path/to/config.cfg
export FLASK_APP=/path/to/autoapp.py
From this point onwards flask will find your application.
Custom Scripts
While the most common way is to use the flask command, you can also make your
own “driver scripts”. Since Flask uses click for the scripts there is no reason you cannot
hook these scripts into any click application. There is one big caveat and that is, that
commands registered to Flask.cli will expect to be (indirectly at least) launched from
a flask.cli.FlaskGroup click group. This is necessary so that the commands know
which Flask application they have to work with.
To understand why you might want custom scripts you need to understand how click
finds and executes the Flask application. If you use the flask script you specify the
application to work with on the command line or environment variable as an import
name. This is simple but it has some limitations. Primarily it does not work with
application factory functions (see Application Factories).
With a custom script you don’t have this problem as you can fully customize how the
application will be created. This is very useful if you write reusable applications that
you want to ship to users and they should be presented with a custom management
To explain all of this, here is an example manage.py script that manages a hypothetical
wiki application. We will go through the details afterwards:
import os
import click
from flask.cli import FlaskGroup
def create_wiki_app(info):
from yourwiki import create_app
return create_app(
config=os.environ.get('WIKI_CONFIG', 'wikiconfig.py'))
@click.group(cls=FlaskGroup, create_app=create_wiki_app)
def cli():
"""This is a management script for the wiki application."""
if __name__ == '__main__':
That’s a lot of code for not much, so let’s go through all parts step by step.
1. First we import the click library as well as the click extensions from the flask.
cli package. Primarily we are here interested in the FlaskGroup click group.
2. The next thing we do is defining a function that is invoked with the script info
object (ScriptInfo) from Flask and its purpose is to fully import and create the
application. This can either directly import an application object or create it (see
Application Factories). In this case we load the config from an environment variable.
3. Next step is to create a FlaskGroup. In this case we just make an empty function
with a help doc string that just does nothing and then pass the create_wiki_app
function as a factory function.
Whenever click now needs to operate on a Flask application it will call that function with the script info and ask for it to be created.
4. All is rounded up by invoking the script.
CLI Plugins
Flask extensions can always patch the Flask.cli instance with more commands if they
want. However there is a second way to add CLI plugins to Flask which is through
setuptools. If you make a Python package that should export a Flask command line
plugin you can ship a setup.py file that declares an entrypoint that points to a click
Example setup.py:
from setuptools import setup
Inside mypackage/commands.py you can then export a Click object:
import click
def cli():
"""This is an example command."""
Once that package is installed in the same virtualenv as Flask itself you can run flask
my-command to invoke your command. This is useful to provide extra functionality that
Flask itself cannot ship.
Development Server
Starting with Flask 0.11 there are multiple built-in ways to run a development server.
The best one is the flask command line utility but you can also continue using the
Flask.run() method.
Command Line
The flask command line script (Command Line Interface) is strongly recommended for
development because it provides a superior reload experience due to how it loads the
application. The basic usage is like this:
$ export FLASK_APP=my_application
$ export FLASK_DEBUG=1
$ flask run
This will enable the debugger, the reloader and then start the server on
The individual features of the server can be controlled by passing more arguments to
the run option. For instance the reloader can be disabled:
$ flask run --no-reload
In Code
The alternative way to start the application is through the Flask.run() method. This
will immediately launch a local server exactly the same way the flask script does.
if __name__ == '__main__':
This works well for the common case but it does not work well for development which
is why from Flask 0.11 onwards the flask method is recommended. The reason for this
is that due to how the reload mechanism works there are some bizarre side-effects (like
executing certain code twice, sometimes crashing without message or dying when a
syntax or import error happens).
It is however still a perfectly valid method for invoking a non automatic reloading
Working with the Shell
New in version 0.3.
One of the reasons everybody loves Python is the interactive shell. It basically allows
you to execute Python commands in real time and immediately get results back. Flask
itself does not come with an interactive shell, because it does not require any specific
setup upfront, just import your application and start playing around.
There are however some handy helpers to make playing around in the shell a more
pleasant experience. The main issue with interactive console sessions is that you’re
not triggering a request like a browser does which means that g, request and others
are not available. But the code you want to test might depend on them, so what can
you do?
This is where some helper functions come in handy. Keep in mind however that these
functions are not only there for interactive shell usage, but also for unittesting and
other situations that require a faked request context.
Generally it’s recommended that you read the The Request Context chapter of the documentation first.
Command Line Interface
Starting with Flask 0.11 the recommended way to work with the shell is the flask
shell command which does a lot of this automatically for you. For instance the shell
is automatically initialized with a loaded application context.
For more information see Command Line Interface.
Creating a Request Context
The easiest way to create a proper request context from the shell is by using the
test_request_context method which creates us a RequestContext:
>>> ctx = app.test_request_context()
Normally you would use the with statement to make this request object active, but in
the shell it’s easier to use the push() and pop() methods by hand:
>>> ctx.push()
From that point onwards you can work with the request object until you call pop:
>>> ctx.pop()
Firing Before/After Request
By just creating a request context, you still don’t have run the code that is normally
run before a request. This might result in your database being unavailable if you are
connecting to the database in a before-request callback or the current user not being
stored on the g object etc.
This however can easily be done yourself. Just call preprocess_request():
>>> ctx = app.test_request_context()
>>> ctx.push()
>>> app.preprocess_request()
Keep in mind that the preprocess_request() function might return a response object,
in that case just ignore it.
To shutdown a request, you need to trick a bit before the after request functions (triggered by process_response()) operate on a response object:
>>> app.process_response(app.response_class())
<Response 0 bytes [200 OK]>
>>> ctx.pop()
The functions registered as teardown_request() are automatically called when the
context is popped. So this is the perfect place to automatically tear down resources
that were needed by the request context (such as database connections).
Further Improving the Shell Experience
If you like the idea of experimenting in a shell, create yourself a module with stuff you
want to star import into your interactive session. There you could also define some
more helper methods for common things such as initializing the database, dropping
tables etc.
Just put them into a module (like shelltools) and import from there:
>>> from shelltools import *
Patterns for Flask
Certain things are common enough that the chances are high you will find them in
most web applications. For example quite a lot of applications are using relational
databases and user authentication. In that case, chances are they will open a database
connection at the beginning of the request and get the information of the currently
logged in user. At the end of the request, the database connection is closed again.
There are more user contributed snippets and patterns in the Flask Snippet Archives.
Larger Applications
For larger applications it’s a good idea to use a package instead of a module. That is
quite simple. Imagine a small application looks like this:
If you find yourself stuck on something, feel free to take a look at the source code for
this example. You’ll find the full src for this example here.
Simple Packages
To convert that into a larger one, just create a new folder yourapplication inside
the existing one and move everything below it. Then rename yourapplication.py
to __init__.py. (Make sure to delete all .pyc files first, otherwise things would most
likely break)
You should then end up with something like that:
But how do you run your application now? The naive python yourapplication/
__init__.py will not work. Let’s just say that Python does not want modules in packages to be the startup file. But that is not a big problem, just add a new file called
setup.py next to the inner yourapplication folder with the following contents:
from setuptools import setup
In order to run the application you need to export an environment variable that tells
Flask where to find the application instance:
export FLASK_APP=yourapplication
If you are outside of the project directory make sure to provide the exact path to your
application directory. Similarly you can turn on “debug mode” with this environment
export FLASK_DEBUG=true
In order to install and run the application you need to issue the following commands:
pip install -e .
flask run
What did we gain from this? Now we can restructure the application a bit into multiple
modules. The only thing you have to remember is the following quick checklist:
1. the Flask application object creation has to be in the __init__.py file. That way
each module can import it safely and the __name__ variable will resolve to the
correct package.
2. all the view functions (the ones with a route() decorator on top) have to be
imported in the __init__.py file. Not the object itself, but the module it is in.
Import the view module after the application object is created.
Here’s an example __init__.py:
from flask import Flask
app = Flask(__name__)
import yourapplication.views
And this is what views.py would look like:
from yourapplication import app
def index():
return 'Hello World!'
You should then end up with something like that:
Circular Imports
Every Python programmer hates them, and yet we just added some: circular imports
(That’s when two modules depend on each other. In this case views.py depends on
__init__.py). Be advised that this is a bad idea in general but here it is actually fine.
The reason for this is that we are not actually using the views in __init__.py and just
ensuring the module is imported and we are doing that at the bottom of the file.
There are still some problems with that approach but if you want to use decorators
there is no way around that. Check out the Becoming Big section for some inspiration
how to deal with that.
Working with Blueprints
If you have larger applications it’s recommended to divide them into smaller groups
where each group is implemented with the help of a blueprint. For a gentle introduction into this topic refer to the Modular Applications with Blueprints chapter of the
Application Factories
If you are already using packages and blueprints for your application (Modular Applications with Blueprints) there are a couple of really nice ways to further improve the
experience. A common pattern is creating the application object when the blueprint
is imported. But if you move the creation of this object into a function, you can then
create multiple instances of this app later.
So why would you want to do this?
1. Testing. You can have instances of the application with different settings to test
every case.
2. Multiple instances. Imagine you want to run different versions of the same application. Of course you could have multiple instances with different configs set
up in your webserver, but if you use factories, you can have multiple instances
of the same application running in the same application process which can be
So how would you then actually implement that?
Basic Factories
The idea is to set up the application in a function. Like this:
def create_app(config_filename):
app = Flask(__name__)
from yourapplication.model import db
from yourapplication.views.admin import admin
from yourapplication.views.frontend import frontend
return app
The downside is that you cannot use the application object in the blueprints at import
time. You can however use it from within a request. How do you get access to the
application with the config? Use current_app:
from flask import current_app, Blueprint, render_template
admin = Blueprint('admin', __name__, url_prefix='/admin')
def index():
return render_template(current_app.config['INDEX_TEMPLATE'])
Here we look up the name of a template in the config.
Factories & Extensions
It’s preferable to create your extensions and app factories so that the extension object
does not initially get bound to the application.
Using Flask-SQLAlchemy, as an example, you should not do something along those
def create_app(config_filename):
app = Flask(__name__)
db = SQLAlchemy(app)
But, rather, in model.py (or equivalent):
db = SQLAlchemy()
and in your application.py (or equivalent):
def create_app(config_filename):
app = Flask(__name__)
from yourapplication.model import db
Using this design pattern, no application-specific state is stored on the extension object, so one extension object can be used for multiple apps. For more information about
the design of extensions refer to Flask Extension Development.
Using Applications
So to use such an application you then have to create the application first in a separate file otherwise the flask command won’t be able to find it. Here an example
exampleapp.py file that creates such an application:
from yourapplication import create_app
app = create_app('/path/to/config.cfg')
It can then be used with the flask command:
export FLASK_APP=exampleapp
flask run
Factory Improvements
The factory function from above is not very clever so far, you can improve it. The
following changes are straightforward and possible:
1. make it possible to pass in configuration values for unittests so that you don’t
have to create config files on the filesystem
2. call a function from a blueprint when the application is setting up so that you
have a place to modify attributes of the application (like hooking in before /
after request handlers etc.)
3. Add in WSGI middlewares when the application is creating if necessary.
Application Dispatching
Application dispatching is the process of combining multiple Flask applications on the
WSGI level. You can combine not only Flask applications but any WSGI application.
This would allow you to run a Django and a Flask application in the same interpreter
side by side if you want. The usefulness of this depends on how the applications work
The fundamental difference from the module approach is that in this case you are running the same or different Flask applications that are entirely isolated from each other.
They run different configurations and are dispatched on the WSGI level.
Working with this Document
Each of the techniques and examples below results in an application object that can be
run with any WSGI server. For production, see Deployment Options. For development,
Werkzeug provides a builtin server for development available at werkzeug.serving.
from werkzeug.serving import run_simple
run_simple('localhost', 5000, application, use_reloader=True)
Note that run_simple is not intended for use in production. Use a full-blown WSGI
In order to use the interactive debugger, debugging must be enabled both on the application and the simple server. Here is the “hello world” example with debugging
and run_simple:
from flask import Flask
from werkzeug.serving import run_simple
app = Flask(__name__)
app.debug = True
def hello_world():
return 'Hello World!'
if __name__ == '__main__':
run_simple('localhost', 5000, app,
use_reloader=True, use_debugger=True, use_evalex=True)
Combining Applications
If you have entirely separated applications and you want them to work next to each
other in the same Python interpreter process you can take advantage of the werkzeug.
wsgi.DispatcherMiddleware. The idea here is that each Flask application is a valid
WSGI application and they are combined by the dispatcher middleware into a larger
one that is dispatched based on prefix.
For example you could have your main application run on / and your backend interface on /backend:
from werkzeug.wsgi import DispatcherMiddleware
from frontend_app import application as frontend
from backend_app import application as backend
application = DispatcherMiddleware(frontend, {
Dispatch by Subdomain
Sometimes you might want to use multiple instances of the same application with
different configurations. Assuming the application is created inside a function and
you can call that function to instantiate it, that is really easy to implement. In order to
develop your application to support creating new instances in functions have a look
at the Application Factories pattern.
A very common example would be creating applications per subdomain. For instance
you configure your webserver to dispatch all requests for all subdomains to your
application and you then use the subdomain information to create user-specific instances. Once you have your server set up to listen on all subdomains you can use a
very simple WSGI application to do the dynamic application creation.
The perfect level for abstraction in that regard is the WSGI layer. You write your own
WSGI application that looks at the request that comes and delegates it to your Flask
application. If that application does not exist yet, it is dynamically created and remembered:
from threading import Lock
class SubdomainDispatcher(object):
def __init__(self, domain, create_app):
self.domain = domain
self.create_app = create_app
self.lock = Lock()
self.instances = {}
def get_application(self, host):
host = host.split(':')[0]
assert host.endswith(self.domain), 'Configuration error'
subdomain = host[:-len(self.domain)].rstrip('.')
with self.lock:
app = self.instances.get(subdomain)
if app is None:
app = self.create_app(subdomain)
self.instances[subdomain] = app
return app
def __call__(self, environ, start_response):
app = self.get_application(environ['HTTP_HOST'])
return app(environ, start_response)
This dispatcher can then be used like this:
from myapplication import create_app, get_user_for_subdomain
from werkzeug.exceptions import NotFound
def make_app(subdomain):
user = get_user_for_subdomain(subdomain)
if user is None:
# if there is no user for that subdomain we still have
# to return a WSGI application that handles that request.
# We can then just return the NotFound() exception as
# application which will render a default 404 page.
# You might also redirect the user to the main page then
return NotFound()
# otherwise create the application for the specific user
return create_app(user)
application = SubdomainDispatcher('example.com', make_app)
Dispatch by Path
Dispatching by a path on the URL is very similar. Instead of looking at the Host header
to figure out the subdomain one simply looks at the request path up to the first slash:
from threading import Lock
from werkzeug.wsgi import pop_path_info, peek_path_info
class PathDispatcher(object):
def __init__(self, default_app, create_app):
self.default_app = default_app
self.create_app = create_app
self.lock = Lock()
self.instances = {}
def get_application(self, prefix):
with self.lock:
app = self.instances.get(prefix)
if app is None:
app = self.create_app(prefix)
if app is not None:
self.instances[prefix] = app
return app
def __call__(self, environ, start_response):
app = self.get_application(peek_path_info(environ))
if app is not None:
app = self.default_app
return app(environ, start_response)
The big difference between this and the subdomain one is that this one falls back to
another application if the creator function returns None:
from myapplication import create_app, default_app, get_user_for_prefix
def make_app(prefix):
user = get_user_for_prefix(prefix)
if user is not None:
return create_app(user)
application = PathDispatcher(default_app, make_app)
Implementing API Exceptions
It’s very common to implement RESTful APIs on top of Flask. One of the first things
that developers run into is the realization that the builtin exceptions are not expressive
enough for APIs and that the content type of text/html they are emitting is not very
useful for API consumers.
The better solution than using abort to signal errors for invalid API usage is to implement your own exception type and install an error handler for it that produces the
errors in the format the user is expecting.
Simple Exception Class
The basic idea is to introduce a new exception that can take a proper human readable
message, a status code for the error and some optional payload to give more context
for the error.
This is a simple example:
from flask import jsonify
class InvalidUsage(Exception):
status_code = 400
def __init__(self, message, status_code=None, payload=None):
self.message = message
if status_code is not None:
self.status_code = status_code
self.payload = payload
def to_dict(self):
rv = dict(self.payload or ())
rv['message'] = self.message
return rv
A view can now raise that exception with an error message. Additionally some extra
payload can be provided as a dictionary through the payload parameter.
Registering an Error Handler
At that point views can raise that error, but it would immediately result in an internal
server error. The reason for this is that there is no handler registered for this error
class. That however is easy to add:
def handle_invalid_usage(error):
response = jsonify(error.to_dict())
response.status_code = error.status_code
return response
Usage in Views
Here is how a view can use that functionality:
def get_foo():
raise InvalidUsage('This view is gone', status_code=410)
Using URL Processors
New in version 0.7.
Flask 0.7 introduces the concept of URL processors. The idea is that you might have
a bunch of resources with common parts in the URL that you don’t always explicitly
want to provide. For instance you might have a bunch of URLs that have the language
code in it but you don’t want to have to handle it in every single function yourself.
URL processors are especially helpful when combined with blueprints. We will handle
both application specific URL processors here as well as blueprint specifics.
Internationalized Application URLs
Consider an application like this:
from flask import Flask, g
app = Flask(__name__)
def index(lang_code):
g.lang_code = lang_code
def about(lang_code):
g.lang_code = lang_code
This is an awful lot of repetition as you have to handle the language code setting on the
g object yourself in every single function. Sure, a decorator could be used to simplify
this, but if you want to generate URLs from one function to another you would have
to still provide the language code explicitly which can be annoying.
For the latter, this is where url_defaults() functions come in. They can automatically
inject values into a call for url_for() automatically. The code below checks if the
language code is not yet in the dictionary of URL values and if the endpoint wants a
value named 'lang_code':
def add_language_code(endpoint, values):
if 'lang_code' in values or not g.lang_code:
if app.url_map.is_endpoint_expecting(endpoint, 'lang_code'):
values['lang_code'] = g.lang_code
The method is_endpoint_expecting() of the URL map can be used to figure out if it
would make sense to provide a language code for the given endpoint.
The reverse of that function are url_value_preprocessor()s. They are executed right
after the request was matched and can execute code based on the URL values. The
idea is that they pull information out of the values dictionary and put it somewhere
def pull_lang_code(endpoint, values):
g.lang_code = values.pop('lang_code', None)
That way you no longer have to do the lang_code assignment to g in every function.
You can further improve that by writing your own decorator that prefixes URLs with
the language code, but the more beautiful solution is using a blueprint. Once the
'lang_code' is popped from the values dictionary and it will no longer be forwarded
to the view function reducing the code to this:
from flask import Flask, g
app = Flask(__name__)
def add_language_code(endpoint, values):
if 'lang_code' in values or not g.lang_code:
if app.url_map.is_endpoint_expecting(endpoint, 'lang_code'):
values['lang_code'] = g.lang_code
def pull_lang_code(endpoint, values):
g.lang_code = values.pop('lang_code', None)
def index():
def about():
Internationalized Blueprint URLs
Because blueprints can automatically prefix all URLs with a common string it’s easy
to automatically do that for every function. Furthermore blueprints can have perblueprint URL processors which removes a whole lot of logic from the url_defaults()
function because it no longer has to check if the URL is really interested in a
'lang_code' parameter:
from flask import Blueprint, g
bp = Blueprint('frontend', __name__, url_prefix='/<lang_code>')
def add_language_code(endpoint, values):
values.setdefault('lang_code', g.lang_code)
def pull_lang_code(endpoint, values):
g.lang_code = values.pop('lang_code')
def index():
def about():
Deploying with Setuptools
Setuptools, is an extension library that is commonly used to distribute Python libraries
and extensions. It extends distutils, a basic module installation system shipped with
Python to also support various more complex constructs that make larger applications
easier to distribute:
• support for dependencies: a library or application can declare a list of other
libraries it depends on which will be installed automatically for you.
• package registry: setuptools registers your package with your Python installation. This makes it possible to query information provided by one package from
another package. The best known feature of this system is the entry point support which allows one package to declare an “entry point” that another package
can hook into to extend the other package.
• installation manager: pip can install other libraries for you.
If you have Python 2 (>=2.7.9) or Python 3 (>=3.4) installed from python.org, you will
already have pip and setuptools on your system. Otherwise, you will need to install
them yourself.
Flask itself, and all the libraries you can find on PyPI are distributed with either setuptools or distutils.
In this case we assume your application is called yourapplication.py and you are not
using a module, but a package. If you have not yet converted your application into a
package, head over to the Larger Applications pattern to see how this can be done.
A working deployment with setuptools is the first step into more complex and more
automated deployment scenarios. If you want to fully automate the process, also read
the Deploying with Fabric chapter.
Basic Setup Script
Because you have Flask installed, you have setuptools available on your system. Flask
already depends upon setuptools.
Standard disclaimer applies: you better use a virtualenv.
Your setup code always goes into a file named setup.py next to your application. The
name of the file is only convention, but because everybody will look for a file with that
name, you better not change it.
A basic setup.py file for a Flask application looks like this:
from setuptools import setup
name='Your Application',
Please keep in mind that you have to list subpackages explicitly. If you want setuptools
to lookup the packages for you automatically, you can use the find_packages function:
from setuptools import setup, find_packages
Most parameters to the setup function should be self explanatory,
include_package_data and zip_safe might not be. include_package_data tells
setuptools to look for a MANIFEST.in file and install all the entries that match as
package data. We will use this to distribute the static files and templates along with
the Python module (see Distributing Resources). The zip_safe flag can be used to force
or prevent zip Archive creation. In general you probably don’t want your packages
to be installed as zip files because some tools do not support them and they make
debugging a lot harder.
Tagging Builds
It is useful to distinguish between release and development builds. Add a setup.cfg
file to configure these options.
[egg_info] tag_build = .dev tag_date = 1
[aliases] release = egg_info -RDb ‘’
Running python setup.py sdist will create a development package with ”.dev” and
the current date appended: flaskr-1.0.dev20160314.tar.gz. Running python setup.
py release sdist will create a release package with only the version: flaskr-1.0.tar.
Distributing Resources
If you try to install the package you just created, you will notice that folders like static
or templates are not installed for you. The reason for this is that setuptools does not
know which files to add for you. What you should do, is to create a MANIFEST.in file
next to your setup.py file. This file lists all the files that should be added to your
recursive-include yourapplication/templates *
recursive-include yourapplication/static *
Don’t forget that even if you enlist them in your MANIFEST.in file, they won’t be installed for you unless you set the include_package_data parameter of the setup function
to True!
Declaring Dependencies
Dependencies are declared in the install_requires parameter as a list. Each item in
that list is the name of a package that should be pulled from PyPI on installation. By
default it will always use the most recent version, but you can also provide minimum
and maximum version requirements. Here some examples:
As mentioned earlier, dependencies are pulled from PyPI. What if you want to depend
on a package that cannot be found on PyPI and won’t be because it is an internal
package you don’t want to share with anyone? Just do it as if there was a PyPI entry
and provide a list of alternative locations where setuptools should look for tarballs:
Make sure that page has a directory listing and the links on the page are pointing to
the actual tarballs with their correct filenames as this is how setuptools will find the
files. If you have an internal company server that contains the packages, provide the
URL to that server.
Installing / Developing
To install your application (ideally into a virtualenv) just run the setup.py script
with the install parameter. It will install your application into the virtualenv’s sitepackages folder and also download and install all dependencies:
$ python setup.py install
If you are developing on the package and also want the requirements to be installed,
you can use the develop command instead:
$ python setup.py develop
This has the advantage of just installing a link to the site-packages folder instead of
copying the data over. You can then continue to work on the code without having to
run install again after each change.
Deploying with Fabric
Fabric is a tool for Python similar to Makefiles but with the ability to execute commands on a remote server. In combination with a properly set up Python package
(Larger Applications) and a good concept for configurations (Configuration Handling) it
is very easy to deploy Flask applications to external servers.
Before we get started, here a quick checklist of things we have to ensure upfront:
• Fabric 1.0 has to be installed locally. This tutorial assumes the latest version of
• The application already has to be a package and requires a working setup.py file
(Deploying with Setuptools).
• In the following example we are using mod_wsgi for the remote servers. You
can of course use your own favourite server there, but for this example we chose
Apache + mod_wsgi because it’s very easy to setup and has a simple way to reload
applications without root access.
Creating the first Fabfile
A fabfile is what controls what Fabric executes. It is named fabfile.py and executed
by the fab command. All the functions defined in that file will show up as fab subcommands. They are executed on one or more hosts. These hosts can be defined either in
the fabfile or on the command line. In this case we will add them to the fabfile.
This is a basic first example that has the ability to upload the current source code to
the server and install it into a pre-existing virtual environment:
from fabric.api import *
# the user to use for the remote commands
env.user = 'appuser'
# the servers where the commands are executed
env.hosts = ['server1.example.com', 'server2.example.com']
def pack():
# build the package
local('python setup.py sdist --formats=gztar', capture=False)
def deploy():
# figure out the package name and version
dist = local('python setup.py --fullname', capture=True).strip()
filename = '%s.tar.gz' % dist
# upload the package to the temporary folder on the server
put('dist/%s' % filename, '/tmp/%s' % filename)
# install the package in the application's virtualenv with pip
run('/var/www/yourapplication/env/bin/pip install /tmp/%s' % filename)
# remove the uploaded package
run('rm -r /tmp/%s' % filename)
# touch the .wsgi file to trigger a reload in mod_wsgi
run('touch /var/www/yourapplication.wsgi')
Running Fabfiles
Now how do you execute that fabfile? You use the fab command. To deploy the current
version of the code on the remote server you would use this command:
$ fab pack deploy
However this requires that our server already has the /var/www/yourapplication
folder created and /var/www/yourapplication/env to be a virtual environment. Furthermore are we not creating the configuration or .wsgi file on the server. So how do
we bootstrap a new server into our infrastructure?
This now depends on the number of servers we want to set up. If we just have one
application server (which the majority of applications will have), creating a command
in the fabfile for this is overkill. But obviously you can do that. In that case you
would probably call it setup or bootstrap and then pass the servername explicitly on the
command line:
$ fab -H newserver.example.com bootstrap
To setup a new server you would roughly do these steps:
1. Create the directory structure in /var/www:
$ mkdir /var/www/yourapplication
$ cd /var/www/yourapplication
$ virtualenv --distribute env
2. Upload a new application.wsgi file to the server and the configuration file for
the application (eg: application.cfg)
3. Create a new Apache config for yourapplication and activate it. Make sure to activate watching for changes of the .wsgi file so that we can automatically reload
the application by touching it. (See mod_wsgi (Apache) for more information)
So now the question is, where do the application.wsgi and application.cfg files
come from?
The WSGI File
The WSGI file has to import the application and also to set an environment variable so
that the application knows where to look for the config. This is a short example that
does exactly that:
import os
os.environ['YOURAPPLICATION_CONFIG'] = '/var/www/yourapplication/application.cfg'
from yourapplication import app
The application itself then has to initialize itself like this to look for the config at that
environment variable:
app = Flask(__name__)
This approach is explained in detail in the Configuration Handling section of the documentation.
The Configuration File
Now as mentioned above, the application will find the correct configuration file by
looking up the YOURAPPLICATION_CONFIG environment variable. So we have to put the
configuration in a place where the application will able to find it. Configuration files
have the unfriendly quality of being different on all computers, so you do not version
them usually.
A popular approach is to store configuration files for different servers in a separate version control repository and check them out on all servers. Then symlink the
file that is active for the server into the location where it’s expected (eg: /var/www/
Either way, in our case here we only expect one or two servers and we can upload
them ahead of time by hand.
First Deployment
Now we can do our first deployment. We have set up the servers so that they have
their virtual environments and activated apache configs. Now we can pack up the
application and deploy it:
$ fab pack deploy
Fabric will now connect to all servers and run the commands as written down in the
fabfile. First it will execute pack so that we have our tarball ready and then it will
execute deploy and upload the source code to all servers and install it there. Thanks
to the setup.py file we will automatically pull in the required libraries into our virtual
Next Steps
From that point onwards there is so much that can be done to make deployment actually fun:
• Create a bootstrap command that initializes new servers. It could initialize a new
virtual environment, setup apache appropriately etc.
• Put configuration files into a separate version control repository and symlink the
active configs into place.
• You could also put your application code into a repository and check out the
latest version on the server and then install. That way you can also easily go
back to older versions.
• hook in testing functionality so that you can deploy to an external server and run
the test suite.
Working with Fabric is fun and you will notice that it’s quite magical to type fab
deploy and see your application being deployed automatically to one or more remote
Using SQLite 3 with Flask
In Flask you can easily implement the opening of database connections on demand
and closing them when the context dies (usually at the end of the request).
Here is a simple example of how you can use SQLite 3 with Flask:
import sqlite3
from flask import g
DATABASE = '/path/to/database.db'
def get_db():
db = getattr(g, '_database', None)
if db is None:
db = g._database = sqlite3.connect(DATABASE)
return db
def close_connection(exception):
db = getattr(g, '_database', None)
if db is not None:
Now, to use the database, the application must either have an active application context (which is always true if there is a request in flight) or create an application context
itself. At that point the get_db function can be used to get the current database connection. Whenever the context is destroyed the database connection will be terminated.
Note: if you use Flask 0.9 or older you need to use flask._app_ctx_stack.top instead
of g as the flask.g object was bound to the request and not application context.
def index():
cur = get_db().cursor()
Note: Please keep in mind that the teardown request and appcontext functions are always executed, even if a before-request handler failed or was never executed. Because
of this we have to make sure here that the database is there before we close it.
Connect on Demand
The upside of this approach (connecting on first use) is that this will only open the
connection if truly necessary. If you want to use this code outside a request context
you can use it in a Python shell by opening the application context by hand:
with app.app_context():
# now you can use get_db()
Easy Querying
Now in each request handling function you can access g.db to get the current open
database connection. To simplify working with SQLite, a row factory function is useful. It is executed for every result returned from the database to convert the result. For
instance, in order to get dictionaries instead of tuples, this could be inserted into the
get_db function we created above:
def make_dicts(cursor, row):
return dict((cursor.description[idx][0], value)
for idx, value in enumerate(row))
db.row_factory = make_dicts
This will make the sqlite3 module return dicts for this database connection, which are
much nicer to deal with. Even more simply, we could place this in get_db instead:
db.row_factory = sqlite3.Row
This would use Row objects rather than dicts to return the results of queries. These
are namedtuple s, so we can access them either by index or by key. For example, assuming we have a sqlite3.Row called r for the rows id, FirstName, LastName, and
>>> # You can get values based on the row's name
>>> r['FirstName']
>>> # Or, you can get them based on index
>>> r[1]
# Row objects are also iterable:
>>> for value in r:
Additionally, it is a good idea to provide a query function that combines getting the
cursor, executing and fetching the results:
def query_db(query, args=(), one=False):
cur = get_db().execute(query, args)
rv = cur.fetchall()
return (rv[0] if rv else None) if one else rv
This handy little function, in combination with a row factory, makes working with the
database much more pleasant than it is by just using the raw cursor and connection
Here is how you can use it:
for user in query_db('select * from users'):
print user['username'], 'has the id', user['user_id']
Or if you just want a single result:
user = query_db('select * from users where username = ?',
[the_username], one=True)
if user is None:
print 'No such user'
print the_username, 'has the id', user['user_id']
To pass variable parts to the SQL statement, use a question mark in the statement and
pass in the arguments as a list. Never directly add them to the SQL statement with
string formatting because this makes it possible to attack the application using SQL
Initial Schemas
Relational databases need schemas, so applications often ship a schema.sql file that
creates the database. It’s a good idea to provide a function that creates the database
based on that schema. This function can do that for you:
def init_db():
with app.app_context():
db = get_db()
with app.open_resource('schema.sql', mode='r') as f:
You can then create such a database from the Python shell:
>>> from yourapplication import init_db
>>> init_db()
SQLAlchemy in Flask
Many people prefer SQLAlchemy for database access. In this case it’s encouraged to
use a package instead of a module for your flask application and drop the models into
a separate module (Larger Applications). While that is not necessary, it makes a lot of
There are four very common ways to use SQLAlchemy. I will outline each of them
Flask-SQLAlchemy Extension
Because SQLAlchemy is a common database abstraction layer and object relational
mapper that requires a little bit of configuration effort, there is a Flask extension that
handles that for you. This is recommended if you want to get started quickly.
You can download Flask-SQLAlchemy from PyPI.
The declarative extension in SQLAlchemy is the most recent method of using
SQLAlchemy. It allows you to define tables and models in one go, similar to how
Django works. In addition to the following text I recommend the official documentation on the declarative extension.
Here’s the example database.py module for your application:
from sqlalchemy import create_engine
from sqlalchemy.orm import scoped_session, sessionmaker
from sqlalchemy.ext.declarative import declarative_base
engine = create_engine('sqlite:////tmp/test.db', convert_unicode=True)
db_session = scoped_session(sessionmaker(autocommit=False,
Base = declarative_base()
Base.query = db_session.query_property()
def init_db():
# import all modules here that might define models so that
# they will be registered properly on the metadata. Otherwise
# you will have to import them first before calling init_db()
import yourapplication.models
To define your models, just subclass the Base class that was created by the code above.
If you are wondering why we don’t have to care about threads here (like we did in the
SQLite3 example above with the g object): that’s because SQLAlchemy does that for
us already with the scoped_session.
To use SQLAlchemy in a declarative way with your application, you just have to put
the following code into your application module. Flask will automatically remove
database sessions at the end of the request or when the application shuts down:
from yourapplication.database import db_session
def shutdown_session(exception=None):
Here is an example model (put this into models.py, e.g.):
from sqlalchemy import Column, Integer, String
from yourapplication.database import Base
class User(Base):
__tablename__ = 'users'
id = Column(Integer, primary_key=True)
name = Column(String(50), unique=True)
email = Column(String(120), unique=True)
def __init__(self, name=None, email=None):
self.name = name
self.email = email
def __repr__(self):
return '<User %r>' % (self.name)
To create the database you can use the init_db function:
>>> from yourapplication.database import init_db
>>> init_db()
You can insert entries into the database like this:
from yourapplication.database import db_session
from yourapplication.models import User
u = User('admin', '[email protected]')
Querying is simple as well:
>>> User.query.all()
[<User u'admin'>]
>>> User.query.filter(User.name == 'admin').first()
<User u'admin'>
Manual Object Relational Mapping
Manual object relational mapping has a few upsides and a few downsides versus the
declarative approach from above. The main difference is that you define tables and
classes separately and map them together. It’s more flexible but a little more to type.
In general it works like the declarative approach, so make sure to also split up your
application into multiple modules in a package.
Here is an example database.py module for your application:
from sqlalchemy import create_engine, MetaData
from sqlalchemy.orm import scoped_session, sessionmaker
engine = create_engine('sqlite:////tmp/test.db', convert_unicode=True)
metadata = MetaData()
db_session = scoped_session(sessionmaker(autocommit=False,
def init_db():
As in the declarative approach, you need to close the session after each request or
application context shutdown. Put this into your application module:
from yourapplication.database import db_session
def shutdown_session(exception=None):
Here is an example table and model (put this into models.py):
from sqlalchemy import Table, Column, Integer, String
from sqlalchemy.orm import mapper
from yourapplication.database import metadata, db_session
class User(object):
query = db_session.query_property()
def __init__(self, name=None, email=None):
self.name = name
self.email = email
def __repr__(self):
return '<User %r>' % (self.name)
users = Table('users', metadata,
Column('id', Integer, primary_key=True),
Column('name', String(50), unique=True),
Column('email', String(120), unique=True)
mapper(User, users)
Querying and inserting works exactly the same as in the example above.
SQL Abstraction Layer
If you just want to use the database system (and SQL) abstraction layer you basically
only need the engine:
from sqlalchemy import create_engine, MetaData, Table
engine = create_engine('sqlite:////tmp/test.db', convert_unicode=True)
metadata = MetaData(bind=engine)
Then you can either declare the tables in your code like in the examples above, or
automatically load them:
from sqlalchemy import Table
users = Table('users', metadata, autoload=True)
To insert data you can use the insert method. We have to get a connection first so that
we can use a transaction:
>>> con = engine.connect()
>>> con.execute(users.insert(), name='admin', email='[email protected]')
SQLAlchemy will automatically commit for us.
To query your database, you use the engine directly or use a connection:
>>> users.select(users.c.id == 1).execute().first()
(1, u'admin', u'[email protected]')
These results are also dict-like tuples:
>>> r = users.select(users.c.id == 1).execute().first()
>>> r['name']
You can also pass strings of SQL statements to the execute() method:
>>> engine.execute('select * from users where id = :1', [1]).first()
(1, u'admin', u'[email protected]')
For more information about SQLAlchemy, head over to the website.
Uploading Files
Ah yes, the good old problem of file uploads. The basic idea of file uploads is actually
quite simple. It basically works like this:
1. A <form> tag is marked with enctype=multipart/form-data and an <input
type=file> is placed in that form.
2. The application accesses the file from the files dictionary on the request object.
3. use the save() method of the file to save the file permanently somewhere on the
A Gentle Introduction
Let’s start with a very basic application that uploads a file to a specific upload folder
and displays a file to the user. Let’s look at the bootstrapping code for our application:
import os
from flask import Flask, flash, request, redirect, url_for
from werkzeug.utils import secure_filename
UPLOAD_FOLDER = '/path/to/the/uploads'
ALLOWED_EXTENSIONS = set(['txt', 'pdf', 'png', 'jpg', 'jpeg', 'gif'])
app = Flask(__name__)
So first we need a couple of imports. Most should be straightforward, the werkzeug.
secure_filename() is explained a little bit later. The UPLOAD_FOLDER is where we will
store the uploaded files and the ALLOWED_EXTENSIONS is the set of allowed file extensions.
Why do we limit the extensions that are allowed? You probably don’t want your users
to be able to upload everything there if the server is directly sending out the data to the
client. That way you can make sure that users are not able to upload HTML files that
would cause XSS problems (see Cross-Site Scripting (XSS)). Also make sure to disallow
.php files if the server executes them, but who has PHP installed on their server, right?
Next the functions that check if an extension is valid and that uploads the file and
redirects the user to the URL for the uploaded file:
def allowed_file(filename):
return '.' in filename and \
filename.rsplit('.', 1)[1].lower() in ALLOWED_EXTENSIONS
@app.route('/', methods=['GET', 'POST'])
def upload_file():
if request.method == 'POST':
# check if the post request has the file part
if 'file' not in request.files:
flash('No file part')
return redirect(request.url)
file = request.files['file']
# if user does not select file, browser also
# submit an empty part without filename
if file.filename == '':
flash('No selected file')
return redirect(request.url)
if file and allowed_file(file.filename):
filename = secure_filename(file.filename)
file.save(os.path.join(app.config['UPLOAD_FOLDER'], filename))
return redirect(url_for('uploaded_file',
return '''
<!doctype html>
<title>Upload new File</title>
<h1>Upload new File</h1>
<form method=post enctype=multipart/form-data>
<input type=file name=file>
<input type=submit value=Upload>
So what does that secure_filename() function actually do? Now the problem is that
there is that principle called “never trust user input”. This is also true for the filename
of an uploaded file. All submitted form data can be forged, and filenames can be dangerous. For the moment just remember: always use that function to secure a filename
before storing it directly on the filesystem.
Information for the Pros
So you’re interested in what that secure_filename() function does and what the problem is if you’re not using it? So just imagine someone would send the following information as filename to your application:
filename = "../../../../home/username/.bashrc"
Assuming the number of ../ is correct and you would join this with the UPLOAD_FOLDER
the user might have the ability to modify a file on the server’s filesystem he or she
should not modify. This does require some knowledge about how the application
looks like, but trust me, hackers are patient :)
Now let’s look how that function works:
>>> secure_filename('../../../../home/username/.bashrc')
Now one last thing is missing: the serving of the uploaded files. In the upload_file()
we redirect the user to url_for('uploaded_file', filename=filename), that is, /
uploads/filename. So we write the uploaded_file() function to return the file of that
name. As of Flask 0.5 we can use a function that does that for us:
from flask import send_from_directory
def uploaded_file(filename):
return send_from_directory(app.config['UPLOAD_FOLDER'],
Alternatively you can register uploaded_file as build_only rule and use the
SharedDataMiddleware. This also works with older versions of Flask:
from werkzeug import SharedDataMiddleware
app.add_url_rule('/uploads/<filename>', 'uploaded_file',
app.wsgi_app = SharedDataMiddleware(app.wsgi_app, {
'/uploads': app.config['UPLOAD_FOLDER']
If you now run the application everything should work as expected.
Improving Uploads
New in version 0.6.
So how exactly does Flask handle uploads? Well it will store them in the webserver’s
memory if the files are reasonable small otherwise in a temporary location (as returned
by tempfile.gettempdir()). But how do you specify the maximum file size after
which an upload is aborted? By default Flask will happily accept file uploads to an unlimited amount of memory, but you can limit that by setting the MAX_CONTENT_LENGTH
config key:
from flask import Flask, Request
app = Flask(__name__)
app.config['MAX_CONTENT_LENGTH'] = 16 * 1024 * 1024
The code above will limited the maximum allowed payload to 16 megabytes. If a
larger file is transmitted, Flask will raise an RequestEntityTooLarge exception.
This feature was added in Flask 0.6 but can be achieved in older versions as well by
subclassing the request object. For more information on that consult the Werkzeug
documentation on file handling.
Upload Progress Bars
A while ago many developers had the idea to read the incoming file in small chunks
and store the upload progress in the database to be able to poll the progress with
JavaScript from the client. Long story short: the client asks the server every 5 seconds
how much it has transmitted already. Do you realize the irony? The client is asking
for something it should already know.
An Easier Solution
Now there are better solutions that work faster and are more reliable. There are
JavaScript libraries like jQuery that have form plugins to ease the construction of
progress bar.
Because the common pattern for file uploads exists almost unchanged in all applications dealing with uploads, there is also a Flask extension called Flask-Uploads that
implements a full fledged upload mechanism with white and blacklisting of extensions and more.
When your application runs slow, throw some caches in. Well, at least it’s the easiest
way to speed up things. What does a cache do? Say you have a function that takes
some time to complete but the results would still be good enough if they were 5 minutes old. So then the idea is that you actually put the result of that calculation into a
cache for some time.
Flask itself does not provide caching for you, but Werkzeug, one of the libraries it is
based on, has some very basic cache support. It supports multiple cache backends,
normally you want to use a memcached server.
Setting up a Cache
You create a cache object once and keep it around, similar to how Flask objects are
created. If you are using the development server you can create a SimpleCache object,
that one is a simple cache that keeps the item stored in the memory of the Python
from werkzeug.contrib.cache import SimpleCache
cache = SimpleCache()
If you want to use memcached, make sure to have one of the memcache modules
supported (you get them from PyPI) and a memcached server running somewhere.
This is how you connect to such an memcached server then:
from werkzeug.contrib.cache import MemcachedCache
cache = MemcachedCache([''])
If you are using App Engine, you can connect to the App Engine memcache server
from werkzeug.contrib.cache import GAEMemcachedCache
cache = GAEMemcachedCache()
Using a Cache
Now how can one use such a cache? There are two very important operations: get()
and set(). This is how to use them:
To get an item from the cache call get() with a string as key name. If something is in
the cache, it is returned. Otherwise that function will return None:
rv = cache.get('my-item')
To add items to the cache, use the set() method instead. The first argument is the
key and the second the value that should be set. Also a timeout can be provided after
which the cache will automatically remove item.
Here a full example how this looks like normally:
def get_my_item():
rv = cache.get('my-item')
if rv is None:
rv = calculate_value()
cache.set('my-item', rv, timeout=5 * 60)
return rv
View Decorators
Python has a really interesting feature called function decorators. This allows some
really neat things for web applications. Because each view in Flask is a function, decorators can be used to inject additional functionality to one or more functions. The
route() decorator is the one you probably used already. But there are use cases for
implementing your own decorator. For instance, imagine you have a view that should
only be used by people that are logged in. If a user goes to the site and is not logged
in, they should be redirected to the login page. This is a good example of a use case
where a decorator is an excellent solution.
Login Required Decorator
So let’s implement such a decorator. A decorator is a function that wraps and replaces
another function. Since the original function is replaced, you need to remember to
copy the original function’s information to the new function. Use functools.wraps()
to handle this for you.
This example assumes that the login page is called 'login' and that the current user
is stored in g.user and is None if there is no-one logged in.
from functools import wraps
from flask import g, request, redirect, url_for
def login_required(f):
def decorated_function(*args, **kwargs):
if g.user is None:
return redirect(url_for('login', next=request.url))
return f(*args, **kwargs)
return decorated_function
To use the decorator, apply it as innermost decorator to a view function. When applying further decorators, always remember that the route() decorator is the outermost.
def secret_page():
Note: The next value will exist in request.args after a GET request for the login page.
You’ll have to pass it along when sending the POST request from the login form. You
can do this with a hidden input tag, then retrieve it from request.form when logging
the user in.
<input type="hidden" value="{{ request.args.get('next', '') }}"/>
Caching Decorator
Imagine you have a view function that does an expensive calculation and because
of that you would like to cache the generated results for a certain amount of time.
A decorator would be nice for that. We’re assuming you have set up a cache like
mentioned in Caching.
Here is an example cache function. It generates the cache key from a specific prefix
(actually a format string) and the current path of the request. Notice that we are using
a function that first creates the decorator that then decorates the function. Sounds aw-
ful? Unfortunately it is a little bit more complex, but the code should still be straightforward to read.
The decorated function will then work as follows
1. get the unique cache key for the current request base on the current path.
2. get the value for that key from the cache. If the cache returned something we
will return that value.
3. otherwise the original function is called and the return value is stored in the
cache for the timeout provided (by default 5 minutes).
Here the code:
from functools import wraps
from flask import request
def cached(timeout=5 * 60, key='view/%s'):
def decorator(f):
def decorated_function(*args, **kwargs):
cache_key = key % request.path
rv = cache.get(cache_key)
if rv is not None:
return rv
rv = f(*args, **kwargs)
cache.set(cache_key, rv, timeout=timeout)
return rv
return decorated_function
return decorator
Notice that this assumes an instantiated cache object is available, see Caching for more
Templating Decorator
A common pattern invented by the TurboGears guys a while back is a templating
decorator. The idea of that decorator is that you return a dictionary with the values
passed to the template from the view function and the template is automatically rendered. With that, the following three examples do exactly the same:
def index():
return render_template('index.html', value=42)
def index():
return dict(value=42)
def index():
return dict(value=42)
As you can see, if no template name is provided it will use the endpoint of the URL
map with dots converted to slashes + '.html'. Otherwise the provided template name
is used. When the decorated function returns, the dictionary returned is passed to the
template rendering function. If None is returned, an empty dictionary is assumed, if
something else than a dictionary is returned we return it from the function unchanged.
That way you can still use the redirect function or return simple strings.
Here is the code for that decorator:
from functools import wraps
from flask import request, render_template
def templated(template=None):
def decorator(f):
def decorated_function(*args, **kwargs):
template_name = template
if template_name is None:
template_name = request.endpoint \
.replace('.', '/') + '.html'
ctx = f(*args, **kwargs)
if ctx is None:
ctx = {}
elif not isinstance(ctx, dict):
return ctx
return render_template(template_name, **ctx)
return decorated_function
return decorator
Endpoint Decorator
When you want to use the werkzeug routing system for more flexibility you need to
map the endpoint as defined in the Rule to a view function. This is possible with this
decorator. For example:
from flask import Flask
from werkzeug.routing import Rule
app = Flask(__name__)
app.url_map.add(Rule('/', endpoint='index'))
def my_index():
return "Hello world"
Form Validation with WTForms
When you have to work with form data submitted by a browser view, code quickly
becomes very hard to read. There are libraries out there designed to make this process
easier to manage. One of them is WTForms which we will handle here. If you find
yourself in the situation of having many forms, you might want to give it a try.
When you are working with WTForms you have to define your forms as classes first.
I recommend breaking up the application into multiple modules (Larger Applications)
for that and adding a separate module for the forms.
Getting the most out of WTForms with an Extension
The Flask-WTF extension expands on this pattern and adds a few little helpers that
make working with forms and Flask more fun. You can get it from PyPI.
The Forms
This is an example form for a typical registration page:
from wtforms import Form, BooleanField, StringField, PasswordField, validators
class RegistrationForm(Form):
username = StringField('Username', [validators.Length(min=4, max=25)])
email = StringField('Email Address', [validators.Length(min=6, max=35)])
password = PasswordField('New Password', [
validators.EqualTo('confirm', message='Passwords must match')
confirm = PasswordField('Repeat Password')
accept_tos = BooleanField('I accept the TOS', [validators.DataRequired()])
In the View
In the view function, the usage of this form looks like this:
@app.route('/register', methods=['GET', 'POST'])
def register():
form = RegistrationForm(request.form)
if request.method == 'POST' and form.validate():
user = User(form.username.data, form.email.data,
flash('Thanks for registering')
return redirect(url_for('login'))
return render_template('register.html', form=form)
Notice we’re implying that the view is using SQLAlchemy here (SQLAlchemy in Flask),
but that’s not a requirement, of course. Adapt the code as necessary.
Things to remember:
1. create the form from the request form value if the data is submitted via the HTTP
POST method and args if the data is submitted as GET.
2. to validate the data, call the validate() method, which will return True if the
data validates, False otherwise.
3. to access individual values from the form, access form.<NAME>.data.
Forms in Templates
Now to the template side. When you pass the form to the templates, you can easily
render them there. Look at the following example template to see how easy this is.
WTForms does half the form generation for us already. To make it even nicer, we can
write a macro that renders a field with label and a list of errors if there are any.
Here’s an example _formhelpers.html template with such a macro:
{% macro render_field(field) %}
<dt>{{ field.label }}
<dd>{{ field(**kwargs)|safe }}
{% if field.errors %}
<ul class=errors>
{% for error in field.errors %}
<li>{{ error }}</li>
{% endfor %}
{% endif %}
{% endmacro %}
This macro accepts a couple of keyword arguments that are forwarded to WTForm’s
field function, which renders the field for us. The keyword arguments will be inserted
as HTML attributes. So, for example, you can call render_field(form.username,
class='username') to add a class to the input element. Note that WTForms returns
standard Python unicode strings, so we have to tell Jinja2 that this data is already
HTML-escaped with the |safe filter.
Here is the register.html template for the function we used above, which takes advantage of the _formhelpers.html template:
{% from "_formhelpers.html" import render_field %}
<form method=post>
{{ render_field(form.username) }}
{{ render_field(form.email) }}
{{ render_field(form.password) }}
{{ render_field(form.confirm) }}
{{ render_field(form.accept_tos) }}
<p><input type=submit value=Register>
For more information about WTForms, head over to the WTForms website.
Template Inheritance
The most powerful part of Jinja is template inheritance. Template inheritance allows
you to build a base “skeleton” template that contains all the common elements of your
site and defines blocks that child templates can override.
Sounds complicated but is very basic. It’s easiest to understand it by starting with an
Base Template
This template, which we’ll call layout.html, defines a simple HTML skeleton document that you might use for a simple two-column page. It’s the job of “child” templates to fill the empty blocks with content:
<!doctype html>
{% block head %}
<link rel="stylesheet" href="{{ url_for('static', filename='style.css') }}">
<title>{% block title %}{% endblock %} - My Webpage</title>
{% endblock %}
<div id="content">{% block content %}{% endblock %}</div>
<div id="footer">
{% block footer %}
&copy; Copyright 2010 by <a href="http://domain.invalid/">you</a>.
{% endblock %}
In this example, the {% block %} tags define four blocks that child templates can fill
in. All the block tag does is tell the template engine that a child template may override
those portions of the template.
Child Template
A child template might look like this:
{% extends "layout.html" %}
{% block title %}Index{% endblock %}
{% block head %}
{{ super() }}
<style type="text/css">
.important { color: #336699; }
{% endblock %}
{% block content %}
<p class="important">
Welcome on my awesome homepage.
{% endblock %}
The {% extends %} tag is the key here. It tells the template engine that this template
“extends” another template. When the template system evaluates this template, first
it locates the parent. The extends tag must be the first tag in the template. To render
the contents of a block defined in the parent template, use {{ super() }}.
Message Flashing
Good applications and user interfaces are all about feedback. If the user does not get
enough feedback they will probably end up hating the application. Flask provides a
really simple way to give feedback to a user with the flashing system. The flashing
system basically makes it possible to record a message at the end of a request and
access it next request and only next request. This is usually combined with a layout
template that does this. Note that browsers and sometimes web servers enforce a
limit on cookie sizes. This means that flashing messages that are too large for session
cookies causes message flashing to fail silently.
Simple Flashing
So here is a full example:
from flask import Flask, flash, redirect, render_template, \
request, url_for
app = Flask(__name__)
app.secret_key = 'some_secret'
def index():
return render_template('index.html')
@app.route('/login', methods=['GET', 'POST'])
def login():
error = None
if request.method == 'POST':
if request.form['username'] != 'admin' or \
request.form['password'] != 'secret':
error = 'Invalid credentials'
flash('You were successfully logged in')
return redirect(url_for('index'))
return render_template('login.html', error=error)
And here is the layout.html template which does the magic:
<!doctype html>
<title>My Application</title>
{% with messages = get_flashed_messages() %}
{% if messages %}
<ul class=flashes>
{% for message in messages %}
<li>{{ message }}</li>
{% endfor %}
{% endif %}
{% endwith %}
{% block body %}{% endblock %}
Here is the index.html template which inherits from layout.html:
{% extends "layout.html" %}
{% block body %}
<p>Do you want to <a href="{{ url_for('login') }}">log in?</a>
{% endblock %}
And here is the login.html template which also inherits from layout.html:
{% extends "layout.html" %}
{% block body %}
{% if error %}
<p class=error><strong>Error:</strong> {{ error }}
{% endif %}
<form method=post>
<dd><input type=text name=username value="{{
request.form.username }}">
<dd><input type=password name=password>
<p><input type=submit value=Login>
{% endblock %}
Flashing With Categories
New in version 0.3.
It is also possible to provide categories when flashing a message. The default category if nothing is provided is 'message'. Alternative categories can be used to give
the user better feedback. For example error messages could be displayed with a red
To flash a message with a different category, just use the second argument to the
flash() function:
flash(u'Invalid password provided', 'error')
Inside the template you then have to tell the get_flashed_messages() function to also
return the categories. The loop looks slightly different in that situation then:
{% with messages = get_flashed_messages(with_categories=true) %}
{% if messages %}
<ul class=flashes>
{% for category, message in messages %}
<li class="{{ category }}">{{ message }}</li>
{% endfor %}
{% endif %}
{% endwith %}
This is just one example of how to render these flashed messages. One might also use
the category to add a prefix such as <strong>Error:</strong> to the message.
Filtering Flash Messages
New in version 0.9.
Optionally you can pass a list of categories which filters the results of
get_flashed_messages(). This is useful if you wish to render each category in a separate block.
{% with errors = get_flashed_messages(category_filter=["error"]) %}
{% if errors %}
<div class="alert-message block-message error">
<a class="close" href="#">×</a>
{%- for msg in errors %}
<li>{{ msg }}</li>
{% endfor -%}
{% endif %}
{% endwith %}
AJAX with jQuery
jQuery is a small JavaScript library commonly used to simplify working with the DOM
and JavaScript in general. It is the perfect tool to make web applications more dynamic
by exchanging JSON between server and client.
JSON itself is a very lightweight transport format, very similar to how Python primitives (numbers, strings, dicts and lists) look like which is widely supported and very
easy to parse. It became popular a few years ago and quickly replaced XML as transport format in web applications.
Loading jQuery
In order to use jQuery, you have to download it first and place it in the static folder of
your application and then ensure it’s loaded. Ideally you have a layout template that
is used for all pages where you just have to add a script statement to the bottom of
your <body> to load jQuery:
<script type=text/javascript src="{{
url_for('static', filename='jquery.js') }}"></script>
Another method is using Google’s AJAX Libraries API to load jQuery:
<script src="//ajax.googleapis.com/ajax/libs/jquery/1.9.1/jquery.min.js"></script>
<script>window.jQuery || document.write('<script src="{{
url_for('static', filename='jquery.js') }}">\x3C/script>')</script>
In this case you have to put jQuery into your static folder as a fallback, but it will
first try to load it directly from Google. This has the advantage that your website will
probably load faster for users if they went to at least one other website before using
the same jQuery version from Google because it will already be in the browser cache.
Where is My Site?
Do you know where your application is? If you are developing the answer is quite
simple: it’s on localhost port something and directly on the root of that server. But
what if you later decide to move your application to a different location? For example
to http://example.com/myapp? On the server side this never was a problem because
we were using the handy url_for() function that could answer that question for us,
but if we are using jQuery we should not hardcode the path to the application but
make that dynamic, so how can we do that?
A simple method would be to add a script tag to our page that sets a global variable
to the prefix to the root of the application. Something like this:
<script type=text/javascript>
$SCRIPT_ROOT = {{ request.script_root|tojson|safe }};
The |safe is necessary in Flask before 0.10 so that Jinja does not escape the JSON
encoded string with HTML rules. Usually this would be necessary, but we are inside
a script block here where different rules apply.
Information for Pros
In HTML the script tag is declared CDATA which means that entities will not be parsed.
Everything until </script> is handled as script. This also means that there must
never be any </ between the script tags. |tojson is kind enough to do the right thing
here and escape slashes for you ({{ "</script>"|tojson|safe }} is rendered as "<\/
In Flask 0.10 it goes a step further and escapes all HTML tags with unicode escapes.
This makes it possible for Flask to automatically mark the result as HTML safe.
JSON View Functions
Now let’s create a server side function that accepts two URL arguments of numbers
which should be added together and then sent back to the application in a JSON object.
This is a really ridiculous example and is something you usually would do on the
client side alone, but a simple example that shows how you would use jQuery and
Flask nonetheless:
from flask import Flask, jsonify, render_template, request
app = Flask(__name__)
def add_numbers():
a = request.args.get('a', 0, type=int)
b = request.args.get('b', 0, type=int)
return jsonify(result=a + b)
def index():
return render_template('index.html')
As you can see I also added an index method here that renders a template. This template will load jQuery as above and have a little form we can add two numbers and a
link to trigger the function on the server side.
Note that we are using the get() method here which will never fail. If the key is
missing a default value (here 0) is returned. Furthermore it can convert values to a
specific type (like in our case int). This is especially handy for code that is triggered by
a script (APIs, JavaScript etc.) because you don’t need special error reporting in that
Your index.html template either has to extend a layout.html template with jQuery
loaded and the $SCRIPT_ROOT variable set, or do that on the top. Here’s the HTML
code needed for our little application (index.html). Notice that we also drop the script
directly into the HTML here. It is usually a better idea to have that in a separate script
<script type=text/javascript>
$(function() {
$('a#calculate').bind('click', function() {
$.getJSON($SCRIPT_ROOT + '/_add_numbers', {
a: $('input[name="a"]').val(),
b: $('input[name="b"]').val()
}, function(data) {
return false;
<h1>jQuery Example</h1>
<p><input type=text size=5 name=a> +
<input type=text size=5 name=b> =
<span id=result>?</span>
<p><a href=# id=calculate>calculate server side</a>
I won’t go into detail here about how jQuery works, just a very quick explanation of
the little bit of code above:
1. $(function() { ... }) specifies code that should run once the browser is done
loading the basic parts of the page.
2. $('selector') selects an element and lets you operate on it.
3. element.bind('event', func) specifies a function that should run when the user
clicked on the element. If that function returns false, the default behavior will not
kick in (in this case, navigate to the # URL).
4. $.getJSON(url, data, func) sends a GET request to url and will send the contents of the data object as query parameters. Once the data arrived, it will call
the given function with the return value as argument. Note that we can use the
$SCRIPT_ROOT variable here that we set earlier.
If you don’t get the whole picture, download the sourcecode for this example from
Custom Error Pages
Flask comes with a handy abort() function that aborts a request with an HTTP error
code early. It will also provide a plain black and white error page for you with a basic
description, but nothing fancy.
Depending on the error code it is less or more likely for the user to actually see such
an error.
Common Error Codes
The following error codes are some that are often displayed to the user, even if the
application behaves correctly:
404 Not Found The good old “chap, you made a mistake typing that URL” message.
So common that even novices to the internet know that 404 means: damn, the
thing I was looking for is not there. It’s a very good idea to make sure there is
actually something useful on a 404 page, at least a link back to the index.
403 Forbidden If you have some kind of access control on your website, you will have
to send a 403 code for disallowed resources. So make sure the user is not lost
when they try to access a forbidden resource.
410 Gone Did you know that there the “404 Not Found” has a brother named “410
Gone”? Few people actually implement that, but the idea is that resources that
previously existed and got deleted answer with 410 instead of 404. If you are
not deleting documents permanently from the database but just mark them as
deleted, do the user a favour and use the 410 code instead and display a message
that what they were looking for was deleted for all eternity.
500 Internal Server Error Usually happens on programming errors or if the server is
overloaded. A terribly good idea is to have a nice page there, because your
application will fail sooner or later (see also: Application Errors).
Error Handlers
An error handler is a function that returns a response when a type of error is raised,
similar to how a view is a function that returns a response when a request URL is
matched. It is passed the instance of the error being handled, which is most likely a
HTTPException. An error handler for “500 Internal Server Error” will be passed uncaught exceptions in addition to explicit 500 errors.
An error handler is registered with the errorhandler() decorator or the
register_error_handler() method. A handler can be registered for a status code,
like 404, or for an exception class.
The status code of the response will not be set to the handler’s code. Make sure to
provide the appropriate HTTP status code when returning a response from a handler.
A handler for “500 Internal Server Error” will not be used when running in debug
mode. Instead, the interactive debugger will be shown.
Here is an example implementation for a “404 Page Not Found” exception:
from flask import render_template
def page_not_found(e):
# note that we set the 404 status explicitly
return render_template('404.html'), 404
When using the application factory pattern:
from flask import Flask, render_template
def page_not_found(e):
return render_template('404.html'), 404
def create_app(config_filename):
app = Flask(__name__)
app.register_error_handler(404, page_not_found)
return app
An example template might be this:
{% extends "layout.html" %}
{% block title %}Page Not Found{% endblock %}
{% block body %}
<h1>Page Not Found</h1>
<p>What you were looking for is just not there.
<p><a href="{{ url_for('index') }}">go somewhere nice</a>
{% endblock %}
Lazily Loading Views
Flask is usually used with the decorators. Decorators are simple and you have the
URL right next to the function that is called for that specific URL. However there is
a downside to this approach: it means all your code that uses decorators has to be
imported upfront or Flask will never actually find your function.
This can be a problem if your application has to import quick. It might have to do
that on systems like Google’s App Engine or other systems. So if you suddenly notice
that your application outgrows this approach you can fall back to a centralized URL
The system that enables having a central URL map is the add_url_rule() function.
Instead of using decorators, you have a file that sets up the application with all URLs.
Converting to Centralized URL Map
Imagine the current application looks somewhat like this:
from flask import Flask
app = Flask(__name__)
def index():
def user(username):
Then, with the centralized approach you would have one file with the views (views.
py) but without any decorator:
def index():
def user(username):
And then a file that sets up an application which maps the functions to URLs:
from flask import Flask
from yourapplication import views
app = Flask(__name__)
app.add_url_rule('/', view_func=views.index)
app.add_url_rule('/user/<username>', view_func=views.user)
Loading Late
So far we only split up the views and the routing, but the module is still loaded upfront. The trick is to actually load the view function as needed. This can be accomplished with a helper class that behaves just like a function but internally imports the
real function on first use:
from werkzeug import import_string, cached_property
class LazyView(object):
def __init__(self, import_name):
self.__module__, self.__name__ = import_name.rsplit('.', 1)
self.import_name = import_name
def view(self):
return import_string(self.import_name)
def __call__(self, *args, **kwargs):
return self.view(*args, **kwargs)
What’s important here is is that __module__ and __name__ are properly set. This is
used by Flask internally to figure out how to name the URL rules in case you don’t
provide a name for the rule yourself.
Then you can define your central place to combine the views like this:
from flask import Flask
from yourapplication.helpers import LazyView
app = Flask(__name__)
You can further optimize this in terms of amount of keystrokes needed to write this by
having a function that calls into add_url_rule() by prefixing a string with the project
name and a dot, and by wrapping view_func in a LazyView as needed.
def url(import_name, url_rules=[], **options):
view = LazyView('yourapplication.' + import_name)
for url_rule in url_rules:
app.add_url_rule(url_rule, view_func=view, **options)
# add a single route to the index view
url('views.index', ['/'])
# add two routes to a single function endpoint
url_rules = ['/user/','/user/<username>']
url('views.user', url_rules)
One thing to keep in mind is that before and after request handlers have to be in a file
that is imported upfront to work properly on the first request. The same goes for any
kind of remaining decorator.
MongoKit in Flask
Using a document database rather than a full DBMS gets more common these days.
This pattern shows how to use MongoKit, a document mapper library, to integrate
with MongoDB.
This pattern requires a running MongoDB server and the MongoKit library installed.
There are two very common ways to use MongoKit. I will outline each of them here:
The default behavior of MongoKit is the declarative one that is based on common
ideas from Django or the SQLAlchemy declarative extension.
Here an example app.py module for your application:
from flask import Flask
from mongokit import Connection, Document
# configuration
MONGODB_HOST = 'localhost'
# create the little application object
app = Flask(__name__)
# connect to the database
connection = Connection(app.config['MONGODB_HOST'],
To define your models, just subclass the Document class that is imported from MongoKit. If you’ve seen the SQLAlchemy pattern you may wonder why we do not have
a session and even do not define a init_db function here. On the one hand, MongoKit
does not have something like a session. This sometimes makes it more to type but
also makes it blazingly fast. On the other hand, MongoDB is schemaless. This means
you can modify the data structure from one insert query to the next without any problem. MongoKit is just schemaless too, but implements some validation to ensure data
Here is an example document (put this also into app.py, e.g.):
from mongokit import ValidationError
def max_length(length):
def validate(value):
if len(value) <= length:
return True
# must have %s in error format string to have mongokit place key in there
raise ValidationError('%s must be at most {} characters long'.
return validate
class User(Document):
structure = {
'name': unicode,
'email': unicode,
validators = {
'name': max_length(50),
'email': max_length(120)
use_dot_notation = True
def __repr__(self):
return '<User %r>' % (self.name)
# register the User document with our current connection
This example shows you how to define your schema (named structure), a validator for the maximum character length and uses a special MongoKit feature called
use_dot_notation. Per default MongoKit behaves like a python dictionary but with
use_dot_notation set to True you can use your documents like you use models in nearly
any other ORM by using dots to separate between attributes.
You can insert entries into the database like this:
from yourapplication.database import connection
from yourapplication.models import User
collection = connection['test'].users
user = collection.User()
user['name'] = u'admin'
user['email'] = u'[email protected]'
Note that MongoKit is kinda strict with used column types, you must not use a common str type for either name or email but unicode.
Querying is simple as well:
>>> list(collection.User.find())
[<User u'admin'>]
>>> collection.User.find_one({'name': u'admin'})
<User u'admin'>
PyMongo Compatibility Layer
If you just want to use PyMongo, you can do that with MongoKit as well. You may
use this process if you need the best performance to get. Note that this example does
not show how to couple it with Flask, see the above MongoKit code for examples:
from MongoKit import Connection
connection = Connection()
To insert data you can use the insert method. We have to get a collection first, this is
somewhat the same as a table in the SQL world.
>>> collection = connection['test'].users
>>> user = {'name': u'admin', 'email': u'[email protected]'}
>>> collection.insert(user)
MongoKit will automatically commit for us.
To query your database, you use the collection directly:
>>> list(collection.find())
[{u'_id': ObjectId('4c271729e13823182f000000'), u'name': u'admin', u'email': u
,→'[email protected]'}]
>>> collection.find_one({'name': u'admin'})
{u'_id': ObjectId('4c271729e13823182f000000'), u'name': u'admin', u'email': u
,→'[email protected]'}
These results are also dict-like objects:
>>> r = collection.find_one({'name': u'admin'})
>>> r['email']
u'[email protected]'
For more information about MongoKit, head over to the website.
Adding a favicon
A “favicon” is an icon used by browsers for tabs and bookmarks. This helps to distinguish your website and to give it a unique brand.
A common question is how to add a favicon to a Flask application. First, of course,
you need an icon. It should be 16 × 16 pixels and in the ICO file format. This is not a
requirement but a de-facto standard supported by all relevant browsers. Put the icon
in your static directory as favicon.ico.
Now, to get browsers to find your icon, the correct way is to add a link tag in your
HTML. So, for example:
<link rel="shortcut icon" href="{{ url_for('static', filename='favicon.ico') }}">
That’s all you need for most browsers, however some really old ones do not support
this standard. The old de-facto standard is to serve this file, with this name, at the
website root. If your application is not mounted at the root path of the domain you
either need to configure the web server to serve the icon at the root or if you can’t do
that you’re out of luck. If however your application is the root you can simply route a
redirect_to=url_for('static', filename='favicon.ico'))
If you want to save the extra redirect request you can also write a view using
import os
from flask import send_from_directory
def favicon():
return send_from_directory(os.path.join(app.root_path, 'static'),
'favicon.ico', mimetype='image/vnd.microsoft.icon')
We can leave out the explicit mimetype and it will be guessed, but we may as well
specify it to avoid the extra guessing, as it will always be the same.
The above will serve the icon via your application and if possible it’s better to configure your dedicated web server to serve it; refer to the web server’s documentation.
See also
• The Favicon article on Wikipedia
Streaming Contents
Sometimes you want to send an enormous amount of data to the client, much more
than you want to keep in memory. When you are generating the data on the fly though,
how do you send that back to the client without the roundtrip to the filesystem?
The answer is by using generators and direct responses.
Basic Usage
This is a basic view function that generates a lot of CSV data on the fly. The trick is to
have an inner function that uses a generator to generate data and to then invoke that
function and pass it to a response object:
from flask import Response
def generate_large_csv():
def generate():
for row in iter_all_rows():
yield ','.join(row) + '\n'
return Response(generate(), mimetype='text/csv')
Each yield expression is directly sent to the browser. Note though that some WSGI
middlewares might break streaming, so be careful there in debug environments with
profilers and other things you might have enabled.
Streaming from Templates
The Jinja2 template engine also supports rendering templates piece by piece. This
functionality is not directly exposed by Flask because it is quite uncommon, but you
can easily do it yourself:
from flask import Response
def stream_template(template_name, **context):
t = app.jinja_env.get_template(template_name)
rv = t.stream(context)
return rv
def render_large_template():
rows = iter_all_rows()
return Response(stream_template('the_template.html', rows=rows))
The trick here is to get the template object from the Jinja2 environment on the application and to call stream() instead of render() which returns a stream object instead
of a string. Since we’re bypassing the Flask template render functions and using the
template object itself we have to make sure to update the render context ourselves by
calling update_template_context(). The template is then evaluated as the stream is
iterated over. Since each time you do a yield the server will flush the content to the
client you might want to buffer up a few items in the template which you can do with
rv.enable_buffering(size). 5 is a sane default.
Streaming with Context
New in version 0.9.
Note that when you stream data, the request context is already gone the moment the
function executes. Flask 0.9 provides you with a helper that can keep the request
context around during the execution of the generator:
from flask import stream_with_context, request, Response
def streamed_response():
def generate():
yield 'Hello '
yield request.args['name']
yield '!'
return Response(stream_with_context(generate()))
Without the stream_with_context() function you would get a RuntimeError at that
Deferred Request Callbacks
One of the design principles of Flask is that response objects are created and passed
down a chain of potential callbacks that can modify them or replace them. When the
request handling starts, there is no response object yet. It is created as necessary either
by a view function or by some other component in the system.
What happens if you want to modify the response at a point where the response does
not exist yet? A common example for that would be a before_request() callback that
wants to set a cookie on the response object.
One way is to avoid the situation. Very often that is possible. For instance you can
try to move that logic into a after_request() callback instead. However, sometimes
moving code there makes it more more complicated or awkward to reason about.
As an alternative, you can use after_this_request() to register callbacks that will
execute after only the current request. This way you can defer code execution from
anywhere in the application, based on the current request.
At any time during a request, we can register a function to be called at the end of the
request. For example you can remember the current language of the user in a cookie
in a before_request() callback:
from flask import request, after_this_request
def detect_user_language():
language = request.cookies.get('user_lang')
if language is None:
language = guess_language_from_request()
# when the response exists, set a cookie with the language
def remember_language(response):
response.set_cookie('user_lang', language)
g.language = language
Adding HTTP Method Overrides
Some HTTP proxies do not support arbitrary HTTP methods or newer HTTP methods
(such as PATCH). In that case it’s possible to “proxy” HTTP methods through another
HTTP method in total violation of the protocol.
The way this works is by letting the client do an HTTP POST request and set the
X-HTTP-Method-Override header and set the value to the intended HTTP method (such
as PATCH).
This can easily be accomplished with an HTTP middleware:
class HTTPMethodOverrideMiddleware(object):
allowed_methods = frozenset([
bodyless_methods = frozenset(['GET', 'HEAD', 'OPTIONS', 'DELETE'])
def __init__(self, app):
self.app = app
def __call__(self, environ, start_response):
method = environ.get('HTTP_X_HTTP_METHOD_OVERRIDE', '').upper()
if method in self.allowed_methods:
method = method.encode('ascii', 'replace')
environ['REQUEST_METHOD'] = method
if method in self.bodyless_methods:
environ['CONTENT_LENGTH'] = '0'
return self.app(environ, start_response)
To use this with Flask this is all that is necessary:
from flask import Flask
app = Flask(__name__)
app.wsgi_app = HTTPMethodOverrideMiddleware(app.wsgi_app)
Request Content Checksums
Various pieces of code can consume the request data and preprocess it. For instance
JSON data ends up on the request object already read and processed, form data ends
up there as well but goes through a different code path. This seems inconvenient when
you want to calculate the checksum of the incoming request data. This is necessary
sometimes for some APIs.
Fortunately this is however very simple to change by wrapping the input stream.
The following example calculates the SHA1 checksum of the incoming data as it gets
read and stores it in the WSGI environment:
import hashlib
class ChecksumCalcStream(object):
def __init__(self, stream):
self._stream = stream
self._hash = hashlib.sha1()
def read(self, bytes):
rv = self._stream.read(bytes)
return rv
def readline(self, size_hint):
rv = self._stream.readline(size_hint)
return rv
def generate_checksum(request):
env = request.environ
stream = ChecksumCalcStream(env['wsgi.input'])
env['wsgi.input'] = stream
return stream._hash
To use this, all you need to do is to hook the calculating stream in before the request
starts consuming data. (Eg: be careful accessing request.form or anything of that
nature. before_request_handlers for instance should be careful not to access it).
Example usage:
@app.route('/special-api', methods=['POST'])
def special_api():
hash = generate_checksum(request)
# Accessing this parses the input stream
files = request.files
# At this point the hash is fully constructed.
checksum = hash.hexdigest()
return 'Hash was: %s' % checksum
Celery Background Tasks
If your application has a long running task, such as processing some uploaded data or
sending email, you don’t want to wait for it to finish during a request. Instead, use a
task queue to send the necessary data to another process that will run the task in the
background while the request returns immediately.
Celery is a powerful task queue that can be used for simple background tasks as well
as complex multi-stage programs and schedules. This guide will show you how to
configure Celery using Flask, but assumes you’ve already read the First Steps with
Celery guide in the Celery documentation.
Celery is a separate Python package. Install it from PyPI using pip:
$ pip install celery
The first thing you need is a Celery instance, this is called the celery application. It
serves the same purpose as the Flask object in Flask, just for Celery. Since this instance
is used as the entry-point for everything you want to do in Celery, like creating tasks
and managing workers, it must be possible for other modules to import it.
For instance you can place this in a tasks module. While you can use Celery without
any reconfiguration with Flask, it becomes a bit nicer by subclassing tasks and adding
support for Flask’s application contexts and hooking it up with the Flask configuration.
This is all that is necessary to properly integrate Celery with Flask:
from celery import Celery
def make_celery(app):
celery = Celery(
class ContextTask(celery.Task):
def __call__(self, *args, **kwargs):
with app.app_context():
return self.run(*args, **kwargs)
celery.Task = ContextTask
return celery
The function creates a new Celery object, configures it with the broker from the application config, updates the rest of the Celery config from the Flask config and then
creates a subclass of the task that wraps the task execution in an application context.
An example task
Let’s write a task that adds two numbers together and returns the result. We configure
Celery’s broker and backend to use Redis, create a celery application using the factor
from above, and then use it to define the task.
from flask import Flask
flask_app = Flask(__name__)
celery = make_celery(flask_app)
def add_together(a, b):
return a + b
This task can now be called in the background:
result = add_together.delay(23, 42)
result.wait() # 65
Run a worker
If you jumped in and already executed the above code you will be disappointed to
learn that .wait() will never actually return. That’s because you also need to run a
Celery worker to receive and execute the task.
$ celery -A your_application.celery worker
The your_application string has to point to your application’s package or module that
creates the celery object.
Now that the worker is running, wait will return the result once the task is finished.
Subclassing Flask
The Flask class is designed for subclassing.
For example, you may want to override how request parameters are handled to preserve their order:
from flask import Flask, Request
from werkzeug.datastructures import ImmutableOrderedMultiDict
class MyRequest(Request):
"""Request subclass to override request parameter storage"""
parameter_storage_class = ImmutableOrderedMultiDict
class MyFlask(Flask):
"""Flask subclass using the custom request class"""
request_class = MyRequest
This is the recommended approach for overriding or augmenting Flask’s internal functionality.
Deployment Options
While lightweight and easy to use, Flask’s built-in server is not suitable for production as it doesn’t scale well and by default serves only one request at a time. Some of
the options available for properly running Flask in production are documented here.
If you want to deploy your Flask application to a WSGI server not listed here, look up
the server documentation about how to use a WSGI app with it. Just remember that
your Flask application object is the actual WSGI application.
Hosted options
• Deploying Flask on Heroku
• Deploying Flask on OpenShift
• Deploying Flask on Webfaction
• Deploying Flask on Google App Engine
• Deploying Flask on AWS Elastic Beanstalk
• Sharing your Localhost Server with Localtunnel
• Deploying on Azure (IIS)
• Deploying on PythonAnywhere
Self-hosted options
Standalone WSGI Containers
There are popular servers written in Python that contain WSGI applications and serve
HTTP. These servers stand alone when they run; you can proxy to them from your
web server. Note the section on Proxy Setups if you run into issues.
Gunicorn ‘Green Unicorn’ is a WSGI HTTP Server for UNIX. It’s a pre-fork worker
model ported from Ruby’s Unicorn project. It supports both eventlet and greenlet.
Running a Flask application on this server is quite simple:
gunicorn myproject:app
Gunicorn provides many command-line options – see gunicorn -h. For example, to
run a Flask application with 4 worker processes (-w 4) binding to localhost port 4000
gunicorn -w 4 -b myproject:app
uWSGI is a fast application server written in C. It is very configurable which makes it
more complicated to setup than gunicorn.
Running uWSGI HTTP Router:
uwsgi --http --module myproject:app
For a more optimized setup, see configuring uWSGI and NGINX.
Gevent is a coroutine-based Python networking library that uses greenlet to provide a
high-level synchronous API on top of libev event loop:
from gevent.wsgi import WSGIServer
from yourapplication import app
http_server = WSGIServer(('', 5000), app)
Twisted Web
Twisted Web is the web server shipped with Twisted, a mature, non-blocking eventdriven networking library. Twisted Web comes with a standard WSGI container which
can be controlled from the command line using the twistd utility:
twistd web --wsgi myproject.app
This example will run a Flask application called app from a module named myproject.
Twisted Web supports many flags and options, and the twistd utility does as well; see
twistd -h and twistd web -h for more information. For example, to run a Twisted
Web server in the foreground, on port 8080, with an application from myproject:
twistd -n web --port 8080 --wsgi myproject.app
Proxy Setups
If you deploy your application using one of these servers behind an HTTP proxy you
will need to rewrite a few headers in order for the application to work. The two problematic values in the WSGI environment usually are REMOTE_ADDR and HTTP_HOST. You
can configure your httpd to pass these headers, or you can fix them in middleware.
Werkzeug ships a fixer that will solve some common setups, but you might want to
write your own WSGI middleware for specific setups.
Here’s a simple nginx configuration which proxies to an application served on localhost at port 8000, setting appropriate headers:
server {
listen 80;
server_name _;
access_log /var/log/nginx/access.log;
error_log /var/log/nginx/error.log;
location / {
If your httpd is not providing these headers, the most common setup invokes the host
being set from X-Forwarded-Host and the remote address from X-Forwarded-For:
from werkzeug.contrib.fixers import ProxyFix
app.wsgi_app = ProxyFix(app.wsgi_app)
Trusting Headers
Please keep in mind that it is a security issue to use such a middleware in a non-proxy
setup because it will blindly trust the incoming headers which might be forged by
malicious clients.
If you want to rewrite the headers from another header, you might want to use a fixer
like this:
class CustomProxyFix(object):
def __init__(self, app):
self.app = app
def __call__(self, environ, start_response):
host = environ.get('HTTP_X_FHOST', '')
if host:
environ['HTTP_HOST'] = host
return self.app(environ, start_response)
app.wsgi_app = CustomProxyFix(app.wsgi_app)
uWSGI is a deployment option on servers like nginx, lighttpd, and cherokee; see
FastCGI and Standalone WSGI Containers for other options. To use your WSGI application with uWSGI protocol you will need a uWSGI server first. uWSGI is both a
protocol and an application server; the application server can serve uWSGI, FastCGI,
and HTTP protocols.
The most popular uWSGI server is uwsgi, which we will use for this guide. Make sure
to have it installed to follow along.
Watch Out
Please make sure in advance that any app.run() calls you might have in your application file are inside an if __name__ == '__main__': block or moved to a separate file.
Just make sure it’s not called because this will always start a local WSGI server which
we do not want if we deploy that application to uWSGI.
Starting your app with uwsgi
uwsgi is designed to operate on WSGI callables found in python modules.
Given a flask application in myapp.py, use the following command:
$ uwsgi -s /tmp/yourapplication.sock --manage-script-name --mount /
The --manage-script-name will move the handling of SCRIPT_NAME to uwsgi, since its
smarter about that. It is used together with the --mount directive which will make
requests to /yourapplication be directed to myapp:app. If your application is accessible at root level, you can use a single / instead of /yourapplication. myapp refers to
the name of the file of your flask application (without extension) or the module which
provides app. app is the callable inside of your application (usually the line reads app
= Flask(__name__).
If you want to deploy your flask application inside of a virtual environment, you need
to also add --virtualenv /path/to/virtual/environment. You might also need to
add --plugin python or --plugin python3 depending on which python version you
use for your project.
Configuring nginx
A basic flask nginx configuration looks like this:
location = /yourapplication { rewrite ^ /yourapplication/; }
location /yourapplication { try_files $uri @yourapplication; }
location @yourapplication {
include uwsgi_params;
uwsgi_pass unix:/tmp/yourapplication.sock;
This configuration binds the application to /yourapplication. If you want to have it
in the URL root its a bit simpler:
location / { try_files $uri @yourapplication; }
location @yourapplication {
include uwsgi_params;
uwsgi_pass unix:/tmp/yourapplication.sock;
mod_wsgi (Apache)
If you are using the Apache webserver, consider using mod_wsgi.
Watch Out
Please make sure in advance that any app.run() calls you might have in your application file are inside an if __name__ == '__main__': block or moved to a separate file.
Just make sure it’s not called because this will always start a local WSGI server which
we do not want if we deploy that application to mod_wsgi.
Installing mod_wsgi
If you don’t have mod_wsgi installed yet you have to either install it using a package
manager or compile it yourself. The mod_wsgi installation instructions cover source
installations on UNIX systems.
If you are using Ubuntu/Debian you can apt-get it and activate it as follows:
# apt-get install libapache2-mod-wsgi
If you are using a yum based distribution (Fedora, OpenSUSE, etc..) you can install it
as follows:
# yum install mod_wsgi
On FreeBSD install mod_wsgi by compiling the www/mod_wsgi port or by using
# pkg install ap22-mod_wsgi2
If you are using pkgsrc you can install mod_wsgi by compiling the www/ap2-wsgi package.
If you encounter segfaulting child processes after the first apache reload you can safely
ignore them. Just restart the server.
Creating a .wsgi file
To run your application you need a yourapplication.wsgi file. This file contains the
code mod_wsgi is executing on startup to get the application object. The object called
application in that file is then used as application.
For most applications the following file should be sufficient:
from yourapplication import app as application
If you don’t have a factory function for application creation but a singleton instance
you can directly import that one as application.
Store that file somewhere that you will find it again (e.g.: /var/www/yourapplication)
and make sure that yourapplication and all the libraries that are in use are on the python
load path. If you don’t want to install it system wide consider using a virtual python
instance. Keep in mind that you will have to actually install your application into the
virtualenv as well. Alternatively there is the option to just patch the path in the .wsgi
file before the import:
import sys
sys.path.insert(0, '/path/to/the/application')
Configuring Apache
The last thing you have to do is to create an Apache configuration file for your application. In this example we are telling mod_wsgi to execute the application under a
different user for security reasons:
<VirtualHost *>
ServerName example.com
WSGIDaemonProcess yourapplication user=user1 group=group1 threads=5
WSGIScriptAlias / /var/www/yourapplication/yourapplication.wsgi
<Directory /var/www/yourapplication>
WSGIProcessGroup yourapplication
WSGIApplicationGroup %{GLOBAL}
Order deny,allow
Allow from all
Note: WSGIDaemonProcess isn’t implemented in Windows and Apache will refuse to
run with the above configuration. On a Windows system, eliminate those lines:
<VirtualHost *>
ServerName example.com
WSGIScriptAlias / C:\yourdir\yourapp.wsgi
<Directory C:\yourdir>
Order deny,allow
Allow from all
Note: There have been some changes in access control configuration for Apache 2.4.
Most notably, the syntax for directory permissions has changed from httpd 2.2
Order allow,deny
Allow from all
to httpd 2.4 syntax
Require all granted
For more information consult the mod_wsgi documentation.
If your application does not run, follow this guide to troubleshoot:
Problem: application does not run, errorlog shows SystemExit ignored You have
an app.run() call in your application file that is not guarded by an if __name__
== '__main__': condition. Either remove that run() call from the file and move
it into a separate run.py file or put it into such an if block.
Problem: application gives permission errors Probably caused by your application
running as the wrong user. Make sure the folders the application needs access to
have the proper privileges set and the application runs as the correct user (user
and group parameter to the WSGIDaemonProcess directive)
Problem: application dies with an error on print Keep in mind that mod_wsgi disallows doing anything with sys.stdout and sys.stderr. You can disable this
protection from the config by setting the WSGIRestrictStdout to off:
WSGIRestrictStdout Off
Alternatively you can also replace the standard out in the .wsgi file with a different stream:
import sys
sys.stdout = sys.stderr
Problem: accessing resources gives IO errors Your application probably is a single
.py file you symlinked into the site-packages folder. Please be aware that this
does not work, instead you either have to put the folder into the pythonpath the
file is stored in, or convert your application into a package.
The reason for this is that for non-installed packages, the module filename is
used to locate the resources and for symlinks the wrong filename is picked up.
Support for Automatic Reloading
To help deployment tools you can activate support for automatic reloading. Whenever
something changes the .wsgi file, mod_wsgi will reload all the daemon processes for
For that, just add the following directive to your Directory section:
WSGIScriptReloading On
Working with Virtual Environments
Virtual environments have the advantage that they never install the required dependencies system wide so you have a better control over what is used where. If you
want to use a virtual environment with mod_wsgi you have to modify your .wsgi file
Add the following lines to the top of your .wsgi file:
activate_this = '/path/to/env/bin/activate_this.py'
execfile(activate_this, dict(__file__=activate_this))
For Python 3 add the following lines to the top of your .wsgi file:
activate_this = '/path/to/env/bin/activate_this.py'
with open(activate_this) as file_:
exec(file_.read(), dict(__file__=activate_this))
This sets up the load paths according to the settings of the virtual environment. Keep
in mind that the path has to be absolute.
FastCGI is a deployment option on servers like nginx, lighttpd, and cherokee; see
uWSGI and Standalone WSGI Containers for other options. To use your WSGI application with any of them you will need a FastCGI server first. The most popular one is
flup which we will use for this guide. Make sure to have it installed to follow along.
Watch Out
Please make sure in advance that any app.run() calls you might have in your application file are inside an if __name__ == '__main__': block or moved to a separate file.
Just make sure it’s not called because this will always start a local WSGI server which
we do not want if we deploy that application to FastCGI.
Creating a .fcgi file
First you need to create the FastCGI server file. Let’s call it yourapplication.fcgi:
from flup.server.fcgi import WSGIServer
from yourapplication import app
if __name__ == '__main__':
This is enough for Apache to work, however nginx and older versions of lighttpd need
a socket to be explicitly passed to communicate with the FastCGI server. For that to
work you need to pass the path to the socket to the WSGIServer:
WSGIServer(application, bindAddress='/path/to/fcgi.sock').run()
The path has to be the exact same path you define in the server config.
Save the yourapplication.fcgi file somewhere you will find it again. It makes sense
to have that in /var/www/yourapplication or something similar.
Make sure to set the executable bit on that file so that the servers can execute it:
# chmod +x /var/www/yourapplication/yourapplication.fcgi
Configuring Apache
The example above is good enough for a basic Apache deployment but your .fcgi file
will appear in your application URL e.g. example.com/yourapplication.fcgi/news/.
There are few ways to configure your application so that yourapplication.fcgi does not
appear in the URL. A preferable way is to use the ScriptAlias and SetHandler configuration directives to route requests to the FastCGI server. The following example uses
FastCgiServer to start 5 instances of the application which will handle all incoming
LoadModule fastcgi_module /usr/lib64/httpd/modules/mod_fastcgi.so
FastCgiServer /var/www/html/yourapplication/app.fcgi -idle-timeout 300 -processes 5
<VirtualHost *>
ServerName webapp1.mydomain.com
DocumentRoot /var/www/html/yourapplication
AddHandler fastcgi-script fcgi
ScriptAlias / /var/www/html/yourapplication/app.fcgi/
<Location />
SetHandler fastcgi-script
These processes will be managed by Apache. If you’re using a standalone FastCGI
server, you can use the FastCgiExternalServer directive instead. Note that in the following the path is not real, it’s simply used as an identifier to other directives such as
FastCgiServer /var/www/html/yourapplication -host
If you cannot set ScriptAlias, for example on a shared web host, you can use WSGI
middleware to remove yourapplication.fcgi from the URLs. Set .htaccess:
<IfModule mod_fcgid.c>
AddHandler fcgid-script .fcgi
<Files ~ (\.fcgi)>
SetHandler fcgid-script
Options +FollowSymLinks +ExecCGI
<IfModule mod_rewrite.c>
Options +FollowSymlinks
RewriteEngine On
RewriteBase /
RewriteCond %{REQUEST_FILENAME} !-f
RewriteRule ^(.*)$ yourapplication.fcgi/$1 [QSA,L]
Set yourapplication.fcgi:
#: optional path to your local python site-packages folder
import sys
sys.path.insert(0, '<your_local_path>/lib/python2.6/site-packages')
from flup.server.fcgi import WSGIServer
from yourapplication import app
class ScriptNameStripper(object):
def __init__(self, app):
self.app = app
def __call__(self, environ, start_response):
environ['SCRIPT_NAME'] = ''
return self.app(environ, start_response)
app = ScriptNameStripper(app)
if __name__ == '__main__':
Configuring lighttpd
A basic FastCGI configuration for lighttpd looks like that:
fastcgi.server = ("/yourapplication.fcgi" =>
"socket" => "/tmp/yourapplication-fcgi.sock",
"bin-path" => "/var/www/yourapplication/yourapplication.fcgi",
"check-local" => "disable",
"max-procs" => 1
alias.url = (
"/static/" => "/path/to/your/static/"
url.rewrite-once = (
"^(/static($|/.*))$" => "$1",
"^(/.*)$" => "/yourapplication.fcgi$1"
Remember to enable the FastCGI, alias and rewrite modules. This configuration binds
the application to /yourapplication. If you want the application to work in the URL
root you have to work around a lighttpd bug with the LighttpdCGIRootFix middleware.
Make sure to apply it only if you are mounting the application the URL root. Also,
see the Lighty docs for more information on FastCGI and Python (note that explicitly
passing a socket to run() is no longer necessary).
Configuring nginx
Installing FastCGI applications on nginx is a bit different because by default no
FastCGI parameters are forwarded.
A basic Flask FastCGI configuration for nginx looks like this:
location = /yourapplication { rewrite ^ /yourapplication/ last; }
location /yourapplication { try_files $uri @yourapplication; }
location @yourapplication {
include fastcgi_params;
fastcgi_split_path_info ^(/yourapplication)(.*)$;
fastcgi_param PATH_INFO $fastcgi_path_info;
fastcgi_param SCRIPT_NAME $fastcgi_script_name;
fastcgi_pass unix:/tmp/yourapplication-fcgi.sock;
This configuration binds the application to /yourapplication. If you want to have it
in the URL root it’s a bit simpler because you don’t have to figure out how to calculate
location / { try_files $uri @yourapplication; }
location @yourapplication {
include fastcgi_params;
fastcgi_param PATH_INFO $fastcgi_script_name;
fastcgi_param SCRIPT_NAME "";
fastcgi_pass unix:/tmp/yourapplication-fcgi.sock;
Running FastCGI Processes
Since nginx and others do not load FastCGI apps, you have to do it by yourself. Supervisor can manage FastCGI processes. You can look around for other FastCGI process
managers or write a script to run your .fcgi file at boot, e.g. using a SysV init.d script.
For a temporary solution, you can always run the .fcgi script inside GNU screen. See
man screen for details, and note that this is a manual solution which does not persist
across system restart:
$ screen
$ /var/www/yourapplication/yourapplication.fcgi
FastCGI deployments tend to be hard to debug on most web servers. Very often the
only thing the server log tells you is something along the lines of “premature end of
headers”. In order to debug the application the only thing that can really give you
ideas why it breaks is switching to the correct user and executing the application by
This example assumes your application is called application.fcgi and that your web
server user is www-data:
$ su www-data
$ cd /var/www/yourapplication
$ python application.fcgi
Traceback (most recent call last):
File "yourapplication.fcgi", line 4, in <module>
ImportError: No module named yourapplication
In this case the error seems to be “yourapplication” not being on the python path.
Common problems are:
• Relative paths being used. Don’t rely on the current working directory.
• The code depending on environment variables that are not set by the web server.
• Different python interpreters being used.
If all other deployment methods do not work, CGI will work for sure. CGI is supported by all major servers but usually has a sub-optimal performance.
This is also the way you can use a Flask application on Google’s App Engine, where
execution happens in a CGI-like environment.
Watch Out
Please make sure in advance that any app.run() calls you might have in your application file are inside an if __name__ == '__main__': block or moved to a separate file.
Just make sure it’s not called because this will always start a local WSGI server which
we do not want if we deploy that application to CGI / app engine.
With CGI, you will also have to make sure that your code does not contain any print
statements, or that sys.stdout is overridden by something that doesn’t write into the
HTTP response.
Creating a .cgi file
First you need to create the CGI application file. Let’s call it yourapplication.cgi:
from wsgiref.handlers import CGIHandler
from yourapplication import app
Server Setup
Usually there are two ways to configure the server. Either just copy the .cgi into
a cgi-bin (and use mod_rewrite or something similar to rewrite the URL) or let the
server point to the file directly.
In Apache for example you can put something like this into the config:
ScriptAlias /app /path/to/the/application.cgi
On shared webhosting, though, you might not have access to your Apache config. In
this case, a file called .htaccess, sitting in the public directory you want your app to
be available, works too but the ScriptAlias directive won’t work in that case:
RewriteEngine On
RewriteCond %{REQUEST_FILENAME} !-f # Don't interfere with static files
RewriteRule ^(.*)$ /path/to/the/application.cgi/$1 [L]
For more information consult the documentation of your webserver.
Becoming Big
Here are your options when growing your codebase or scaling your application.
Read the Source.
Flask started in part to demonstrate how to build your own framework on top of
existing well-used tools Werkzeug (WSGI) and Jinja (templating), and as it developed,
it became useful to a wide audience. As you grow your codebase, don’t just use Flask
– understand it. Read the source. Flask’s code is written to be read; its documentation
is published so you can use its internal APIs. Flask sticks to documented APIs in
upstream libraries, and documents its internal utilities so that you can find the hook
points needed for your project.
Hook. Extend.
The API docs are full of available overrides, hook points, and Signals. You can provide
custom classes for things like the request and response objects. Dig deeper on the
APIs you use, and look for the customizations which are available out of the box in a
Flask release. Look for ways in which your project can be refactored into a collection
of utilities and Flask extensions. Explore the many extensions in the community, and
look for patterns to build your own extensions if you do not find the tools you need.
The Flask class has many methods designed for subclassing. You can quickly add or
customize behavior by subclassing Flask (see the linked method docs) and using that
subclass wherever you instantiate an application class. This works well with Application Factories. See Subclassing Flask for an example.
Wrap with middleware.
The Application Dispatching chapter shows in detail how to apply middleware. You
can introduce WSGI middleware to wrap your Flask instances and introduce fixes and
changes at the layer between your Flask application and your HTTP server. Werkzeug
includes several middlewares.
If none of the above options work, fork Flask. The majority of code of Flask is within
Werkzeug and Jinja2. These libraries do the majority of the work. Flask is just the paste
that glues those together. For every project there is the point where the underlying
framework gets in the way (due to assumptions the original developers had). This is
natural because if this would not be the case, the framework would be a very complex
system to begin with which causes a steep learning curve and a lot of user frustration.
This is not unique to Flask. Many people use patched and modified versions of their
framework to counter shortcomings. This idea is also reflected in the license of Flask.
You don’t have to contribute any changes back if you decide to modify the framework.
The downside of forking is of course that Flask extensions will most likely break because the new framework has a different import name. Furthermore integrating upstream changes can be a complex process, depending on the number of changes. Because of that, forking should be the very last resort.
Scale like a pro.
For many web applications the complexity of the code is less an issue than the scaling
for the number of users or data entries expected. Flask by itself is only limited in terms
of scaling by your application code, the data store you want to use and the Python
implementation and webserver you are running on.
Scaling well means for example that if you double the amount of servers you get about
twice the performance. Scaling bad means that if you add a new server the application
won’t perform any better or would not even support a second server.
There is only one limiting factor regarding scaling in Flask which are the context local
proxies. They depend on context which in Flask is defined as being either a thread,
process or greenlet. If your server uses some kind of concurrency that is not based
on threads or greenlets, Flask will no longer be able to support these global proxies.
However the majority of servers are using either threads, greenlets or separate processes to achieve concurrency which are all methods well supported by the underlying
Werkzeug library.
Discuss with the community.
The Flask developers keep the framework accessible to users with codebases big and
small. If you find an obstacle in your way, caused by Flask, don’t hesitate to contact
the developers on the mailinglist or IRC channel. The best way for the Flask and Flask
extension developers to improve the tools for larger applications is getting feedback
from users.
Part II
If you are looking for information on a specific function, class or method, this part of
the documentation is for you.
This part of the documentation covers all the interfaces of Flask. For parts where Flask
depends on external libraries, we document the most important right here and provide
links to the canonical documentation.
Application Object
class flask.Flask(import_name,
static_folder=’static’, static_host=None, host_matching=False,
instance_relative_config=False, root_path=None)
The flask object implements a WSGI application and acts as the central object.
It is passed the name of the module or package of the application. Once it is
created it will act as a central registry for the view functions, the URL rules,
template configuration and much more.
The name of the package is used to resolve resources from inside the package
or the folder the module is contained in depending on if the package parameter
resolves to an actual python package (a folder with an __init__.py file inside)
or a standard module (just a .py file).
For more information about resource loading, see open_resource().
Usually you create a Flask instance in your main module or in the __init__.py
file of your package like this:
from flask import Flask
app = Flask(__name__)
About the First Parameter
The idea of the first parameter is to give Flask an idea of what belongs to your
application. This name is used to find resources on the filesystem, can be used
by extensions to improve debugging information and a lot more.
So it’s important what you provide there. If you are using a single module,
__name__ is always the correct value. If you however are using a package, it’s
usually recommended to hardcode the name of your package there.
For example if your application is defined in yourapplication/app.py you
should create it with one of the two versions below:
app = Flask('yourapplication')
app = Flask(__name__.split('.')[0])
Why is that? The application will work even with __name__, thanks to how resources are looked up. However it will make debugging more painful. Certain
extensions can make assumptions based on the import name of your application.
For example the Flask-SQLAlchemy extension will look for the code in your application that triggered an SQL query in debug mode. If the import name is not
properly set up, that debugging information is lost. (For example it would only
pick up SQL queries in yourapplication.app and not yourapplication.views.frontend)
New in version 0.7: The static_url_path, static_folder, and template_folder parameters were added.
New in version 0.8: The instance_path and instance_relative_config parameters
were added.
New in version 0.11: The root_path parameter was added.
New in version 0.13: The host_matching and static_host parameters were added.
• import_name – the name of the application package
• static_url_path – can be used to specify a different path
for the static files on the web. Defaults to the name of the
static_folder folder.
• static_folder – the folder with static files that should be
served at static_url_path. Defaults to the 'static' folder in the
root path of the application.
• host_matching – sets the app’s url_map.host_matching to the
given given value. Defaults to False.
• static_host – the host to use when adding the static route.
Defaults to None. Required when using host_matching=True
with a static_folder configured.
• template_folder – the folder that contains the templates that
should be used by the application. Defaults to 'templates'
folder in the root path of the application.
• instance_path – An alternative instance path for the application. By default the folder 'instance' next to the package or
module is assumed to be the instance path.
• instance_relative_config – if set to True relative filenames
for loading the config are assumed to be relative to the instance
path instead of the application root.
• root_path – Flask by default will automatically calculate the
path to the root of the application. In certain situations this
cannot be achieved (for instance if the package is a Python 3
namespace package) and needs to be manually defined.
add_template_filter(f, name=None)
Register a custom template filter. Works exactly like the template_filter()
Parameters name – the optional name of the filter, otherwise the
function name will be used.
add_template_global(f, name=None)
Register a custom template global function.
template_global() decorator.
Works exactly like the
New in version 0.10.
Parameters name – the optional name of the global function, otherwise the function name will be used.
add_template_test(f, name=None)
Register a custom template test. Works exactly like the template_test()
New in version 0.10.
Parameters name – the optional name of the test, otherwise the
function name will be used.
provide_automatic_options=None, **options)
Connects a URL rule. Works exactly like the route() decorator. If a
view_func is provided it will be registered with the endpoint.
Basically this example:
def index():
Is equivalent to the following:
def index():
app.add_url_rule('/', 'index', index)
If the view_func is not provided you will need to connect the endpoint to a
view function like so:
app.view_functions['index'] = index
Internally route() invokes add_url_rule() so if you want to customize the
behavior via subclassing you only need to change this method.
For more information refer to URL Route Registrations.
Changed in version 0.2: view_func parameter added.
Changed in version 0.6: OPTIONS is added automatically as method.
• rule – the URL rule as string
• endpoint – the endpoint for the registered URL rule. Flask
itself assumes the name of the view function as endpoint
• view_func – the function to call when serving a request to the
provided endpoint
• provide_automatic_options – controls whether the OPTIONS
method should be added automatically. This can also be controlled by setting the view_func.provide_automatic_options
= False before adding the rule.
• options – the options to be forwarded to the underlying Rule
object. A change to Werkzeug is handling of method options.
methods is a list of methods this rule should be limited to
(GET, POST etc.). By default a rule just listens for GET (and
implicitly HEAD). Starting with Flask 0.6, OPTIONS is implicitly
added and handled by the standard request handling.
after_request(f )
Register a function to be run after each request.
Your function must take one parameter, an instance of response_class and
return a new response object or the same (see process_response()).
As of Flask 0.7 this function might not be executed at the end of the request
in case an unhandled exception occurred.
after_request_funcs = None
A dictionary with lists of functions that should be called after each request.
The key of the dictionary is the name of the blueprint this function is active
for, None for all requests. This can for example be used to close database
connections. To register a function here, use the after_request() decorator.
Binds the application only. For as long as the application is bound to the
current context the flask.current_app points to that application. An application context is automatically created when a request context is pushed if
Example usage:
with app.app_context():
New in version 0.9.
The class that is used for the g instance.
Example use cases for a custom class:
1.Store arbitrary attributes on flask.g.
2.Add a property for lazy per-request database connectors.
3.Return None instead of AttributeError on unexpected attributes.
4.Raise exception if an unexpected attr is set, a “controlled” flask.g.
In Flask 0.9 this property was called request_globals_class but it was changed
in 0.10 to app_ctx_globals_class because the flask.g object is now application context scoped.
New in version 0.10.
alias of _AppCtxGlobals
Tries to locate the instance path if it was not provided to the constructor of
the application class. It will basically calculate the path to a folder named
instance next to your main file or the package.
New in version 0.8.
before_first_request(f )
Registers a function to be run before the first request to this instance of the
The function will be called without any arguments and its return value is
New in version 0.8.
before_first_request_funcs = None
A list of functions that will be called at the beginning of the first request to
this instance. To register a function, use the before_first_request() decorator.
New in version 0.8.
before_request(f )
Registers a function to run before each request.
For example, this can be used to open a database connection, or to load the
logged in user from the session.
The function will be called without any arguments. If it returns a non-None
value, the value is handled as if it was the return value from the view, and
further request handling is stopped.
before_request_funcs = None
A dictionary with lists of functions that will be called at the beginning of
each request. The key of the dictionary is the name of the blueprint this
function is active for, or None for all requests. To register a function, use the
before_request() decorator.
blueprints = None
all the attached blueprints in a dictionary by name. Blueprints can be attached multiple times so this dictionary does not tell you how often they
got attached.
New in version 0.7.
cli = None
The click command line context for this application. Commands registered
here show up in the flask command once the application has been discovered. The default commands are provided by Flask itself and can be
This is an instance of a click.Group object.
config = None
The configuration dictionary as Config. This behaves exactly like a regular
dictionary but supports additional methods to load a config from files.
The class that is used for the config attribute of this app. Defaults to Config.
Example use cases for a custom class:
1.Default values for certain config options.
2.Access to config values through attributes in addition to keys.
New in version 0.11.
alias of Config
context_processor(f )
Registers a template context processor function.
Creates the loader for the Jinja2 environment. Can be used to override just
the loader and keeping the rest unchanged. It’s discouraged to override this
function. Instead one should override the jinja_loader() function instead.
The global loader dispatches between the loaders of the application and the
individual blueprints.
New in version 0.7.
Creates the Jinja2 environment based on jinja_options and
select_jinja_autoescape(). Since 0.7 this also adds the Jinja2 globals
and filters after initialization. Override this function to customize the
New in version 0.5.
Changed in version 0.11: Environment.auto_reload set in accordance with
TEMPLATES_AUTO_RELOAD configuration option.
Creates a URL adapter for the given request. The URL adapter is created at
a point where the request context is not yet set up so the request is passed
New in version 0.6.
Changed in version 0.9: This can now also be called without a request object
when the URL adapter is created for the application context.
The debug flag. Set this to True to enable debugging of the application.
In debug mode the debugger will kick in when an unhandled exception
occurs and the integrated server will automatically reload the application if
changes in the code are detected.
This attribute can also be configured from the config with the DEBUG configuration key. Defaults to False.
default_config = ImmutableDict({‘JSON_AS_ASCII’: True, ‘USE_X_SENDFILE’: False, ‘SE
Default configuration parameters.
Does the request dispatching. Matches the URL and returns the return
value of the view or error handler. This does not have to be a response
object. In order to convert the return value to a proper response object, call
Changed in version 0.7: This no longer does the exception handling, this
code was moved to the new full_dispatch_request().
do_teardown_appcontext(exc=<object object>)
Called when an application context is popped. This works pretty much the
same as do_teardown_request() but for the application context.
New in version 0.9.
do_teardown_request(exc=<object object>)
Called after the actual request dispatching and will call every as
teardown_request() decorated function. This is not actually called by
the Flask object itself but is always triggered when the request context is
popped. That way we have a tighter control over certain resources under
testing environments.
Changed in version 0.9: Added the exc argument. Previously this was always using the current exception information.
A decorator to register a function as an endpoint. Example:
def example():
return "example"
Parameters endpoint – the name of the endpoint
error_handler_spec = None
A dictionary of all registered error handlers. The key is None for error
handlers active on the application, otherwise the key is the name of the
blueprint. Each key points to another dictionary where the key is the status
code of the http exception. The special key None points to a list of tuples
where the first item is the class for the instance check and the second the
error handler function.
To register an error handler, use the errorhandler() decorator.
A decorator that is used to register a function given an error code. Example:
def page_not_found(error):
return 'This page does not exist', 404
You can also register handlers for arbitrary exceptions:
def special_exception_handler(error):
return 'Database connection failed', 500
You can also register a function as error handler without using the
errorhandler() decorator. The following example is equivalent to the one
def page_not_found(error):
return 'This page does not exist', 404
app.error_handler_spec[None][404] = page_not_found
Setting error handlers via assignments to error_handler_spec however is
discouraged as it requires fiddling with nested dictionaries and the special
case for arbitrary exception types.
The first None refers to the active blueprint. If the error handler should be
application wide None shall be used.
New in version 0.7: Use register_error_handler() instead of modifying
error_handler_spec directly, for application wide error handlers.
New in version 0.7: One can now additionally also register custom exception types that do not necessarily have to be a subclass of the HTTPException
Parameters code_or_exception – the code as integer for the handler, or an arbitrary exception
extensions = None
a place where extensions can store application specific state. For example this is where an extension could store database engines and similar
things. For backwards compatibility extensions should register themselves
like this:
if not hasattr(app, 'extensions'):
app.extensions = {}
app.extensions['extensionname'] = SomeObject()
The key must match the name of the extension module. For example in case
of a “Flask-Foo” extension in flask_foo, the key would be 'foo'.
New in version 0.7.
Dispatches the request and on top of that performs request pre and postprocessing as well as HTTP exception catching and error handling.
New in version 0.7.
Provides default cache_timeout for the send_file() functions.
By default, this function returns SEND_FILE_MAX_AGE_DEFAULT from the configuration of current_app.
Static file functions such as send_from_directory() use this function,
and send_file() calls this function on current_app when the given
cache_timeout is None. If a cache_timeout is given in send_file(), that timeout is used; otherwise, this method is called.
This allows subclasses to change the behavior when sending files based on
the filename. For example, to set the cache timeout for .js files to 60 seconds:
class MyFlask(flask.Flask):
def get_send_file_max_age(self, name):
if name.lower().endswith('.js'):
return 60
return flask.Flask.get_send_file_max_age(self, name)
New in version 0.9.
This attribute is set to True if the application started handling the first request.
New in version 0.8.
Default exception handling that kicks in when an exception occurs that is
not caught. In debug mode the exception will be re-raised immediately, otherwise it is logged and the handler for a 500 internal server error is used. If
no such handler exists, a default 500 internal server error message is displayed.
New in version 0.3.
Handles an HTTP exception. By default this will invoke the registered error
handlers and fall back to returning the exception as response.
New in version 0.3.
handle_url_build_error(error, endpoint, values)
Handle BuildError on url_for().
This method is called whenever an exception occurs that should be handled.
A special case are HTTPExceptions which are forwarded by this function to
the handle_http_exception() method. This function will either return a
response value or reraise the exception with the same traceback.
New in version 0.7.
This is True if the package bound object’s container has a folder for static
New in version 0.5.
Deprecated. Used to initialize the Jinja2 globals.
New in version 0.5.
Changed in version 0.7: This method is deprecated with 0.7. Override
create_jinja_environment() instead.
inject_url_defaults(endpoint, values)
Injects the URL defaults for the given endpoint directly into the values dictionary passed. This is used internally and automatically called on URL
New in version 0.7.
instance_path = None
Holds the path to the instance folder.
New in version 0.8.
Iterates over all blueprints by the order they were registered.
New in version 0.11.
The Jinja2 environment used to load templates.
The class that is used for the Jinja environment.
New in version 0.11.
alias of Environment
The Jinja loader for this package bound object.
New in version 0.5.
jinja_options = ImmutableDict({‘extensions’: [’jinja2.ext.autoescape’, ‘jinja2.ext.with_’]})
Options that are passed directly to the Jinja2 environment.
The JSON decoder class to use. Defaults to JSONDecoder.
New in version 0.10.
alias of JSONDecoder
The JSON encoder class to use. Defaults to JSONEncoder.
New in version 0.10.
alias of JSONEncoder
Logs an exception. This is called by handle_exception() if debugging is
disabled and right before the handler is called. The default implementation
logs the exception as error on the logger.
New in version 0.8.
A logging.Logger object for this application. The default configuration is to
log to stderr if the application is in debug mode. This logger can be used to
(surprise) log messages. Here some examples:
app.logger.debug('A value for debugging')
app.logger.warning('A warning occurred (%d apples)', 42)
app.logger.error('An error occurred')
New in version 0.3.
The name of the logger to use. By default the logger name is the package
name passed to the constructor.
New in version 0.4.
Used to create the config attribute by the Flask constructor. The instance_relative parameter is passed in from the constructor of Flask (there
named instance_relative_config) and indicates if the config should be relative
to the instance path or the root path of the application.
New in version 0.8.
This method is called to create the default OPTIONS response. This can be
changed through subclassing to change the default behavior of OPTIONS responses.
New in version 0.7.
Creates a new instance of a missing session. Instead of overriding this
method we recommend replacing the session_interface.
New in version 0.7.
Convert the return value from a view function to an instance of
Parameters rv – the return value from the view function. The view
function must return a response. Returning None, or the view
ending without returning, is not allowed. The following types
are allowed for view_rv:
str (unicode in Python 2) A response object is created with the
string encoded to UTF-8 as the body.
bytes (str in Python 2) A response object is created with the
bytes as the body.
tuple Either (body, status, headers), (body, status), or
(body, headers), where body is any of the other types allowed here, status is a string or an integer, and headers is
a dictionary or a list of (key, value) tuples. If body is a
response_class instance, status overwrites the exiting value
and headers are extended.
response_class The object is returned unchanged.
other Response class The object is coerced to response_class.
callable() The function is called as a WSGI application. The
result is used to create a response object.
Changed in version 0.9: Previously a tuple was interpreted as the arguments
for the response object.
Returns the shell context for an interactive shell for this application. This
runs all the registered shell context processors.
New in version 0.11.
The name of the application. This is usually the import name with the difference that it’s guessed from the run file if the import name is main. This
name is used as a display name when Flask needs the name of the application. It can be set and overridden to change the value.
New in version 0.8.
open_instance_resource(resource, mode=’rb’)
Opens a resource from the application’s instance folder (instance_path).
Otherwise works like open_resource(). Instance resources can also be
opened for writing.
• resource – the name of the resource. To access resources
within subfolders use forward slashes as separator.
• mode – resource file opening mode, default is ‘rb’.
open_resource(resource, mode=’rb’)
Opens a resource from the application’s resource folder. To see how this
works, consider the following folder structure:
If you want to open the schema.sql file you would do the following:
with app.open_resource('schema.sql') as f:
contents = f.read()
• resource – the name of the resource. To access resources
within subfolders use forward slashes as separator.
• mode – resource file opening mode, default is ‘rb’.
Creates or opens a new session. Default implementation stores all session
data in a signed cookie. This requires that the secret_key is set. Instead of
overriding this method we recommend replacing the session_interface.
Parameters request – an instance of request_class.
A timedelta which is used to set the expiration date of a permanent session.
The default is 31 days which makes a permanent session survive for roughly
one month.
This attribute can also be configured from the config with the
Defaults to
Called before the request is dispatched. Calls url_value_preprocessors
registered with the app and the current blueprint (if any). Then calls
before_request_funcs registered with the app and the blueprint.
If any before_request() handler returns a non-None value, the value is
handled as if it was the return value from the view, and further request
handling is stopped.
Returns the value of the PRESERVE_CONTEXT_ON_EXCEPTION configuration
value in case it’s set, otherwise a sensible default is returned.
New in version 0.7.
Can be overridden in order to modify the response object before it’s sent to
the WSGI server. By default this will call all the after_request() decorated
Changed in version 0.5: As of Flask 0.5 the functions registered for after
request execution are called in reverse order of registration.
Parameters response – a response_class object.
Returns a new response object or the same, has to be an instance of
Returns the value of the PROPAGATE_EXCEPTIONS configuration value in case
it’s set, otherwise a sensible default is returned.
New in version 0.7.
register_blueprint(blueprint, **options)
Registers a blueprint on the application.
New in version 0.7.
register_error_handler(code_or_exception, f )
Alternative error attach function to the errorhandler() decorator that is
more straightforward to use for non decorator usage.
New in version 0.7.
The class that is used for request objects. See Request for more information.
alias of Request
Creates a RequestContext from the given environment and binds it to the
current context. This must be used in combination with the with statement
because the request is only bound to the current context for the duration of
the with block.
Example usage:
with app.request_context(environ):
The object returned can also be used without the with statement which is
useful for working in the shell. The example above is doing exactly the
same as this code:
ctx = app.request_context(environ)
Changed in version 0.3: Added support for non-with statement usage and
with statement is now passed the ctx object.
Parameters environ – a WSGI environment
The class that is used for response objects. See Response for more information.
alias of Response
route(rule, **options)
A decorator that is used to register a view function for a given URL rule.
This does the same thing as add_url_rule() but is intended for decorator
def index():
return 'Hello World'
For more information refer to URL Route Registrations.
• rule – the URL rule as string
• endpoint – the endpoint for the registered URL rule. Flask
itself assumes the name of the view function as endpoint
• options – the options to be forwarded to the underlying Rule
object. A change to Werkzeug is handling of method options.
methods is a list of methods this rule should be limited to
(GET, POST etc.). By default a rule just listens for GET (and
implicitly HEAD). Starting with Flask 0.6, OPTIONS is implicitly
added and handled by the standard request handling.
run(host=None, port=None, debug=None, **options)
Runs the application on a local development server.
Do not use run() in a production setting. It is not intended to meet security
and performance requirements for a production server. Instead, see Deployment Options for WSGI server recommendations.
If the debug flag is set the server will automatically reload for code changes
and show a debugger in case an exception happened.
If you want to run the application in debug mode, but disable the code execution on the interactive debugger, you can pass use_evalex=False as parameter. This will keep the debugger’s traceback screen active, but disable
code execution.
It is not recommended to use this function for development with automatic
reloading as this is badly supported. Instead you should be using the flask
command line script’s run support.
Keep in Mind
Flask will suppress any server error with a generic error page unless it
is in debug mode. As such to enable just the interactive debugger without the code reloading, you have to invoke run() with debug=True and
use_reloader=False. Setting use_debugger to True without being in debug
mode won’t catch any exceptions because there won’t be any to catch.
Changed in version 0.10:
SERVER_NAME variable.
The default port is now picked from the
• host – the hostname to listen on. Set this to '' to have
the server available externally as well. Defaults to '127.0.0.
1' or the host in the SERVER_NAME config variable if present.
• port – the port of the webserver. Defaults to 5000 or the port
defined in the SERVER_NAME config variable if present.
• debug – if given, enable or disable debug mode. See debug.
• options – the options to be forwarded to the underlying
Werkzeug server. See werkzeug.serving.run_simple() for
more information.
save_session(session, response)
Saves the session if it needs updates. For the default implementation, check
open_session(). Instead of overriding this method we recommend replacing the session_interface.
• session – the session to be saved (a SecureCookie object)
• response – an instance of response_class
If a secret key is set, cryptographic components can use this to sign cookies
and other things. Set this to a complex random value when you want to use
the secure cookie for instance.
This attribute can also be configured from the config with the SECRET_KEY
configuration key. Defaults to None.
Returns True if autoescaping should be active for the given template name.
If no template name is given, returns True.
New in version 0.5.
A timedelta which is used as default cache_timeout for the send_file()
functions. The default is 12 hours.
This attribute can also be configured from the config with the
SEND_FILE_MAX_AGE_DEFAULT configuration key. This configuration variable can also be set with an integer value used as seconds. Defaults to
Function used internally to send static files from the static folder to the
New in version 0.5.
The secure cookie uses this for the name of the session cookie.
This attribute can also be configured from the config with the
SESSION_COOKIE_NAME configuration key. Defaults to 'session'
session_interface = <flask.sessions.SecureCookieSessionInterface object>
the session interface to use.
By default an instance of
SecureCookieSessionInterface is used here.
New in version 0.8.
shell_context_processor(f )
Registers a shell context processor function.
New in version 0.11.
shell_context_processors = None
A list of shell context processor functions that should be run when a shell
context is created.
New in version 0.11.
This is called to figure out if an error should be ignored or not as far as
the teardown system is concerned. If this function returns True then the
teardown handlers will not be passed the error.
New in version 0.10.
The absolute path to the configured static folder.
teardown_appcontext(f )
Registers a function to be called when the application context ends. These
functions are typically also called when the request context is popped.
ctx = app.app_context()
When ctx.pop() is executed in the above example, the teardown functions
are called just before the app context moves from the stack of active contexts. This becomes relevant if you are using such constructs in tests.
Since a request context typically also manages an application context it
would also be called when you pop a request context.
When a teardown function was called because of an exception it will be
passed an error object.
The return values of teardown functions are ignored.
New in version 0.9.
teardown_appcontext_funcs = None
A list of functions that are called when the application context is destroyed.
Since the application context is also torn down if the request ends this is the
place to store code that disconnects from databases.
New in version 0.9.
teardown_request(f )
Register a function to be run at the end of each request, regardless of
whether there was an exception or not. These functions are executed when
the request context is popped, even if not an actual request was performed.
ctx = app.test_request_context()
When ctx.pop() is executed in the above example, the teardown functions
are called just before the request context moves from the stack of active contexts. This becomes relevant if you are using such constructs in tests.
Generally teardown functions must take every necessary step to avoid that
they will fail. If they do execute code that might fail they will have to surround the execution of these code by try/except statements and log occurring errors.
When a teardown function was called because of an exception it will be
passed an error object.
The return values of teardown functions are ignored.
Debug Note
In debug mode Flask will not tear down a request on an exception
immediately. Instead it will keep it alive so that the interactive debugger can still access it. This behavior can be controlled by the
PRESERVE_CONTEXT_ON_EXCEPTION configuration variable.
teardown_request_funcs = None
A dictionary with lists of functions that are called after each request, even
if an exception has occurred. The key of the dictionary is the name of
the blueprint this function is active for, None for all requests. These functions are not allowed to modify the request, and their return values are ignored. If an exception occurred while processing the request, it gets passed
to each teardown_request function. To register a function here, use the
teardown_request() decorator.
New in version 0.7.
template_context_processors = None
A dictionary with list of functions that are called without argument to populate the template context. The key of the dictionary is the name of the
blueprint this function is active for, None for all requests. Each returns a
dictionary that the template context is updated with. To register a function
here, use the context_processor() decorator.
A decorator that is used to register custom template filter. You can specify a
name for the filter, otherwise the function name will be used. Example:
def reverse(s):
return s[::-1]
Parameters name – the optional name of the filter, otherwise the
function name will be used.
A decorator that is used to register a custom template global function. You
can specify a name for the global function, otherwise the function name will
be used. Example:
def double(n):
return 2 * n
New in version 0.10.
Parameters name – the optional name of the global function, otherwise the function name will be used.
A decorator that is used to register custom template test. You can specify a
name for the test, otherwise the function name will be used. Example:
def is_prime(n):
if n == 2:
return True
for i in range(2, int(math.ceil(math.sqrt(n))) + 1):
if n % i == 0:
return False
return True
New in version 0.10.
Parameters name – the optional name of the test, otherwise the
function name will be used.
test_client(use_cookies=True, **kwargs)
Creates a test client for this application. For information about unit testing
head over to Testing Flask Applications.
Note that if you are testing for assertions or exceptions in your application
code, you must set app.testing = True in order for the exceptions to propagate to the test client. Otherwise, the exception will be handled by the
application (not visible to the test client) and the only indication of an AssertionError or other exception will be a 500 status code response to the test
client. See the testing attribute. For example:
app.testing = True
client = app.test_client()
The test client can be used in a with block to defer the closing down of the
context until the end of the with block. This is useful if you want to access
the context locals for testing:
with app.test_client() as c:
rv = c.get('/?vodka=42')
assert request.args['vodka'] == '42'
Additionally, you may pass optional keyword arguments that will then be
passed to the application’s test_client_class constructor. For example:
from flask.testing import FlaskClient
class CustomClient(FlaskClient):
def __init__(self, *args, **kwargs):
self._authentication = kwargs.pop("authentication")
super(CustomClient,self).__init__( *args, **kwargs)
app.test_client_class = CustomClient
client = app.test_client(authentication='Basic ....')
See FlaskClient for more information.
Changed in version 0.4: added support for with block usage for the client.
New in version 0.7: The use_cookies parameter was added as well as the
ability to override the client to be used by setting the test_client_class
Changed in version 0.11: Added **kwargs to support passing additional keyword arguments to the constructor of test_client_class.
test_client_class = None
the test client that is used with when test_client is used.
New in version 0.7.
test_request_context(*args, **kwargs)
Creates a WSGI environment from the given values (see werkzeug.test.
EnvironBuilder for more information, this function accepts the same arguments).
The testing flag. Set this to True to enable the test mode of Flask extensions
(and in the future probably also Flask itself). For example this might activate test helpers that have an additional runtime cost which should not be
enabled by default.
If this is enabled and PROPAGATE_EXCEPTIONS is not changed from the
default it’s implicitly enabled.
This attribute can also be configured from the config with the TESTING configuration key. Defaults to False.
Checks if an HTTP exception should be trapped or not. By default
this will return False for all exceptions except for a bad request key error if TRAP_BAD_REQUEST_ERRORS is set to True. It also returns True if
This is called for all HTTP exceptions raised by a view function. If it returns
True for any exception the error handler for this exception is not called and
it shows up as regular exception in the traceback. This is helpful for debugging implicitly raised HTTP exceptions.
New in version 0.8.
Update the template context with some commonly used variables. This injects request, session, config and g into the template context as well as everything template context processors want to inject. Note that the as of Flask
0.6, the original values in the context will not be overridden if a context
processor decides to return a value with the same key.
Parameters context – the context as a dictionary that is updated in
place to add extra variables.
url_build_error_handlers = None
A list of functions that are called when url_for() raises a BuildError. Each
function registered here is called with error, endpoint and values. If a function
returns None or raises a BuildError the next function is tried.
New in version 0.9.
url_default_functions = None
A dictionary with lists of functions that can be used as URL value preprocessors. The key None here is used for application wide callbacks, otherwise
the key is the name of the blueprint. Each of these functions has the chance
to modify the dictionary of URL values before they are used as the keyword arguments of the view function. For each function registered this one
should also provide a url_defaults() function that adds the parameters
automatically again that were removed that way.
New in version 0.7.
url_defaults(f )
Callback function for URL defaults for all view functions of the application. It’s called with the endpoint and values and should update the values
passed in place.
url_map = None
The Map for this instance. You can use this to change the routing converters
after the class was created but before any routes are connected. Example:
from werkzeug.routing import BaseConverter
class ListConverter(BaseConverter):
def to_python(self, value):
return value.split(',')
def to_url(self, values):
return ','.join(super(ListConverter, self).to_url(value)
for value in values)
app = Flask(__name__)
app.url_map.converters['list'] = ListConverter
The rule object to use for URL rules created. This is used by add_url_rule().
Defaults to werkzeug.routing.Rule.
New in version 0.7.
alias of Rule
url_value_preprocessor(f )
Register a URL value preprocessor function for all view functions in the application. These functions will be called before the before_request() functions.
The function can modify the values captured from the matched url before
they are passed to the view. For example, this can be used to pop a common
language code value and place it in g rather than pass it to every view.
The function is passed the endpoint name and values dict. The return value
is ignored.
url_value_preprocessors = None
A dictionary with lists of functions that are called before the
before_request_funcs functions. The key of the dictionary is the name of
the blueprint this function is active for, or None for all requests. To register a
function, use url_value_preprocessor().
New in version 0.7.
Enable this if you want to use the X-Sendfile feature. Keep in mind that the
server has to support this. This only affects files sent with the send_file()
New in version 0.2.
This attribute can also be configured from the config with the
USE_X_SENDFILE configuration key. Defaults to False.
view_functions = None
A dictionary of all view functions registered. The keys will be function
names which are also used to generate URLs and the values are the function
objects themselves. To register a view function, use the route() decorator.
wsgi_app(environ, start_response)
The actual WSGI application. This is not implemented in __call__ so that
middlewares can be applied without losing a reference to the class. So instead of doing this:
app = MyMiddleware(app)
It’s a better idea to do this instead:
app.wsgi_app = MyMiddleware(app.wsgi_app)
Then you still have the original application object around and can continue
to call methods on it.
Changed in version 0.7: The behavior of the before and after request callbacks was changed under error conditions and a new callback was added
that will always execute at the end of the request, independent on if an error
occurred or not. See Callbacks and Errors.
• environ – a WSGI environment
• start_response – a callable accepting a status code, a list
of headers and an optional exception context to start the response
Blueprint Objects
class flask.Blueprint(name,
url_prefix=None, subdomain=None, url_defaults=None,
Represents a blueprint. A blueprint is an object that records functions that will
be called with the BlueprintSetupState later to register functions or other things
on the main application. See Modular Applications with Blueprints for more information.
New in version 0.7.
add_app_template_filter(f, name=None)
Register a custom template filter, available application wide. Like Flask.
add_template_filter() but for a blueprint.
Works exactly like the
app_template_filter() decorator.
Parameters name – the optional name of the filter, otherwise the
function name will be used.
add_app_template_global(f, name=None)
Register a custom template global, available application wide.
Flask.add_template_global() but for a blueprint. Works exactly like the
app_template_global() decorator.
New in version 0.10.
Parameters name – the optional name of the global, otherwise the
function name will be used.
add_app_template_test(f, name=None)
Register a custom template test, available application wide.
Flask.add_template_test() but for a blueprint. Works exactly like the
app_template_test() decorator.
New in version 0.10.
Parameters name – the optional name of the test, otherwise the
function name will be used.
add_url_rule(rule, endpoint=None, view_func=None, **options)
Like Flask.add_url_rule() but for a blueprint. The endpoint for the
url_for() function is prefixed with the name of the blueprint.
after_app_request(f )
Like Flask.after_request() but for a blueprint. Such a function is executed
after each request, even if outside of the blueprint.
after_request(f )
Like Flask.after_request() but for a blueprint. This function is only executed after each request that is handled by a function of that blueprint.
app_context_processor(f )
Like Flask.context_processor() but for a blueprint. Such a function is executed each request, even if outside of the blueprint.
Like Flask.errorhandler() but for a blueprint. This handler is used for all
requests, even if outside of the blueprint.
Register a custom template filter, available application wide. Like Flask.
template_filter() but for a blueprint.
Parameters name – the optional name of the filter, otherwise the
function name will be used.
Register a custom template global, available application wide. Like Flask.
template_global() but for a blueprint.
New in version 0.10.
Parameters name – the optional name of the global, otherwise the
function name will be used.
Register a custom template test, available application wide. Like Flask.
template_test() but for a blueprint.
New in version 0.10.
Parameters name – the optional name of the test, otherwise the
function name will be used.
app_url_defaults(f )
Same as url_defaults() but application wide.
app_url_value_preprocessor(f )
Same as url_value_preprocessor() but application wide.
before_app_first_request(f )
Like Flask.before_first_request(). Such a function is executed before the
first request to the application.
before_app_request(f )
Like Flask.before_request(). Such a function is executed before each request, even if outside of a blueprint.
before_request(f )
Like Flask.before_request() but for a blueprint. This function is only executed before each request that is handled by a function of that blueprint.
context_processor(f )
Like Flask.context_processor() but for a blueprint. This function is only
executed for requests handled by a blueprint.
Like Flask.endpoint() but for a blueprint. This does not prefix the endpoint with the blueprint name, this has to be done explicitly by the user of
this method. If the endpoint is prefixed with a . it will be registered to the
current blueprint, otherwise it’s an application independent endpoint.
Registers an error handler that becomes active for this blueprint only. Please
be aware that routing does not happen local to a blueprint so an error handler for 404 usually is not handled by a blueprint unless it is caused inside
a view function. Another special case is the 500 internal server error which
is always looked up from the application.
Otherwise works as the errorhandler() decorator of the Flask object.
Provides default cache_timeout for the send_file() functions.
By default, this function returns SEND_FILE_MAX_AGE_DEFAULT from the configuration of current_app.
Static file functions such as send_from_directory() use this function,
and send_file() calls this function on current_app when the given
cache_timeout is None. If a cache_timeout is given in send_file(), that timeout is used; otherwise, this method is called.
This allows subclasses to change the behavior when sending files based on
the filename. For example, to set the cache timeout for .js files to 60 seconds:
class MyFlask(flask.Flask):
def get_send_file_max_age(self, name):
if name.lower().endswith('.js'):
return 60
return flask.Flask.get_send_file_max_age(self, name)
New in version 0.9.
This is True if the package bound object’s container has a folder for static
New in version 0.5.
The Jinja loader for this package bound object.
New in version 0.5.
json_decoder = None
Blueprint local JSON decoder class to use. Set to None to use the app’s
json_encoder = None
Blueprint local JSON decoder class to use. Set to None to use the app’s
make_setup_state(app, options, first_registration=False)
Creates an instance of BlueprintSetupState() object that is later passed to
the register callback functions. Subclasses can override this to return a subclass of the setup state.
open_resource(resource, mode=’rb’)
Opens a resource from the application’s resource folder. To see how this
works, consider the following folder structure:
If you want to open the schema.sql file you would do the following:
with app.open_resource('schema.sql') as f:
contents = f.read()
• resource – the name of the resource. To access resources
within subfolders use forward slashes as separator.
• mode – resource file opening mode, default is ‘rb’.
Registers a function that is called when the blueprint is registered on the
application. This function is called with the state as argument as returned
by the make_setup_state() method.
Works like record() but wraps the function in another function that will
ensure the function is only called once. If the blueprint is registered a second
time on the application, the function passed is not called.
register(app, options, first_registration=False)
Called by Flask.register_blueprint() to register a blueprint on the application. This can be overridden to customize the register behavior. Keyword arguments from register_blueprint() are directly forwarded to this
method in the options dictionary.
register_error_handler(code_or_exception, f )
Non-decorator version of the errorhandler() error attach function, akin to
the register_error_handler() application-wide function of the Flask object but for error handlers limited to this blueprint.
New in version 0.11.
route(rule, **options)
Like Flask.route() but for a blueprint. The endpoint for the url_for()
function is prefixed with the name of the blueprint.
Function used internally to send static files from the static folder to the
New in version 0.5.
The absolute path to the configured static folder.
teardown_app_request(f )
Like Flask.teardown_request() but for a blueprint. Such a function is executed when tearing down each request, even if outside of the blueprint.
teardown_request(f )
Like Flask.teardown_request() but for a blueprint. This function is
only executed when tearing down requests handled by a function of that
blueprint. Teardown request functions are executed when the request context is popped, even when no actual request was performed.
url_defaults(f )
Callback function for URL defaults for this blueprint. It’s called with the
endpoint and values and should update the values passed in place.
url_value_preprocessor(f )
Registers a function as URL value preprocessor for this blueprint. It’s called
before the view functions are called and can modify the url values provided.
Incoming Request Data
class flask.Request(environ, populate_request=True, shallow=False)
The request object used by default in Flask. Remembers the matched endpoint
and view arguments.
It is what ends up as request. If you want to replace the request object used you
can subclass this and set request_class to your subclass.
The request object is a Request subclass and provides all of the attributes
Werkzeug defines plus a few Flask specific ones.
The underlying WSGI environment.
Provides different ways to look at the current IRI. Imagine your application
is listening on the following application root:
And a user requests the following URI:
In this case the values of the above mentioned attributes would be the following:
List of charsets this client supports as CharsetAccept object.
List of encodings this client accepts. Encodings in a HTTP term are compression encodings such as gzip. For charsets have a look at accept_charset.
List of languages this client accepts as LanguageAccept object.
List of mimetypes this client supports as MIMEAccept object.
If a forwarded header exists this is a list of all ip addresses from the client
ip to the last proxy server.
application(f )
Decorate a function as responder that accepts the request as first argument.
This works like the responder() decorator but the function is passed the
request object as first argument and the request object will be closed automatically:
def my_wsgi_app(request):
return Response('Hello World!')
Parameters f – the WSGI callable to decorate
Returns a new WSGI callable
The parsed URL parameters (the part in the URL after the question mark).
By default an ImmutableMultiDict is returned from this function. This can
be changed by setting parameter_storage_class to a different type. This
might be necessary if the order of the form data is important.
The Authorization object in parsed form.
Like url but without the querystring See also: trusted_hosts.
The name of the current blueprint
A RequestCacheControl object for the incoming cache control headers.
Closes associated resources of this request object. This closes all file handles
explicitly. You can also use the request object in a with statement which will
automatically close it.
New in version 0.9.
The Content-Encoding entity-header field is used as a modifier to the
media-type. When present, its value indicates what additional content codings have been applied to the entity-body, and thus what decoding mechanisms must be applied in order to obtain the media-type referenced by the
Content-Type header field.
New in version 0.9.
The Content-Length entity-header field indicates the size of the entity-body
in bytes or, in the case of the HEAD method, the size of the entity-body that
would have been sent had the request been a GET.
The Content-MD5 entity-header field, as defined in RFC 1864, is an
MD5 digest of the entity-body for the purpose of providing an endto-end message integrity check (MIC) of the entity-body. (Note:
a MIC is good for detecting accidental modification of the entitybody in transit, but is not proof against malicious attacks.)
New in version 0.9.
The Content-Type entity-header field indicates the media type of the entitybody sent to the recipient or, in the case of the HEAD method, the media
type that would have been sent had the request been a GET.
A dict with the contents of all cookies transmitted with the request.
Contains the incoming request data as string in case it came with a mimetype Werkzeug does not handle.
The Date general-header field represents the date and time at which the
message was originated, having the same semantics as orig-date in RFC
alias of ImmutableTypeConversionDict
The endpoint that matched the request. This in combination with view_args
can be used to reconstruct the same or a modified URL. If an exception
happened when matching, this will be None.
MultiDict object containing all uploaded files. Each key in files is the
name from the <input type="file" name="">. Each value in files is a
Werkzeug FileStorage object.
It basically behaves like a standard file object you know from Python, with
the difference that it also has a save() function that can store the file on the
Note that files will only contain data if the request method was
POST, PUT or PATCH and the <form> that posted to the request had
enctype="multipart/form-data". It will be empty otherwise.
See the MultiDict / FileStorage documentation for more details about the
used data structure.
The form parameters. By default an ImmutableMultiDict is returned from
this function. This can be changed by setting parameter_storage_class to
a different type. This might be necessary if the order of the form data is
Please keep in mind that file uploads will not end up here, but instead in
the files attribute.
Changed in version 0.9: Previous to Werkzeug 0.9 this would only contain
form data for POST and PUT requests.
alias of FormDataParser
from_values(*args, **kwargs)
Create a new request object based on the values provided. If environ is
given missing values are filled from there. This method is useful for small
scripts when you need to simulate a request from an URL. Do not use this
method for unittesting, there is a full featured client object (Client) that
allows to create multipart requests, support for cookies etc.
This accepts the same options as the EnvironBuilder.
Changed in version 0.5: This method now accepts the same arguments as
EnvironBuilder. Because of this the environ parameter is now called environ_overrides.
Returns request object
Requested path as unicode, including the query string.
get_data(cache=True, as_text=False, parse_form_data=False)
This reads the buffered incoming data from the client into one bytestring.
By default this is cached but that behavior can be changed by setting cache
to False.
Usually it’s a bad idea to call this method without checking the content
length first as a client could send dozens of megabytes or more to cause
memory problems on the server.
Note that if the form data was already parsed this method will not return
anything as form data parsing does not cache the data like this method does.
To implicitly invoke form data parsing function set parse_form_data to True.
When this is done the return value of this method will be an empty string
if the form parser handles the data. This generally is not necessary as if
the whole data is cached (which is the default) the form parser will used
the cached data to parse the form data. Please be generally aware of checking the content length first in any case before calling this method to avoid
exhausting server memory.
If as_text is set to True the return value will be a decoded unicode string.
New in version 0.9.
get_json(force=False, silent=False, cache=True)
Parses the incoming JSON request data and returns it. By default this
function will return None if the mimetype is not application/json but
this can be overridden by the force parameter. If parsing fails the
on_json_loading_failed() method on the request object will be invoked.
• force – if set to True the mimetype is ignored.
• silent – if set to True this method will fail silently and return
• cache – if set to True the parsed JSON data is remembered on
the request.
The headers from the WSGI environ as immutable EnvironHeaders.
Just the host including the port if available. See also: trusted_hosts.
Just the host with scheme as IRI. See also: trusted_hosts.
An object containing all the etags in the If-Match header.
Return type ETags
The parsed If-Modified-Since header as datetime object.
An object containing all the etags in the If-None-Match header.
Return type ETags
The parsed If-Range header.
New in version 0.7.
Return type IfRange
The parsed If-Unmodified-Since header as datetime object.
Indicates if this request is JSON or not. By default a request is considered
to include JSON data if the mimetype is application/json or application/
New in version 0.11.
boolean that is True if the application is served by a WSGI server that
spawns multiple processes.
boolean that is True if the application is served by a multithreaded WSGI
boolean that is True if the application will be executed only once in a process
lifetime. This is the case for CGI for example, but it’s not guaranteed that
the execution only happens one time.
True if the request is secure.
True if the request was triggered via a JavaScript XMLHttpRequest. This
only works with libraries that support the X-Requested-With header and set
it to “XMLHttpRequest”. Libraries that do that are prototype, jQuery and
Mochikit and probably some more.
If the mimetype is application/json this will contain the parsed JSON data.
Otherwise this will be None.
The get_json() method should be used instead.
alias of ImmutableList
Creates the form data parser. Instanciates the form_data_parser_class with
some parameters.
New in version 0.8.
Read-only view of the MAX_CONTENT_LENGTH config key.
The Max-Forwards request-header field provides a mechanism with the
TRACE and OPTIONS methods to limit the number of proxies or gateways
that can forward the request to the next inbound server.
The request method. (For example 'GET' or 'POST').
Like content_type, but without parameters (eg, without charset, type etc.)
and always lowercase. For example if the content type is text/HTML;
charset=utf-8 the mimetype would be 'text/html'.
The mimetype parameters as dict. For example if the content type is text/
html; charset=utf-8 the params would be {'charset': 'utf-8'}.
The name of the current module if the request was dispatched to an actual
module. This is deprecated functionality, use blueprints instead.
Called if decoding of the JSON data failed. The return value of this method
is used by get_json() when an error occurred. The default implementation
just raises a BadRequest exception.
Changed in version 0.10: Removed buggy previous behavior of generating
a random JSON response. If you want that behavior back you can trivially
add it by subclassing.
New in version 0.8.
alias of ImmutableMultiDict
Requested path as unicode. This works a bit like the regular path info in the
WSGI environment but will always include a leading slash, even if the URL
root is accessed.
The Pragma general-header field is used to include implementation-specific
directives that might apply to any recipient along the request/response
chain. All pragma directives specify optional behavior from the viewpoint
of the protocol; however, some systems MAY require that behavior be consistent with the directives.
The URL parameters as raw bytestring.
The parsed Range header.
New in version 0.7.
Return type Range
The Referer[sic] request-header field allows the client to specify, for the
server’s benefit, the address (URI) of the resource from which the RequestURI was obtained (the “referrer”, although the header field is misspelled).
The remote address of the client.
If the server supports user authentication, and the script is protected, this
attribute contains the username the user has authenticated as.
routing_exception = None
If matching the URL failed, this is the exception that will be raised / was
raised as part of the request handling. This is usually a NotFound exception
or something similar.
URL scheme (http or https).
New in version 0.7.
The root path of the script without the trailing slash.
If the incoming form data was not encoded with a known mimetype the
data is stored unmodified in this stream for consumption. Most of the time
it is a better idea to use data which will give you that data as a string. The
stream only returns the data once.
Unlike input_stream this stream is properly guarded that you can’t accidentally read past the length of the input. Werkzeug will internally always
refer to this stream to read data which makes it possible to wrap this object
with a stream that does filtering.
Changed in version 0.9: This stream is now always available but might be
consumed by the form parser later on. Previously the stream was only set
if no parsing happened.
The reconstructed current URL as IRI. See also: trusted_hosts.
The charset that is assumed for URLs. Defaults to the value of charset.
New in version 0.6.
The full URL root (with hostname), this is the application root as IRI. See
also: trusted_hosts.
url_rule = None
The internal URL rule that matched the request. This can be useful to in-
spect which methods are allowed for the URL from a before/after handler
(request.url_rule.methods) etc.
New in version 0.6.
The current user agent.
A werkzeug.datastructures.CombinedMultiDict that combines args and
view_args = None
A dict of view arguments that matched the request. If an exception happened when matching, this will be None.
Returns True if the request method carries content. As of Werkzeug 0.9 this
will be the case if a content type is transmitted.
New in version 0.8.
To access incoming request data, you can use the global request object. Flask
parses incoming request data for you and gives you access to it through that
global object. Internally Flask makes sure that you always get the correct data
for the active thread if you are in a multithreaded environment.
This is a proxy. See Notes On Proxies for more information.
The request object is an instance of a Request subclass and provides all of the
attributes Werkzeug defines. This just shows a quick overview of the most important ones.
Response Objects
class flask.Response(response=None,
mimetype=None, content_type=None, direct_passthrough=False)
The response object that is used by default in Flask. Works like the response
object from Werkzeug but is set to have an HTML mimetype by default. Quite
often you don’t have to create this object yourself because make_response() will
take care of that for you.
If you want to replace the response object used you can subclass this and set
response_class to your subclass.
A Headers object representing the response headers.
A string with a response status.
The response status as integer.
A descriptor that calls get_data() and set_data(). This should not be used
and will eventually get deprecated.
The mimetype (content type without charset etc.)
set_cookie(key, value=’‘, max_age=None, expires=None, path=’/’, domain=None, secure=False, httponly=False)
Sets a cookie. The parameters are the same as in the cookie Morsel object in
the Python standard library but it accepts unicode data, too.
• key – the key (name) of the cookie to be set.
• value – the value of the cookie.
• max_age – should be a number of seconds, or None (default)
if the cookie should last only as long as the client’s browser
• expires – should be a datetime object or UNIX timestamp.
• path – limits the cookie to a given path, per default it will
span the whole domain.
• domain – if you want to set a cross-domain cookie. For example, domain=".example.com" will set a cookie that is readable
by the domain www.example.com, foo.example.com etc. Otherwise, a cookie will only be readable by the domain that set
• secure – If True, the cookie will only be available via HTTPS
• httponly – disallow JavaScript to access the cookie. This is an
extension to the cookie standard and probably not supported
by all browsers.
If you have the Flask.secret_key set you can use sessions in Flask applications. A
session basically makes it possible to remember information from one request to another. The way Flask does this is by using a signed cookie. So the user can look at the
session contents, but not modify it unless they know the secret key, so make sure to
set that to something complex and unguessable.
To access the current session you can use the session object:
class flask.session
The session object works pretty much like an ordinary dict, with the difference
that it keeps track on modifications.
This is a proxy. See Notes On Proxies for more information.
The following attributes are interesting:
True if the session is new, False otherwise.
True if the session object detected a modification. Be advised that modifications on mutable structures are not picked up automatically, in that situation
you have to explicitly set the attribute to True yourself. Here an example:
# this change is not picked up because a mutable object (here
# a list) is changed.
# so mark it as modified yourself
session.modified = True
If set to True the session lives for permanent_session_lifetime seconds. The
default is 31 days. If set to False (which is the default) the session will be
deleted when the user closes the browser.
Session Interface
New in version 0.8.
The session interface provides a simple way to replace the session implementation that
Flask is using.
class flask.sessions.SessionInterface
The basic interface you have to implement in order to replace the default session
interface which uses werkzeug’s securecookie implementation. The only methods you have to implement are open_session() and save_session(), the others
have useful defaults which you don’t need to change.
The session object returned by the open_session() method has to provide a dictionary like interface plus the properties and methods from the SessionMixin.
We recommend just subclassing a dict and adding that mixin:
class Session(dict, SessionMixin):
If open_session() returns None Flask will call into make_null_session() to create
a session that acts as replacement if the session support cannot work because
some requirement is not fulfilled. The default NullSession class that is created
will complain that the secret key was not set.
To replace the session interface on an application all you have to do is to assign
app = Flask(__name__)
app.session_interface = MySessionInterface()
New in version 0.8.
Returns the domain that should be set for the session cookie.
Uses SESSION_COOKIE_DOMAIN if it is configured, otherwise falls back to detecting the domain based on SERVER_NAME.
Once detected (or if not set at all), SESSION_COOKIE_DOMAIN is updated to
avoid re-running the logic.
Returns True if the session cookie should be httponly. This currently just
returns the value of the SESSION_COOKIE_HTTPONLY config var.
Returns the path for which the cookie should be valid. The default implementation uses the value from the SESSION_COOKIE_PATH config var if it’s set,
and falls back to APPLICATION_ROOT or uses / if it’s None.
Returns True if the cookie should be secure. This currently just returns the
value of the SESSION_COOKIE_SECURE setting.
get_expiration_time(app, session)
A helper method that returns an expiration date for the session or None if the
session is linked to the browser session. The default implementation returns
now + the permanent session lifetime configured on the application.
Checks if a given object is a null session. Null sessions are not asked to be
This checks if the object is an instance of null_session_class by default.
Creates a null session which acts as a replacement object if the real session
support could not be loaded due to a configuration error. This mainly aids
the user experience because the job of the null session is to still support
lookup without complaining but modifications are answered with a helpful
error message of what failed.
This creates an instance of null_session_class by default.
make_null_session() will look here for the class that should be created
when a null session is requested. Likewise the is_null_session() method
will perform a typecheck against this type.
alias of NullSession
open_session(app, request)
This method has to be implemented and must either return None in case
the loading failed because of a configuration error or an instance of a session object which implements a dictionary like interface + the methods and
attributes on SessionMixin.
pickle_based = False
A flag that indicates if the session interface is pickle based. This can be used
by Flask extensions to make a decision in regards to how to deal with the
session object.
New in version 0.10.
save_session(app, session, response)
This is called for actual sessions returned by open_session() at the end of
the request. This is still called during a request context so if you absolutely
need access to the request you can do that.
should_set_cookie(app, session)
Used by session backends to determine if a Set-Cookie header should
be set for this session cookie for this response. If the session has
been modified, the cookie is set. If the session is permanent and the
SESSION_REFRESH_EACH_REQUEST config is true, the cookie is always set.
This check is usually skipped if the session was deleted.
New in version 0.11.
class flask.sessions.SecureCookieSessionInterface
The default session interface that stores sessions in signed cookies through the
itsdangerous module.
static digest_method()
the hash function to use for the signature. The default is sha1
key_derivation = ‘hmac’
the name of the itsdangerous supported key derivation. The default is
salt = ‘cookie-session’
the salt that should be applied on top of the secret key for the signing of
cookie based sessions.
serializer = <flask.sessions.TaggedJSONSerializer object>
A python serializer for the payload. The default is a compact JSON derived
serializer with support for some extra Python types such as datetime objects
or tuples.
alias of SecureCookieSession
class flask.sessions.SecureCookieSession(initial=None)
Base class for sessions based on signed cookies.
class flask.sessions.NullSession(initial=None)
Class used to generate nicer error messages if sessions are not available. Will still
allow read-only access to the empty session but fail on setting.
class flask.sessions.SessionMixin
Expands a basic dictionary with an accessors that are expected by Flask extensions and users for the session.
accessed = True
the accessed variable indicates whether or not the session object has been
accessed in that request. This allows flask to append a Vary: Cookie header
to the response if the session is being accessed. This allows caching proxy
servers, like Varnish, to use both the URL and the session cookie as keys
when caching pages, preventing multiple users from being served the same
modified = True
for some backends this will always be True, but some backends will default
this to false and detect changes in the dictionary for as long as changes do
not happen on mutable structures in the session. The default mixin implementation just hardcodes True in.
new = False
some session backends can tell you if a session is new, but that is not necessarily guaranteed. Use with caution. The default mixin implementation
just hardcodes False in.
this reflects the '_permanent' key in the dict.
flask.sessions.session_json_serializer = <flask.sessions.TaggedJSONSerializer object>
A customized JSON serializer that supports a few extra types that we take for
granted when serializing (tuples, markup objects, datetime).
This object provides dumping and loading methods similar to simplejson but it
also tags certain builtin Python objects that commonly appear in sessions. Currently the following extended values are supported in the JSON it dumps:
•Markup objects
•UUID objects
•datetime objects
The PERMANENT_SESSION_LIFETIME config key can also be an integer starting with Flask
0.8. Either catch this down yourself or use the permanent_session_lifetime attribute
on the app which converts the result to an integer automatically.
Test Client
class flask.testing.FlaskClient(*args, **kwargs)
Works like a regular Werkzeug test client but has some knowledge about how
Flask works to defer the cleanup of the request context stack to the end of a with
body when used in a with statement. For general information about how to use
this class refer to werkzeug.test.Client.
Changed in version 0.12: app.test_client() includes preset default environment, which can be set after instantiation of the app.test_client() object in
Basic usage is outlined in the Testing Flask Applications chapter.
session_transaction(*args, **kwargs)
When used in combination with a with statement this opens a session transaction. This can be used to modify the session that the test client uses. Once
the with block is left the session is stored back.
with client.session_transaction() as session:
session['value'] = 42
Internally this is implemented by going through a temporary test request
context and since session handling could depend on request variables this
function accepts the same arguments as test_request_context() which are
directly passed through.
Application Globals
To share data that is valid for one request only from one function to another, a global
variable is not good enough because it would break in threaded environments. Flask
provides you with a special object that ensures it is only valid for the active request
and that will return different values for each request. In a nutshell: it does the right
thing, like it does for request and session.
Just store on this whatever you want. For example a database connection or the
user that is currently logged in.
Starting with Flask 0.10 this is stored on the application context and no longer
on the request context which means it becomes available if only the application
context is bound and not yet a request. This is especially useful when combined
with the Faking Resources and Context pattern for testing.
Additionally as of 0.10 you can use the get() method to get an attribute or None
(or the second argument) if it’s not set. These two usages are now equivalent:
user = getattr(flask.g, 'user', None)
user = flask.g.get('user', None)
It’s now also possible to use the in operator on it to see if an attribute is defined
and it yields all keys on iteration.
As of 0.11 you can use pop() and setdefault() in the same way you would use
them on a dictionary.
This is a proxy. See Notes On Proxies for more information.
Useful Functions and Classes
Points to the application handling the request. This is useful for extensions that
want to support multiple applications running side by side. This is powered by
the application context and not by the request context, so you can change the
value of this proxy by using the app_context() method.
This is a proxy. See Notes On Proxies for more information.
If you have code that wants to test if a request context is there or not this function
can be used. For instance, you may want to take advantage of request information if the request object is available, but fail silently if it is unavailable.
class User(db.Model):
def __init__(self, username, remote_addr=None):
self.username = username
if remote_addr is None and has_request_context():
remote_addr = request.remote_addr
self.remote_addr = remote_addr
Alternatively you can also just test any of the context bound objects (such as
request or g for truthness):
class User(db.Model):
def __init__(self, username, remote_addr=None):
self.username = username
if remote_addr is None and request:
remote_addr = request.remote_addr
self.remote_addr = remote_addr
New in version 0.7.
flask.copy_current_request_context(f )
A helper function that decorates a function to retain the current request context.
This is useful when working with greenlets. The moment the function is decorated a copy of the request context is created and then pushed when the function
is called.
import gevent
from flask import copy_current_request_context
def index():
def do_some_work():
# do some work here, it can access flask.request like you
# would otherwise in the view function.
return 'Regular response'
New in version 0.10.
Works like has_request_context() but for the application context. You can also
just do a boolean check on the current_app object instead.
New in version 0.9.
flask.url_for(endpoint, **values)
Generates a URL to the given endpoint with the method provided.
Variable arguments that are unknown to the target endpoint are appended to the
generated URL as query arguments. If the value of a query argument is None, the
whole pair is skipped. In case blueprints are active you can shortcut references
to the same blueprint by prefixing the local endpoint with a dot (.).
This will reference the index function local to the current blueprint:
For more information, head over to the Quickstart.
To integrate applications, Flask has a hook to intercept URL build errors through
Flask.url_build_error_handlers. The url_for function results in a BuildError
when the current app does not have a URL for the given endpoint and values.
When it does, the current_app calls its url_build_error_handlers if it is not
None, which can return a string to use as the result of url_for (instead of url_for‘s
default to raise the BuildError exception) or re-raise the exception. An example:
def external_url_handler(error, endpoint, values):
"Looks up an external URL when `url_for` cannot build a URL."
# This is an example of hooking the build_error_handler.
# Here, lookup_url is some utility function you've built
# which looks up the endpoint in some external URL registry.
url = lookup_url(endpoint, **values)
if url is None:
# External lookup did not have a URL.
# Re-raise the BuildError, in context of original traceback.
exc_type, exc_value, tb = sys.exc_info()
if exc_value is error:
raise exc_type, exc_value, tb
raise error
# url_for will use this result, instead of raising BuildError.
return url
Here, error is the instance of BuildError, and endpoint and values are the arguments passed into url_for. Note that this is for building URLs outside the current
application, and not for handling 404 NotFound errors.
New in version 0.10: The _scheme parameter was added.
New in version 0.9: The _anchor and _method parameters were added.
New in version 0.9: Calls Flask.handle_build_error() on BuildError.
• endpoint – the endpoint of the URL (name of the function)
• values – the variable arguments of the URL rule
• _external – if set to True, an absolute URL is generated. Server
address can be changed via SERVER_NAME configuration variable
which defaults to localhost.
• _scheme – a string specifying the desired URL scheme. The _external parameter must be set to True or a ValueError is raised.
The default behavior uses the same scheme as the current request, or PREFERRED_URL_SCHEME from the app configuration if no
request context is available. As of Werkzeug 0.10, this also can
be set to an empty string to build protocol-relative URLs.
• _anchor – if provided this is added as anchor to the URL.
• _method – if provided this explicitly specifies an HTTP method.
flask.abort(status, *args, **kwargs)
Raises an HTTPException for the given status code or WSGI application:
abort(404) # 404 Not Found
abort(Response('Hello World'))
Can be passed a WSGI application or a status code. If a status code is given it’s
looked up in the list of exceptions and will raise that exception, if passed a WSGI
application it will wrap it in a proxy WSGI exception and raise that:
abort(Response('Hello World'))
flask.redirect(location, code=302, Response=None)
Returns a response object (a WSGI application) that, if called, redirects the client
to the target location. Supported codes are 301, 302, 303, 305, and 307. 300 is not
supported because it’s not a real redirect and 304 because it’s the answer for a
request with a request with defined If-Modified-Since headers.
New in version 0.6: The location can now be a unicode string that is encoded
using the iri_to_uri() function.
New in version 0.10: The class used for the Response object can now be passed
• location – the location the response should redirect to.
• code – the redirect status code. defaults to 302.
• Response (class) – a Response class to use when instantiating
a response. The default is werkzeug.wrappers.Response if unspecified.
Sometimes it is necessary to set additional headers in a view. Because views
do not have to return response objects but can return a value that is converted
into a response object by Flask itself, it becomes tricky to add headers to it. This
function can be called instead of using a return and you will get a response object
which you can use to attach headers.
If view looked like this and you want to add a new header:
def index():
return render_template('index.html', foo=42)
You can now do something like this:
def index():
response = make_response(render_template('index.html', foo=42))
response.headers['X-Parachutes'] = 'parachutes are cool'
return response
This function accepts the very same arguments you can return from a view function. This for example creates a response with a 404 error code:
response = make_response(render_template('not_found.html'), 404)
The other use case of this function is to force the return value of a view function
into a response which is helpful with view decorators:
response = make_response(view_function())
response.headers['X-Parachutes'] = 'parachutes are cool'
Internally this function does the following things:
•if no arguments are passed, it creates a new response argument
•if one argument is passed, flask.Flask.make_response() is invoked with
•if more than one argument is passed, the arguments are passed to the flask.
Flask.make_response() function as tuple.
New in version 0.6.
flask.after_this_request(f )
Executes a function after this request. This is useful to modify response objects.
The function is passed the response object and has to return the same or a new
def index():
def add_header(response):
response.headers['X-Foo'] = 'Parachute'
return response
return 'Hello World!'
This is more useful if a function other than the view function wants to modify
a response. For instance think of a decorator that wants to add some headers
without converting the return value into a response object.
New in version 0.9.
flask.send_file(filename_or_fp, mimetype=None, as_attachment=False, attachment_filename=None, add_etags=True, cache_timeout=None, conditional=False, last_modified=None)
Sends the contents of a file to the client. This will use the most efficient
method available and configured. By default it will try to use the WSGI
server’s file_wrapper support. Alternatively you can set the application’s
use_x_sendfile attribute to True to directly emit an X-Sendfile header. This
however requires support of the underlying webserver for X-Sendfile.
By default it will try to guess the mimetype for you, but you can also explicitly
provide one. For extra security you probably want to send certain files as attachment (HTML for instance). The mimetype guessing requires a filename or an
attachment_filename to be provided.
ETags will also be attached automatically if a filename is provided. You can turn
this off by setting add_etags=False.
If conditional=True and filename is provided, this method will try to upgrade the
response stream to support range requests. This will allow the request to be
answered with partial content response.
Please never pass filenames to this function from user sources; you should use
send_from_directory() instead.
New in version 0.2.
New in version 0.5: The add_etags, cache_timeout and conditional parameters were
added. The default behavior is now to attach etags.
Changed in version 0.7: mimetype guessing and etag support for file objects was
deprecated because it was unreliable. Pass a filename if you are able to, otherwise
attach an etag yourself. This functionality will be removed in Flask 1.0
Changed in version 0.9: cache_timeout pulls its default from application config,
when None.
Changed in version 0.12: The filename is no longer automatically inferred from
file objects. If you want to use automatic mimetype and etag support, pass a
filepath via filename_or_fp or attachment_filename.
Changed in version 0.12: The attachment_filename is preferred over filename for
MIME-type detection.
Changed in version 0.13: UTF-8 filenames, as specified in RFC 2231, are supported.
• filename_or_fp – the filename of the file to send. This is relative to the root_path if a relative path is specified. Alternatively a file object might be provided in which case X-Sendfile
might not work and fall back to the traditional method. Make
sure that the file pointer is positioned at the start of data to
send before calling send_file().
• mimetype – the mimetype of the file if provided. If a file path is
given, auto detection happens as fallback, otherwise an error
will be raised.
• as_attachment – set to True if you want to send this file with a
Content-Disposition: attachment header.
• attachment_filename – the filename for the attachment if it differs from the file’s filename.
• add_etags – set to False to disable attaching of etags.
• conditional – set to True to enable conditional responses.
• cache_timeout – the timeout in seconds for the headers. When
None (default), this value is set by get_send_file_max_age() of
• last_modified – set the Last-Modified header to this value, a
datetime or timestamp. If a file was passed, this overrides its
flask.send_from_directory(directory, filename, **options)
Send a file from a given directory with send_file(). This is a secure way to
quickly expose static files from an upload folder or something similar.
Example usage:
def download_file(filename):
return send_from_directory(app.config['UPLOAD_FOLDER'],
filename, as_attachment=True)
Sending files and Performance
It is strongly recommended to activate either X-Sendfile support in your webserver or (if no authentication happens) to tell the webserver to serve files for
the given path on its own without calling into the web application for improved
New in version 0.5.
• directory – the directory where all the files are stored.
• filename – the filename relative to that directory to download.
• options – optional keyword arguments that are directly forwarded to send_file().
flask.safe_join(directory, *pathnames)
Safely join directory and zero or more untrusted pathnames components.
Example usage:
def wiki_page(filename):
filename = safe_join(app.config['WIKI_FOLDER'], filename)
with open(filename, 'rb') as fd:
content = fd.read() # Read and process the file content...
• directory – the trusted base directory.
• pathnames – the untrusted pathnames relative to that directory.
Raises NotFound if one or more passed paths fall out of its boundaries.
flask.escape(s) → markup
Convert the characters &, <, >, ‘, and ” in string s to HTML-safe sequences. Use
this if you need to display text that might contain such characters in HTML.
Marks return value as markup string.
class flask.Markup
Marks a string as being safe for inclusion in HTML/XML output without needing to be escaped. This implements the __html__ interface a couple of frameworks and web applications use. Markup is a direct subclass of unicode and pro242
vides all the methods of unicode just that it escapes arguments passed and always
returns Markup.
The escape function returns markup objects so that double escaping can’t happen.
The constructor of the Markup class can be used for three different things: When
passed an unicode object it’s assumed to be safe, when passed an object with
an HTML representation (has an __html__ method) that representation is used,
otherwise the object passed is converted into a unicode string and then assumed
to be safe:
>>> Markup("Hello <em>World</em>!")
Markup(u'Hello <em>World</em>!')
>>> class Foo(object):
... def __html__(self):
... return '<a href="#">foo</a>'
>>> Markup(Foo())
Markup(u'<a href="#">foo</a>')
If you want object passed being always treated as unsafe you can use the
escape() classmethod to create a Markup object:
>>> Markup.escape("Hello <em>World</em>!")
Markup(u'Hello &lt;em&gt;World&lt;/em&gt;!')
Operations on a markup string are markup aware which means that all arguments are passed through the escape() function:
>>> em = Markup("<em>%s</em>")
>>> em % "foo & bar"
Markup(u'<em>foo &amp; bar</em>')
>>> strong = Markup("<strong>%(text)s</strong>")
>>> strong % {'text': '<blink>hacker here</blink>'}
Markup(u'<strong>&lt;blink&gt;hacker here&lt;/blink&gt;</strong>')
>>> Markup("<em>Hello</em> ") + "<foo>"
Markup(u'<em>Hello</em> &lt;foo&gt;')
classmethod escape(s)
Escape the string. Works like escape() with the difference that for subclasses of Markup this function would return the correct subclass.
Unescape markup into an text_type string and strip all tags. This also resolves known HTML4 and XHTML entities. Whitespace is normalized to
>>> Markup("Main &raquo;
u'Main \xbb About'
Unescape markup again into an text_type string. This also resolves known
HTML4 and XHTML entities:
>>> Markup("Main &raquo; <em>About</em>").unescape()
u'Main \xbb <em>About</em>'
Message Flashing
flask.flash(message, category=’message’)
Flashes a message to the next request. In order to remove the flashed message from the session and to display it to the user, the template has to call
Changed in version 0.3: category parameter added.
• message – the message to be flashed.
• category – the category for the message. The following values
are recommended: 'message' for any kind of message, 'error'
for errors, 'info' for information messages and 'warning' for
warnings. However any kind of string can be used as category.
flask.get_flashed_messages(with_categories=False, category_filter=[])
Pulls all flashed messages from the session and returns them. Further calls in the
same request to the function will return the same messages. By default just the
messages are returned, but when with_categories is set to True, the return value
will be a list of tuples in the form (category, message) instead.
Filter the flashed messages to one or more categories by providing those categories in category_filter. This allows rendering categories in separate html blocks.
The with_categories and category_filter arguments are distinct:
•with_categories controls whether categories are returned with message text
(True gives a tuple, where False gives just the message text).
•category_filter filters the messages down to only those matching the provided categories.
See Message Flashing for examples.
Changed in version 0.3: with_categories parameter added.
Changed in version 0.9: category_filter parameter added.
• with_categories – set to True to also receive categories.
• category_filter – whitelist of categories to limit return values
JSON Support
Flask uses simplejson for the JSON implementation. Since simplejson is provided by
both the standard library as well as extension, Flask will try simplejson first and then
fall back to the stdlib json module. On top of that it will delegate access to the current
application’s JSON encoders and decoders for easier customization.
So for starters instead of doing:
import simplejson as json
except ImportError:
import json
You can instead just do this:
from flask import json
For usage examples, read the json documentation in the standard library. The following extensions are by default applied to the stdlib’s JSON module:
1. datetime objects are serialized as RFC 822 strings.
2. Any object with an __html__ method (like Markup) will have that method called
and then the return value is serialized as string.
The htmlsafe_dumps() function of this json module is also available as filter called
|tojson in Jinja2. Note that inside script tags no escaping must take place, so make
sure to disable escaping with |safe if you intend to use it inside script tags unless
you are using Flask 0.10 which implies that:
<script type=text/javascript>
doSomethingWith({{ user.username|tojson|safe }});
Auto-Sort JSON Keys
The configuration variable JSON_SORT_KEYS (Configuration Handling) can be set to false
to stop Flask from auto-sorting keys. By default sorting is enabled and outside of the
app context sorting is turned on.
Notice that disabling key sorting can cause issues when using content based HTTP
caches and Python’s hash randomization feature.
flask.json.jsonify(*args, **kwargs)
This function wraps dumps() to add a few enhancements that make life easier. It
turns the JSON output into a Response object with the application/json mimetype. For convenience, it also converts multiple arguments into an array or multiple keyword arguments into a dict. This means that both jsonify(1,2,3) and
jsonify([1,2,3]) serialize to [1,2,3].
For clarity, the JSON serialization behavior has the following differences from
1.Single argument: Passed straight through to dumps().
2.Multiple arguments: Converted to an array before being passed to dumps().
3.Multiple keyword arguments: Converted to a dict before being passed to
4.Both args and kwargs: Behavior undefined and will throw an exception.
Example usage:
from flask import jsonify
def get_current_user():
return jsonify(username=g.user.username,
This will send a JSON response like this to the browser:
"username": "admin",
"email": "[email protected]",
"id": 42
Changed in version 0.11: Added support for serializing top-level arrays. This
introduces a security risk in ancient browsers. See JSON Security for details.
JSONIFY_PRETTYPRINT_REGULAR config parameter is set to True or the Flask
app is running in debug mode. Compressed (not pretty) formatting currently
means no indents and no spaces after separators.
New in version 0.2.
flask.json.dumps(obj, **kwargs)
Serialize obj to a JSON formatted str by using the application’s configured encoder (json_encoder) if there is an application on the stack.
This function can return unicode strings or ascii-only bytestrings by default
which coerce into unicode strings automatically. That behavior by default is
controlled by the JSON_AS_ASCII configuration variable and can be overridden
by the simplejson ensure_ascii parameter.
flask.json.dump(obj, fp, **kwargs)
Like dumps() but writes into a file object.
flask.json.loads(s, **kwargs)
Unserialize a JSON object from a string s by using the application’s configured
decoder (json_decoder) if there is an application on the stack.
flask.json.load(fp, **kwargs)
Like loads() but reads from a file object.
class flask.json.JSONEncoder(skipkeys=False,
sort_keys=False, indent=None, separators=None,
encoding=’utf-8’, default=None)
The default Flask JSON encoder. This one extends the default simplejson encoder
by also supporting datetime objects, UUID as well as Markup objects which are
serialized as RFC 822 datetime strings (same as the HTTP date format). In order
to support more data types override the default() method.
Implement this method in a subclass such that it returns a serializable object
for o, or calls the base implementation (to raise a TypeError).
For example, to support arbitrary iterators, you could implement default
like this:
def default(self, o):
iterable = iter(o)
except TypeError:
return list(iterable)
return JSONEncoder.default(self, o)
class flask.json.JSONDecoder(encoding=None,
The default JSON decoder. This one does not change the behavior from the default simplejson decoder. Consult the json documentation for more information. This decoder is not only used for the load functions of this module but also
Template Rendering
flask.render_template(template_name_or_list, **context)
Renders a template from the template folder with the given context.
• template_name_or_list – the name of the template to be rendered, or an iterable with template names the first one existing
will be rendered
• context – the variables that should be available in the context
of the template.
flask.render_template_string(source, **context)
Renders a template from the given template source string with the given context.
Template variables will be autoescaped.
• source – the source code of the template to be rendered
• context – the variables that should be available in the context
of the template.
flask.get_template_attribute(template_name, attribute)
Loads a macro (or variable) a template exports. This can be used to invoke a
macro from within Python code. If you for example have a template named
_cider.html with the following contents:
{% macro hello(name) %}Hello {{ name }}!{% endmacro %}
You can access this from Python code like this:
hello = get_template_attribute('_cider.html', 'hello')
return hello('World')
New in version 0.2.
• template_name – the name of the template
• attribute – the name of the variable of macro to access
class flask.Config(root_path, defaults=None)
Works exactly like a dict but provides ways to fill it from files or special dictionaries. There are two common patterns to populate the config.
Either you can fill the config from a config file:
Or alternatively you can define the configuration options in the module that calls
from_object() or provide an import path to a module that should be loaded.
It is also possible to tell it to use the same module and with that provide the
configuration values just before the call:
DEBUG = True
SECRET_KEY = 'development key'
In both cases (loading from any Python file or loading from modules), only uppercase keys are added to the config. This makes it possible to use lowercase
values in the config file for temporary values that are not added to the config or
to define the config keys in the same file that implements the application.
Probably the most interesting way to load configurations is from an environment
variable pointing to a file:
In this case before launching the application you have to set this environment
variable to the file you want to use. On Linux and OS X use the export statement:
export YOURAPPLICATION_SETTINGS='/path/to/config/file'
On windows use set instead.
• root_path – path to which files are read relative from. When
the config object is created by the application, this is the application’s root_path.
• defaults – an optional dictionary of default values
from_envvar(variable_name, silent=False)
Loads a configuration from an environment variable pointing to a configuration file. This is basically just a shortcut with nicer error messages for this
line of code:
• variable_name – name of the environment variable
• silent – set to True if you want silent failure for missing files.
Returns bool. True if able to load config, False otherwise.
from_json(filename, silent=False)
Updates the values in the config from a JSON file. This function behaves
as if the JSON object was a dictionary and passed to the from_mapping()
• filename – the filename of the JSON file. This can either be
an absolute filename or a filename relative to the root path.
• silent – set to True if you want silent failure for missing files.
New in version 0.11.
from_mapping(*mapping, **kwargs)
Updates the config like update() ignoring items with non-upper keys.
New in version 0.11.
Updates the values from the given object. An object can be of one of the
following two types:
•a string: in this case the object with that name will be imported
•an actual object reference: that object is used directly
Objects are usually either modules or classes. from_object() loads only the
uppercase attributes of the module/class. A dict object will not work with
from_object() because the keys of a dict are not attributes of the dict class.
Example of module-based configuration:
from yourapplication import default_config
You should not use this function to load the actual configuration but
rather configuration defaults. The actual config should be loaded with
from_pyfile() and ideally from a location not within the package because
the package might be installed system wide.
See Development / Production for an example of class-based configuration
using from_object().
Parameters obj – an import name or object
from_pyfile(filename, silent=False)
Updates the values in the config from a Python file. This function behaves
as if the file was imported as module with the from_object() function.
• filename – the filename of the config. This can either be an
absolute filename or a filename relative to the root path.
• silent – set to True if you want silent failure for missing files.
New in version 0.7: silent parameter.
get_namespace(namespace, lowercase=True, trim_namespace=True)
Returns a dictionary containing a subset of configuration options that match
the specified namespace/prefix. Example usage:
app.config['IMAGE_STORE_TYPE'] = 'fs'
app.config['IMAGE_STORE_PATH'] = '/var/app/images'
app.config['IMAGE_STORE_BASE_URL'] = 'http://img.website.com'
image_store_config = app.config.get_namespace('IMAGE_STORE_')
The resulting dictionary image_store_config would look like:
'type': 'fs',
'path': '/var/app/images',
'base_url': 'http://img.website.com'
This is often useful when configuration options map directly to keyword
arguments in functions or class constructors.
• namespace – a configuration namespace
• lowercase – a flag indicating if the keys of the resulting dictionary should be lowercase
• trim_namespace – a flag indicating if the keys of the resulting
dictionary should not include the namespace
New in version 0.11.
This module acts as redirect import module to Flask extensions. It was added in
0.8 as the canonical way to import Flask extensions and makes it possible for us
to have more flexibility in how we distribute extensions.
If you want to use an extension named “Flask-Foo” you would import it from
ext as follows:
from flask.ext import foo
New in version 0.8.
Stream Helpers
Request contexts disappear when the response is started on the server. This is
done for efficiency reasons and to make it less likely to encounter memory leaks
with badly written WSGI middlewares. The downside is that if you are using
streamed responses, the generator cannot access request bound information any
This function however can help you keep the context around for longer:
from flask import stream_with_context, request, Response
def streamed_response():
def generate():
yield 'Hello '
yield request.args['name']
yield '!'
return Response(generate())
Alternatively it can also be used around a specific generator:
from flask import stream_with_context, request, Response
def streamed_response():
def generate():
yield 'Hello '
yield request.args['name']
yield '!'
return Response(stream_with_context(generate()))
New in version 0.9.
Useful Internals
class flask.ctx.RequestContext(app, environ, request=None)
The request context contains all request relevant information. It is created at the
beginning of the request and pushed to the _request_ctx_stack and removed at
the end of it. It will create the URL adapter and request object for the WSGI
environment provided.
Do not attempt to use this class directly, instead use test_request_context()
and request_context() to create this object.
When the request context is popped, it will evaluate all the functions registered
on the application for teardown execution (teardown_request()).
The request context is automatically popped at the end of the request for you.
In debug mode the request context is kept around if exceptions happen so that
interactive debuggers have a chance to introspect the data. With 0.4 this can also
be forced for requests that did not fail and outside of DEBUG mode. By setting
'flask._preserve_context' to True on the WSGI environment the context will
not pop itself at the end of the request. This is used by the test_client() for
example to implement the deferred cleanup functionality.
You might find this helpful for unittests where you need the information from
the context local around for a little longer. Make sure to properly pop() the stack
yourself in that situation, otherwise your unittests will leak memory.
Creates a copy of this request context with the same request object. This can
be used to move a request context to a different greenlet. Because the actual
request object is the same this cannot be used to move a request context to
a different thread unless access to the request object is locked.
New in version 0.10.
Can be overridden by a subclass to hook into the matching of the request.
pop(exc=<object object>)
Pops the request context and unbinds it by doing that. This will also trigger
the execution of functions registered by the teardown_request() decorator.
Changed in version 0.9: Added the exc argument.
Binds the request context to the current context.
The internal LocalStack that is used to implement all the context local objects
used in Flask. This is a documented instance and can be used by extensions and
application code but the use is discouraged in general.
The following attributes are always present on each layer of the stack:
app the active Flask application.
url_adapter the URL adapter that was used to match the request.
request the current request object.
session the active session object.
g an object with all the attributes of the flask.g object.
flashes an internal cache for the flashed messages.
Example usage:
from flask import _request_ctx_stack
def get_session():
ctx = _request_ctx_stack.top
if ctx is not None:
return ctx.session
class flask.ctx.AppContext(app)
The application context binds an application object implicitly to the current
thread or greenlet, similar to how the RequestContext binds request information. The application context is also implicitly created if a request context is
created but the application is not on top of the individual application context.
pop(exc=<object object>)
Pops the app context.
Binds the app context to the current context.
Works similar to the request context but only binds the application. This is
mainly there for extensions to store data.
New in version 0.9.
class flask.blueprints.BlueprintSetupState(blueprint,
Temporary holder object for registering a blueprint with the application. An instance of this class is created by the make_setup_state() method and later passed
to all register callback functions.
add_url_rule(rule, endpoint=None, view_func=None, **options)
A helper method to register a rule (and optionally a view function) to the
application. The endpoint is automatically prefixed with the blueprint’s
app = None
a reference to the current application
blueprint = None
a reference to the blueprint that created this setup state.
first_registration = None
as blueprints can be registered multiple times with the application and not
everything wants to be registered multiple times on it, this attribute can be
used to figure out if the blueprint was registered in the past already.
options = None
a dictionary with all options that were passed to the register_blueprint()
subdomain = None
The subdomain that the blueprint should be active for, None otherwise.
url_defaults = None
A dictionary with URL defaults that is added to each and every URL that
was defined with the blueprint.
url_prefix = None
The prefix that should be used for all URLs defined on the blueprint.
New in version 0.6.
True if the signaling system is available. This is the case when blinker is installed.
The following signals exist in Flask:
This signal is sent when a template was successfully rendered. The signal is in254
voked with the instance of the template as template and the context as dictionary
(named context).
Example subscriber:
def log_template_renders(sender, template, context, **extra):
sender.logger.debug('Rendering template "%s" with context %s',
template.name or 'string template',
from flask import template_rendered
template_rendered.connect(log_template_renders, app)
This signal is sent before template rendering process. The signal is invoked with
the instance of the template as template and the context as dictionary (named
Example subscriber:
def log_template_renders(sender, template, context, **extra):
sender.logger.debug('Rendering template "%s" with context %s',
template.name or 'string template',
from flask import before_render_template
before_render_template.connect(log_template_renders, app)
This signal is sent when the request context is set up, before any request processing happens. Because the request context is already bound, the subscriber can
access the request with the standard global proxies such as request.
Example subscriber:
def log_request(sender, **extra):
sender.logger.debug('Request context is set up')
from flask import request_started
request_started.connect(log_request, app)
This signal is sent right before the response is sent to the client. It is passed the
response to be sent named response.
Example subscriber:
def log_response(sender, response, **extra):
sender.logger.debug('Request context is about to close down. '
'Response: %s', response)
from flask import request_finished
request_finished.connect(log_response, app)
This signal is sent when an exception happens during request processing. It is
sent before the standard exception handling kicks in and even in debug mode,
where no exception handling happens. The exception itself is passed to the subscriber as exception.
Example subscriber:
def log_exception(sender, exception, **extra):
sender.logger.debug('Got exception during processing: %s', exception)
from flask import got_request_exception
got_request_exception.connect(log_exception, app)
This signal is sent when the request is tearing down. This is always called, even
if an exception is caused. Currently functions listening to this signal are called
after the regular teardown handlers, but this is not something you can rely on.
Example subscriber:
def close_db_connection(sender, **extra):
from flask import request_tearing_down
request_tearing_down.connect(close_db_connection, app)
As of Flask 0.9, this will also be passed an exc keyword argument that has a
reference to the exception that caused the teardown if there was one.
This signal is sent when the app context is tearing down. This is always called,
even if an exception is caused. Currently functions listening to this signal are
called after the regular teardown handlers, but this is not something you can
rely on.
Example subscriber:
def close_db_connection(sender, **extra):
from flask import appcontext_tearing_down
appcontext_tearing_down.connect(close_db_connection, app)
This will also be passed an exc keyword argument that has a reference to the
exception that caused the teardown if there was one.
This signal is sent when an application context is pushed. The sender is the
application. This is usually useful for unittests in order to temporarily hook in
information. For instance it can be used to set a resource early onto the g object.
Example usage:
from contextlib import contextmanager
from flask import appcontext_pushed
def user_set(app, user):
def handler(sender, **kwargs):
g.user = user
with appcontext_pushed.connected_to(handler, app):
And in the testcode:
def test_user_me(self):
with user_set(app, 'john'):
c = app.test_client()
resp = c.get('/users/me')
assert resp.data == 'username=john'
New in version 0.10.
This signal is sent when an application context is popped. The sender is the
application. This usually falls in line with the appcontext_tearing_down signal.
New in version 0.10.
This signal is sent when the application is flashing a message. The messages is
sent as message keyword argument and the category as category.
Example subscriber:
recorded = []
def record(sender, message, category, **extra):
recorded.append((message, category))
from flask import message_flashed
message_flashed.connect(record, app)
New in version 0.10.
class signals.Namespace
An alias for blinker.base.Namespace if blinker is available, otherwise a dummy
class that creates fake signals. This class is available for Flask extensions that
want to provide the same fallback system as Flask itself.
signal(name, doc=None)
Creates a new signal for this namespace if blinker is available, otherwise
returns a fake signal that has a send method that will do nothing but will
fail with a RuntimeError for all other operations, including connecting.
Class-Based Views
New in version 0.7.
class flask.views.View
Alternative way to use view functions.
A subclass has to implement
dispatch_request() which is called with the view arguments from the URL routing system. If methods is provided the methods do not have to be passed to the
add_url_rule() method explicitly:
class MyView(View):
methods = ['GET']
def dispatch_request(self, name):
return 'Hello %s!' % name
app.add_url_rule('/hello/<name>', view_func=MyView.as_view('myview'))
When you want to decorate a pluggable view you will have to either do that
when the view function is created (by wrapping the return value of as_view())
or you can use the decorators attribute:
class SecretView(View):
methods = ['GET']
decorators = [superuser_required]
def dispatch_request(self):
The decorators stored in the decorators list are applied one after another when
the view function is created. Note that you can not use the class based decorators
since those would decorate the view class and not the generated view function!
classmethod as_view(name, *class_args, **class_kwargs)
Converts the class into an actual view function that can be used with the
routing system. Internally this generates a function on the fly which will instantiate the View on each request and call the dispatch_request() method
on it.
The arguments passed to as_view() are forwarded to the constructor of the
decorators = ()
The canonical way to decorate class-based views is to decorate the return
value of as_view(). However since this moves parts of the logic from the
class declaration to the place where it’s hooked into the routing system.
You can place one or more decorators in this list and whenever the view
function is created the result is automatically decorated.
New in version 0.8.
Subclasses have to override this method to implement the actual view function code. This method is called with all the arguments from the URL rule.
methods = None
A list of methods this view can handle.
provide_automatic_options = None
Setting this disables or force-enables the automatic options handling.
class flask.views.MethodView
A class-based view that dispatches request methods to the corresponding class
methods. For example, if you implement a get method, it will be used to handle
GET requests.
class CounterAPI(MethodView):
def get(self):
return session.get('counter', 0)
def post(self):
session['counter'] = session.get('counter', 0) + 1
return 'OK'
app.add_url_rule('/counter', view_func=CounterAPI.as_view('counter'))
URL Route Registrations
Generally there are three ways to define rules for the routing system:
1. You can use the flask.Flask.route() decorator.
2. You can use the flask.Flask.add_url_rule() function.
3. You can directly access the underlying Werkzeug routing system which is exposed as flask.Flask.url_map.
Variable parts in the route can be specified with angular brackets (/user/<username>).
By default a variable part in the URL accepts any string without a slash however a
different converter can be specified as well by using <converter:name>.
Variable parts are passed to the view function as keyword arguments.
The following converters are available:
accepts any text without a slash (the default)
accepts integers
like int but for floating point values
like the default but also accepts slashes
matches one of the items provided
accepts UUID strings
Custom converters can be defined using flask.Flask.url_map.
Here are some examples:
def index():
def show_user(username):
def show_post(post_id):
An important detail to keep in mind is how Flask deals with trailing slashes. The idea
is to keep each URL unique so the following rules apply:
1. If a rule ends with a slash and is requested without a slash by the user, the user
is automatically redirected to the same page with a trailing slash attached.
2. If a rule does not end with a trailing slash and the user requests the page with a
trailing slash, a 404 not found is raised.
This is consistent with how web servers deal with static files. This also makes it possible to use relative link targets safely.
You can also define multiple rules for the same function. They have to be unique
however. Defaults can also be specified. Here for example is a definition for a URL
that accepts an optional page:
@app.route('/users/', defaults={'page': 1})
def show_users(page):
This specifies that /users/ will be the URL for page one and /users/page/N will be the
URL for page N.
Here are the parameters that route() and add_url_rule() accept. The only difference
is that with the route parameter the view function is defined with the decorator instead
of the view_func parameter.
the URL rule as string
endthe endpoint for the registered URL rule. Flask itself assumes that the name
point of the view function is the name of the endpoint if not explicitly stated.
the function to call when serving a request to the provided endpoint. If this
is not provided one can specify the function later by storing it in the
view_functions dictionary with the endpoint as key.
deA dictionary with defaults for this rule. See the example above for how
faults defaults work.
subspecifies the rule for the subdomain in case subdomain matching is in use.
doIf not specified the default subdomain is assumed.
**op- the options to be forwarded to the underlying Rule object. A change to
tions Werkzeug is handling of method options. methods is a list of methods this
rule should be limited to (GET, POST etc.). By default a rule just listens for
GET (and implicitly HEAD). Starting with Flask 0.6, OPTIONS is implicitly
added and handled by the standard request handling. They have to be
specified as keyword arguments.
View Function Options
For internal usage the view functions can have some attributes attached to customize
behavior the view function would normally not have control over. The following attributes can be provided optionally to either override some defaults to add_url_rule()
or general behavior:
• __name__: The name of a function is by default used as endpoint. If endpoint is
provided explicitly this value is used. Additionally this will be prefixed with the
name of the blueprint by default which cannot be customized from the function
• methods: If methods are not provided when the URL rule is added, Flask will
look on the view function object itself if a methods attribute exists. If it does, it
will pull the information for the methods from there.
• provide_automatic_options: if this attribute is set Flask will either force enable or
disable the automatic implementation of the HTTP OPTIONS response. This can
be useful when working with decorators that want to customize the OPTIONS
response on a per-view basis.
• required_methods: if this attribute is set, Flask will always add these methods
when registering a URL rule even if the methods were explicitly overridden in
the route() call.
Full example:
def index():
if request.method == 'OPTIONS':
# custom options handling here
return 'Hello World!'
index.provide_automatic_options = False
index.methods = ['GET', 'OPTIONS']
app.add_url_rule('/', index)
New in version 0.8: The provide_automatic_options functionality was added.
Command Line Interface
class flask.cli.FlaskGroup(add_default_commands=True,
add_version_option=True, **extra)
Special subclass of the AppGroup group that supports loading more commands
from the configured Flask app. Normally a developer does not have to interface
with this class but there are some very advanced use cases for which it makes
sense to create an instance of this.
For information as of why this is useful see Custom Scripts.
• add_default_commands – if this is True then the default run and
shell commands wil be added.
• add_version_option – adds the --version option.
• create_app – an optional callback that is passed the script info
and returns the loaded app.
class flask.cli.AppGroup(name=None, commands=None, **attrs)
This works similar to a regular click Group but it changes the behavior
of the command() decorator so that it automatically wraps the functions in
Not to be confused with FlaskGroup.
command(*args, **kwargs)
This works exactly like the method of the same name on a regular click.
Group but it wraps callbacks in with_appcontext() unless it’s disabled by
passing with_appcontext=False.
group(*args, **kwargs)
This works exactly like the method of the same name on a regular click.
Group but it defaults the group class to AppGroup.
class flask.cli.ScriptInfo(app_import_path=None, create_app=None)
Help object to deal with Flask applications. This is usually not necessary to interface with as it’s used internally in the dispatching to click. In future versions
of Flask this object will most likely play a bigger role. Typically it’s created automatically by the FlaskGroup but you can also manually create it and pass it
onwards as click object.
app_import_path = None
Optionally the import path for the Flask application.
create_app = None
Optionally a function that is passed the script info to create the instance of
the application.
data = None
A dictionary with arbitrary data that can be associated with this script info.
Loads the Flask app (if not yet loaded) and returns it. Calling this multiple
times will just result in the already loaded app to be returned.
flask.cli.with_appcontext(f )
Wraps a callback so that it’s guaranteed to be executed with the script’s application context. If callbacks are registered directly to the app.cli object then they
are wrapped with this function by default unless it’s disabled.
flask.cli.pass_script_info(f )
Marks a function so that an instance of ScriptInfo is passed as first argument to
the click callback.
flask.cli.run_command = <click.core.Command object>
Runs a local development server for the Flask application.
This local server is recommended for development purposes only but it can also
be used for simple intranet deployments. By default it will not support any sort
of concurrency at all to simplify debugging. This can be changed with the –withthreads option which will enable basic multithreading.
The reloader and debugger are by default enabled if the debug flag of Flask is
enabled and disabled otherwise.
flask.cli.shell_command = <click.core.Command object>
Runs an interactive Python shell in the context of a given Flask application. The
application will populate the default namespace of this shell according to it’s
This is useful for executing small snippets of management code without having
to manually configuring the application.
Part III
Design notes, legal information and changelog are here for the interested.
Design Decisions in Flask
If you are curious why Flask does certain things the way it does and not differently,
this section is for you. This should give you an idea about some of the design decisions
that may appear arbitrary and surprising at first, especially in direct comparison with
other frameworks.
The Explicit Application Object
A Python web application based on WSGI has to have one central callable object that
implements the actual application. In Flask this is an instance of the Flask class. Each
Flask application has to create an instance of this class itself and pass it the name of
the module, but why can’t Flask do that itself?
Without such an explicit application object the following code:
from flask import Flask
app = Flask(__name__)
def index():
return 'Hello World!'
Would look like this instead:
from hypothetical_flask import route
def index():
return 'Hello World!'
There are three major reasons for this. The most important one is that implicit application objects require that there may only be one instance at the time. There are ways
to fake multiple applications with a single application object, like maintaining a stack
of applications, but this causes some problems I won’t outline here in detail. Now
the question is: when does a microframework need more than one application at the
same time? A good example for this is unittesting. When you want to test something
it can be very helpful to create a minimal application to test specific behavior. When
the application object is deleted everything it allocated will be freed again.
Another thing that becomes possible when you have an explicit object lying around in
your code is that you can subclass the base class (Flask) to alter specific behavior. This
would not be possible without hacks if the object were created ahead of time for you
based on a class that is not exposed to you.
But there is another very important reason why Flask depends on an explicit instantiation of that class: the package name. Whenever you create a Flask instance you usually
pass it __name__ as package name. Flask depends on that information to properly load
resources relative to your module. With Python’s outstanding support for reflection it
can then access the package to figure out where the templates and static files are stored
(see open_resource()). Now obviously there are frameworks around that do not need
any configuration and will still be able to load templates relative to your application
module. But they have to use the current working directory for that, which is a very
unreliable way to determine where the application is. The current working directory is
process-wide and if you are running multiple applications in one process (which could
happen in a webserver without you knowing) the paths will be off. Worse: many webservers do not set the working directory to the directory of your application but to the
document root which does not have to be the same folder.
The third reason is “explicit is better than implicit”. That object is your WSGI application, you don’t have to remember anything else. If you want to apply a WSGI
middleware, just wrap it and you’re done (though there are better ways to do that so
that you do not lose the reference to the application object wsgi_app()).
Furthermore this design makes it possible to use a factory function to create the application which is very helpful for unittesting and similar things (Application Factories).
The Routing System
Flask uses the Werkzeug routing system which was designed to automatically order
routes by complexity. This means that you can declare routes in arbitrary order and
they will still work as expected. This is a requirement if you want to properly implement decorator based routing since decorators could be fired in undefined order when
the application is split into multiple modules.
Another design decision with the Werkzeug routing system is that routes in Werkzeug
try to ensure that URLs are unique. Werkzeug will go quite far with that in that it will
automatically redirect to a canonical URL if a route is ambiguous.
One Template Engine
Flask decides on one template engine: Jinja2. Why doesn’t Flask have a pluggable template engine interface? You can obviously use a different template engine, but Flask
will still configure Jinja2 for you. While that limitation that Jinja2 is always configured
will probably go away, the decision to bundle one template engine and use that will
Template engines are like programming languages and each of those engines has a
certain understanding about how things work. On the surface they all work the same:
you tell the engine to evaluate a template with a set of variables and take the return
value as string.
But that’s about where similarities end. Jinja2 for example has an extensive filter system, a certain way to do template inheritance, support for reusable blocks (macros)
that can be used from inside templates and also from Python code, uses Unicode for
all operations, supports iterative template rendering, configurable syntax and more.
On the other hand an engine like Genshi is based on XML stream evaluation, template
inheritance by taking the availability of XPath into account and more. Mako on the
other hand treats templates similar to Python modules.
When it comes to connecting a template engine with an application or framework
there is more than just rendering templates. For instance, Flask uses Jinja2’s extensive
autoescaping support. Also it provides ways to access macros from Jinja2 templates.
A template abstraction layer that would not take the unique features of the template
engines away is a science on its own and a too large undertaking for a microframework
like Flask.
Furthermore extensions can then easily depend on one template language being
present. You can easily use your own templating language, but an extension could
still depend on Jinja itself.
Micro with Dependencies
Why does Flask call itself a microframework and yet it depends on two libraries
(namely Werkzeug and Jinja2). Why shouldn’t it? If we look over to the Ruby side of
web development there we have a protocol very similar to WSGI. Just that it’s called
Rack there, but besides that it looks very much like a WSGI rendition for Ruby. But
nearly all applications in Ruby land do not work with Rack directly, but on top of a
library with the same name. This Rack library has two equivalents in Python: WebOb
(formerly Paste) and Werkzeug. Paste is still around but from my understanding it’s
sort of deprecated in favour of WebOb. The development of WebOb and Werkzeug
started side by side with similar ideas in mind: be a good implementation of WSGI for
other applications to take advantage.
Flask is a framework that takes advantage of the work already done by Werkzeug to
properly interface WSGI (which can be a complex task at times). Thanks to recent
developments in the Python package infrastructure, packages with dependencies are
no longer an issue and there are very few reasons against having libraries that depend
on others.
Thread Locals
Flask uses thread local objects (context local objects in fact, they support greenlet contexts as well) for request, session and an extra object you can put your own things on
(g). Why is that and isn’t that a bad idea?
Yes it is usually not such a bright idea to use thread locals. They cause troubles for
servers that are not based on the concept of threads and make large applications harder
to maintain. However Flask is just not designed for large applications or asynchronous
servers. Flask wants to make it quick and easy to write a traditional web application.
Also see the Becoming Big section of the documentation for some inspiration for larger
applications based on Flask.
What Flask is, What Flask is Not
Flask will never have a database layer. It will not have a form library or anything else
in that direction. Flask itself just bridges to Werkzeug to implement a proper WSGI
application and to Jinja2 to handle templating. It also binds to a few common standard
library packages such as logging. Everything else is up for extensions.
Why is this the case? Because people have different preferences and requirements and
Flask could not meet those if it would force any of this into the core. The majority
of web applications will need a template engine in some sort. However not every
application needs a SQL database.
The idea of Flask is to build a good foundation for all applications. Everything else is
up to you or extensions.
The Flask documentation and example applications are using HTML5. You may notice that in many situations, when end tags are optional they are not used, so that
the HTML is cleaner and faster to load. Because there is much confusion about HTML
and XHTML among developers, this document tries to answer some of the major questions.
History of XHTML
For a while, it appeared that HTML was about to be replaced by XHTML. However,
barely any websites on the Internet are actual XHTML (which is HTML processed using XML rules). There are a couple of major reasons why this is the case. One of them
is Internet Explorer’s lack of proper XHTML support. The XHTML spec states that
XHTML must be served with the MIME type application/xhtml+xml, but Internet Explorer refuses to read files with that MIME type. While it is relatively easy to configure
Web servers to serve XHTML properly, few people do. This is likely because properly
using XHTML can be quite painful.
One of the most important causes of pain is XML’s draconian (strict and ruthless) error handling. When an XML parsing error is encountered, the browser is supposed to
show the user an ugly error message, instead of attempting to recover from the error
and display what it can. Most of the (X)HTML generation on the web is based on
non-XML template engines (such as Jinja, the one used in Flask) which do not protect
you from accidentally creating invalid XHTML. There are XML based template engines, such as Kid and the popular Genshi, but they often come with a larger runtime
overhead and are not as straightforward to use because they have to obey XML rules.
The majority of users, however, assumed they were properly using XHTML. They
wrote an XHTML doctype at the top of the document and self-closed all the necessary
tags (<br> becomes <br/> or <br></br> in XHTML). However, even if the document
properly validates as XHTML, what really determines XHTML/HTML processing in
browsers is the MIME type, which as said before is often not set properly. So the valid
XHTML was being treated as invalid HTML.
XHTML also changed the way JavaScript is used. To properly work with XHTML, programmers have to use the namespaced DOM interface with the XHTML namespace
to query for HTML elements.
History of HTML5
Development of the HTML5 specification was started in 2004 under the name “Web
Applications 1.0” by the Web Hypertext Application Technology Working Group, or
WHATWG (which was formed by the major browser vendors Apple, Mozilla, and
Opera) with the goal of writing a new and improved HTML specification, based on
existing browser behavior instead of unrealistic and backwards-incompatible specifications.
For example, in HTML4 <title/Hello/ theoretically parses exactly the same as
<title>Hello</title>. However, since people were using XHTML-like tags along the
lines of <link />, browser vendors implemented the XHTML syntax over the syntax
defined by the specification.
In 2007, the specification was adopted as the basis of a new HTML specification under
the umbrella of the W3C, known as HTML5. Currently, it appears that XHTML is
losing traction, as the XHTML 2 working group has been disbanded and HTML5 is
being implemented by all major browser vendors.
The following table gives you a quick overview of features available in HTML 4.01,
XHTML 1.1 and HTML5. (XHTML 1.0 is not included, as it was superseded by
XHTML 1.1 and the barely-used XHTML5.)
<tag/value/ == <tag>value</tag>
<br/> supported
<script/> supported
should be served as text/html
should be served as application/xhtml+xml
strict error handling
inline SVG
inline MathML
<video> tag
<audio> tag
New semantic tags like <article>
What does “strict” mean?
HTML5 has strictly defined parsing rules, but it also specifies exactly how a browser
should react to parsing errors - unlike XHTML, which simply states parsing should
abort. Some people are confused by apparently invalid syntax that still generates the
expected results (for example, missing end tags or unquoted attribute values).
Some of these work because of the lenient error handling most browsers use when
they encounter a markup error, others are actually specified. The following constructs
are optional in HTML5 by standard, but have to be supported by browsers:
• Wrapping the document in an <html> tag
• Wrapping header elements in <head> or the body elements in <body>
• Closing the <p>, <li>, <dt>, <dd>, <tr>, <td>, <th>, <tbody>, <thead>, or <tfoot>
• Quoting attributes, so long as they contain no whitespace or special characters
(like <, >, ', or ").
• Requiring boolean attributes to have a value.
This means the following page in HTML5 is perfectly valid:
<!doctype html>
<title>Hello HTML5</title>
This is an obscure feature inherited from SGML. It is usually not supported by browsers, for reasons
detailed above.
2 This is for compatibility with server code that generates XHTML for tags such as <br>. It should
not be used in new code.
3 XHTML 1.0 is the last XHTML standard that allows to be served as text/html for backwards compatibility reasons.
<div class=header>
<h1>Hello HTML5</h1>
<p class=tagline>HTML5 is awesome
<ul class=nav>
<li><a href=/index>Index</a>
<li><a href=/downloads>Downloads</a>
<li><a href=/about>About</a>
<div class=body>
<h2>HTML5 is probably the future</h2>
There might be some other things around but in terms of
browser vendor support, HTML5 is hard to beat.
<dt>Key 1
<dd>Value 1
<dt>Key 2
<dd>Value 2
New technologies in HTML5
HTML5 adds many new features that make Web applications easier to write and to
• The <audio> and <video> tags provide a way to embed audio and video without
complicated add-ons like QuickTime or Flash.
• Semantic elements like <article>, <header>, <nav>, and <time> that make content easier to understand.
• The <canvas> tag, which supports a powerful drawing API, reducing the need
for server-generated images to present data graphically.
• New form control types like <input type="date"> that allow user agents to make
entering and validating values easier.
• Advanced JavaScript APIs like Web Storage, Web Workers, Web Sockets, geolocation, and offline applications.
Many other features have been added, as well. A good guide to new features in
HTML5 is Mark Pilgrim’s soon-to-be-published book, Dive Into HTML5. Not all of
them are supported in browsers yet, however, so use caution.
What should be used?
Currently, the answer is HTML5. There are very few reasons to use XHTML considering the latest developments in Web browsers. To summarize the reasons given above:
• Internet Explorer (which, sadly, currently leads in market share) has poor support for XHTML.
• Many JavaScript libraries also do not support XHTML, due to the more complicated namespacing API it requires.
• HTML5 adds several new features, including semantic tags and the long-awaited
<audio> and <video> tags.
• It has the support of most browser vendors behind it.
• It is much easier to write, and more compact.
For most applications, it is undoubtedly better to use HTML5 than XHTML.
Security Considerations
Web applications usually face all kinds of security problems and it’s very hard to get
everything right. Flask tries to solve a few of these things for you, but there are a
couple more you have to take care of yourself.
Cross-Site Scripting (XSS)
Cross site scripting is the concept of injecting arbitrary HTML (and with it JavaScript)
into the context of a website. To remedy this, developers have to properly escape text
so that it cannot include arbitrary HTML tags. For more information on that have a
look at the Wikipedia article on Cross-Site Scripting.
Flask configures Jinja2 to automatically escape all values unless explicitly told otherwise. This should rule out all XSS problems caused in templates, but there are still
other places where you have to be careful:
• generating HTML without the help of Jinja2
• calling Markup on data submitted by users
• sending out HTML from uploaded files, never do that,
Content-Disposition: attachment header to prevent that problem.
use the
• sending out textfiles from uploaded files. Some browsers are using content-type
guessing based on the first few bytes so users could trick a browser to execute
Another thing that is very important are unquoted attributes. While Jinja2 can protect
you from XSS issues by escaping HTML, there is one thing it cannot protect you from:
XSS by attribute injection. To counter this possible attack vector, be sure to always
quote your attributes with either double or single quotes when using Jinja expressions
in them:
<a href="{{ href }}">the text</a>
Why is this necessary? Because if you would not be doing that, an attacker could easily
inject custom JavaScript handlers. For example an attacker could inject this piece of
When the user would then move with the mouse over the link, the cookie would be
presented to the user in an alert window. But instead of showing the cookie to the
user, a good attacker might also execute any other JavaScript code. In combination
with CSS injections the attacker might even make the element fill out the entire page
so that the user would just have to have the mouse anywhere on the page to trigger
the attack.
Cross-Site Request Forgery (CSRF)
Another big problem is CSRF. This is a very complex topic and I won’t outline it here
in detail just mention what it is and how to theoretically prevent it.
If your authentication information is stored in cookies, you have implicit state management. The state of “being logged in” is controlled by a cookie, and that cookie is
sent with each request to a page. Unfortunately that includes requests triggered by
3rd party sites. If you don’t keep that in mind, some people might be able to trick your
application’s users with social engineering to do stupid things without them knowing.
Say you have a specific URL that, when you sent POST requests to will delete a user’s
profile (say http://example.com/user/delete). If an attacker now creates a page that
sends a post request to that page with some JavaScript they just have to trick some
users to load that page and their profiles will end up being deleted.
Imagine you were to run Facebook with millions of concurrent users and someone
would send out links to images of little kittens. When users would go to that page,
their profiles would get deleted while they are looking at images of fluffy cats.
How can you prevent that? Basically for each request that modifies content on the
server you would have to either use a one-time token and store that in the cookie and
also transmit it with the form data. After receiving the data on the server again, you
would then have to compare the two tokens and ensure they are equal.
Why does Flask not do that for you? The ideal place for this to happen is the form
validation framework, which does not exist in Flask.
JSON Security
In Flask 0.10 and lower, jsonify() did not serialize top-level arrays to JSON. This was
because of a security vulnerability in ECMAScript 4.
ECMAScript 5 closed this vulnerability, so only extremely old browsers are still vulnerable. All of these browsers have other more serious vulnerabilities, so this behavior
was changed and jsonify() now supports serializing arrays.
Unicode in Flask
Flask, like Jinja2 and Werkzeug, is totally Unicode based when it comes to text. Not
only these libraries, also the majority of web related Python libraries that deal with
text. If you don’t know Unicode so far, you should probably read The Absolute Minimum Every Software Developer Absolutely, Positively Must Know About Unicode
and Character Sets. This part of the documentation just tries to cover the very basics
so that you have a pleasant experience with Unicode related things.
Automatic Conversion
Flask has a few assumptions about your application (which you can change of course)
that give you basic and painless Unicode support:
• the encoding for text on your website is UTF-8
• internally you will always use Unicode exclusively for text except for literal
strings with only ASCII character points.
• encoding and decoding happens whenever you are talking over a protocol that
requires bytes to be transmitted.
So what does this mean to you?
HTTP is based on bytes. Not only the protocol, also the system used to address documents on servers (so called URIs or URLs). However HTML which is usually transmitted on top of HTTP supports a large variety of character sets and which ones are
used, are transmitted in an HTTP header. To not make this too complex Flask just
assumes that if you are sending Unicode out you want it to be UTF-8 encoded. Flask
will do the encoding and setting of the appropriate headers for you.
The same is true if you are talking to databases with the help of SQLAlchemy or a
similar ORM system. Some databases have a protocol that already transmits Unicode
and if they do not, SQLAlchemy or your other ORM should take care of that.
The Golden Rule
So the rule of thumb: if you are not dealing with binary data, work with Unicode.
What does working with Unicode in Python 2.x mean?
• as long as you are using ASCII charpoints only (basically numbers, some special
characters of latin letters without umlauts or anything fancy) you can use regular
string literals ('Hello World').
• if you need anything else than ASCII in a string you have to mark this string as
Unicode string by prefixing it with a lowercase u. (like u'Hänsel und Gretel')
• if you are using non-Unicode characters in your Python files you have to tell
Python which encoding your file uses. Again, I recommend UTF-8 for this purpose. To tell the interpreter your encoding you can put the # -*- coding: utf-8
-*- into the first or second line of your Python source file.
• Jinja is configured to decode the template files from UTF-8. So make sure to tell
your editor to save the file as UTF-8 there as well.
Encoding and Decoding Yourself
If you are talking with a filesystem or something that is not really based on Unicode
you will have to ensure that you decode properly when working with Unicode interface. So for example if you want to load a file on the filesystem and embed it into a
Jinja2 template you will have to decode it from the encoding of that file. Here the old
problem that text files do not specify their encoding comes into play. So do yourself a
favour and limit yourself to UTF-8 for text files as well.
Anyways. To load such a file with Unicode you can use the built-in str.decode()
def read_file(filename, charset='utf-8'):
with open(filename, 'r') as f:
return f.read().decode(charset)
To go from Unicode into a specific charset such as UTF-8 you can use the unicode.
encode() method:
def write_file(filename, contents, charset='utf-8'):
with open(filename, 'w') as f:
Configuring Editors
Most editors save as UTF-8 by default nowadays but in case your editor is not configured to do this you have to change it. Here some common ways to set your editor to
store as UTF-8:
• Vim: put set enc=utf-8 to your .vimrc file.
• Emacs: either use an encoding cookie or put this into your .emacs file:
(prefer-coding-system 'utf-8)
(setq default-buffer-file-coding-system 'utf-8)
• Notepad++:
1. Go to Settings -> Preferences ...
2. Select the “New Document/Default Directory” tab
3. Select “UTF-8 without BOM” as encoding
It is also recommended to use the Unix newline format, you can select it in the
same panel but this is not a requirement.
Flask Extension Development
Flask, being a microframework, often requires some repetitive steps to get a third party
library working. Because very often these steps could be abstracted to support multiple projects the Flask Extension Registry was created.
If you want to create your own Flask extension for something that does not exist yet,
this guide to extension development will help you get your extension running in no
time and to feel like users would expect your extension to behave.
Anatomy of an Extension
Extensions are all located in a package called flask_something where “something” is
the name of the library you want to bridge. So for example if you plan to add support
for a library named simplexml to Flask, you would name your extension’s package
The name of the actual extension (the human readable name) however would be something like “Flask-SimpleXML”. Make sure to include the name “Flask” somewhere in
that name and that you check the capitalization. This is how users can then register
dependencies to your extension in their setup.py files.
But what do extensions look like themselves? An extension has to ensure that it works
with multiple Flask application instances at once. This is a requirement because many
people will use patterns like the Application Factories pattern to create their application
as needed to aid unittests and to support multiple configurations. Because of that it is
crucial that your application supports that kind of behavior.
Most importantly the extension must be shipped with a setup.py file and registered
on PyPI. Also the development checkout link should work so that people can easily
install the development version into their virtualenv without having to download the
library by hand.
Flask extensions must be licensed under a BSD, MIT or more liberal license to be able
to be enlisted in the Flask Extension Registry. Keep in mind that the Flask Extension
Registry is a moderated place and libraries will be reviewed upfront if they behave as
“Hello Flaskext!”
So let’s get started with creating such a Flask extension. The extension we want to
create here will provide very basic support for SQLite3.
First we create the following folder structure:
Here’s the contents of the most important files:
The next file that is absolutely required is the setup.py file which is used to install
your Flask extension. The following contents are something you can work with:
------------This is the description for that library
from setuptools import setup
author='Your Name',
author_email='[email protected]',
description='Very short description',
# if you would be using a package instead use packages instead
# of py_modules:
# packages=['flask_sqlite3'],
'Environment :: Web Environment',
'Intended Audience :: Developers',
'License :: OSI Approved :: BSD License',
'Operating System :: OS Independent',
'Programming Language :: Python',
'Topic :: Internet :: WWW/HTTP :: Dynamic Content',
'Topic :: Software Development :: Libraries :: Python Modules'
That’s a lot of code but you can really just copy/paste that from existing extensions
and adapt.
Now this is where your extension code goes. But how exactly should such an extension look like? What are the best practices? Continue reading for some insight.
Initializing Extensions
Many extensions will need some kind of initialization step. For example, consider an
application that’s currently connecting to SQLite like the documentation suggests (Using SQLite 3 with Flask). So how does the extension know the name of the application
Quite simple: you pass it to it.
There are two recommended ways for an extension to initialize:
initialization functions:
If your extension is called helloworld you might have a function called
init_helloworld(app[, extra_args]) that initializes the extension for that
application. It could attach before / after handlers etc.
Classes work mostly like initialization functions but can later be used to
further change the behavior. For an example look at how the OAuth extension works: there is an OAuth object that provides some helper functions
like OAuth.remote_app to create a reference to a remote application that uses
What to use depends on what you have in mind. For the SQLite 3 extension we will
use the class-based approach because it will provide users with an object that handles
opening and closing database connections.
What’s important about classes is that they encourage to be shared around on module level. In that case, the object itself must not under any circumstances store any
application specific state and must be shareable between different application.
The Extension Code
Here’s the contents of the flask_sqlite3.py for copy/paste:
import sqlite3
from flask import current_app
# Find the stack on which we want to store the database connection.
# Starting with Flask 0.9, the _app_ctx_stack is the correct one,
# before that we need to use the _request_ctx_stack.
from flask import _app_ctx_stack as stack
except ImportError:
from flask import _request_ctx_stack as stack
class SQLite3(object):
def __init__(self, app=None):
self.app = app
if app is not None:
def init_app(self, app):
app.config.setdefault('SQLITE3_DATABASE', ':memory:')
# Use the newstyle teardown_appcontext if it's available,
# otherwise fall back to the request context
if hasattr(app, 'teardown_appcontext'):
def connect(self):
return sqlite3.connect(current_app.config['SQLITE3_DATABASE'])
def teardown(self, exception):
ctx = stack.top
if hasattr(ctx, 'sqlite3_db'):
def connection(self):
ctx = stack.top
if ctx is not None:
if not hasattr(ctx, 'sqlite3_db'):
ctx.sqlite3_db = self.connect()
return ctx.sqlite3_db
So here’s what these lines of code do:
1. The __init__ method takes an optional app object and, if supplied, will call
2. The init_app method exists so that the SQLite3 object can be instantiated without requiring an app object. This method supports the factory pattern for creating applications. The init_app will set the configuration for the database, defaulting to an in memory database if no configuration is supplied. In addition,
the init_app method attaches the teardown handler. It will try to use the newstyle app context handler and if it does not exist, falls back to the request context
3. Next, we define a connect method that opens a database connection.
4. Finally, we add a connection property that on first access opens the database
connection and stores it on the context. This is also the recommended way to
handling resources: fetch resources lazily the first time they are used.
Note here that we’re attaching our database connection to the top application
context via _app_ctx_stack.top. Extensions should use the top context for storing their own information with a sufficiently complex name. Note that we’re
falling back to the _request_ctx_stack.top if the application is using an older
version of Flask that does not support it.
So why did we decide on a class-based approach here? Because using our extension
looks something like this:
from flask import Flask
from flask_sqlite3 import SQLite3
app = Flask(__name__)
db = SQLite3(app)
You can then use the database from views like this:
def show_all():
cur = db.connection.cursor()
Likewise if you are outside of a request but you are using Flask 0.9 or later with the
app context support, you can use the database in the same way:
with app.app_context():
cur = db.connection.cursor()
At the end of the with block the teardown handles will be executed automatically.
Additionally, the init_app method is used to support the factory pattern for creating
db = Sqlite3()
# Then later on.
app = create_app('the-config.cfg')
Keep in mind that supporting this factory pattern for creating apps is required for
approved flask extensions (described below).
Note on init_app
As you noticed, init_app does not assign app to self. This is intentional! Class based
Flask extensions must only store the application on the object when the application
was passed to the constructor. This tells the extension: I am not interested in using
multiple applications.
When the extension needs to find the current application and it does not have a reference to it, it must either use the current_app context local or change the API in a way
that you can pass the application explicitly.
Using _app_ctx_stack
In the example above, before every request, a sqlite3_db variable is assigned
to _app_ctx_stack.top. In a view function, this variable is accessible using the
connection property of SQLite3. During the teardown of a request, the sqlite3_db
connection is closed. By using this pattern, the same connection to the sqlite3 database
is accessible to anything that needs it for the duration of the request.
If the _app_ctx_stack does not exist because the user uses an old version of Flask, it is
recommended to fall back to _request_ctx_stack which is bound to a request.
Teardown Behavior
This is only relevant if you want to support Flask 0.6 and older
Due to the change in Flask 0.7 regarding functions that are run at the end of the request
your extension will have to be extra careful there if it wants to continue to support
older versions of Flask. The following pattern is a good way to support both:
def close_connection(response):
ctx = _request_ctx_stack.top
return response
if hasattr(app, 'teardown_request'):
Strictly speaking the above code is wrong, because teardown functions are passed the
exception and typically don’t return anything. However because the return value is
discarded this will just work assuming that the code in between does not touch the
passed parameter.
Learn from Others
This documentation only touches the bare minimum for extension development. If
you want to learn more, it’s a very good idea to check out existing extensions on the
Flask Extension Registry. If you feel lost there is still the mailinglist and the IRC channel to get some ideas for nice looking APIs. Especially if you do something nobody
before you did, it might be a very good idea to get some more input. This not only
to get an idea about what people might want to have from an extension, but also to
avoid having multiple developers working on pretty much the same side by side.
Remember: good API design is hard, so introduce your project on the mailinglist, and
let other developers give you a helping hand with designing the API.
The best Flask extensions are extensions that share common idioms for the API. And
this can only work if collaboration happens early.
Approved Extensions
Flask also has the concept of approved extensions. Approved extensions are tested as
part of Flask itself to ensure extensions do not break on new releases. These approved
extensions are listed on the Flask Extension Registry and marked appropriately. If you
want your own extension to be approved you have to follow these guidelines:
0. An approved Flask extension requires a maintainer. In the event an extension
author would like to move beyond the project, the project should find a new
maintainer including full source hosting transition and PyPI access. If no maintainer is available, give access to the Flask core team.
1. An approved Flask extension must provide exactly one package or module
named flask_extensionname.
2. It must ship a testing suite that can either be invoked with make test or python
setup.py test. For test suites invoked with make test the extension has to
ensure that all dependencies for the test are installed automatically. If tests are
invoked with python setup.py test, test dependencies can be specified in the
setup.py file. The test suite also has to be part of the distribution.
3. APIs of approved extensions will be checked for the following characteristics:
• an approved extension has to support multiple applications running in the
same Python process.
• it must be possible to use the factory pattern for creating applications.
4. The license must be BSD/MIT/WTFPL licensed.
5. The naming scheme for official extensions is Flask-ExtensionName or
6. Approved extensions must define all their dependencies in the setup.py file unless a dependency cannot be met because it is not available on PyPI.
7. The extension must have documentation that uses one of the two Flask themes
for Sphinx documentation.
8. The setup.py description (and thus the PyPI description) has to link to the documentation, website (if there is one) and there must be a link to automatically
install the development version (PackageName==dev).
9. The zip_safe flag in the setup script must be set to False, even if the extension
would be safe for zipping.
10. An extension currently has to support Python 2.6 as well as Python 2.7
Extension Import Transition
In early versions of Flask we recommended using namespace packages for Flask extensions, of the form flaskext.foo. This turned out to be problematic in practice because it meant that multiple flaskext packages coexist. Consequently we have recommended to name extensions flask_foo over flaskext.foo for a long time.
Flask 0.8 introduced a redirect import system as a compatibility aid for app developers: Importing flask.ext.foo would try flask_foo and flaskext.foo in that order.
As of Flask 0.11, most Flask extensions have transitioned to the new naming schema.
The flask.ext.foo compatibility alias is still in Flask 0.11 but is now deprecated – you
should use flask_foo.
Pocoo Styleguide
The Pocoo styleguide is the styleguide for all Pocoo Projects, including Flask. This
styleguide is a requirement for Patches to Flask and a recommendation for Flask extensions.
In general the Pocoo Styleguide closely follows PEP 8 with some small differences and
General Layout
Indentation: 4 real spaces. No tabs, no exceptions.
Maximum line length: 79 characters with a soft limit for 84 if absolutely necessary.
Try to avoid too nested code by cleverly placing break, continue and return statements.
Continuing long statements: To continue a statement you can use backslashes in
which case you should align the next line with the last dot or equal sign, or
indent four spaces:
this_is_a_very_long(function_call, 'with many parameters') \
MyModel.query.filter(MyModel.scalar > 120) \
.order_by(MyModel.name.desc()) \
If you break in a statement with parentheses or braces, align to the braces:
this_is_a_very_long(function_call, 'with many parameters',
23, 42, 'and even more')
For lists or tuples with many items, break immediately after the opening brace:
items = [
'this is the first', 'set of items', 'with more items',
'to come in this line', 'like this'
Blank lines: Top level functions and classes are separated by two lines, everything
else by one. Do not use too many blank lines to separate logical segments in
code. Example:
def hello(name):
print 'Hello %s!' % name
def goodbye(name):
print 'See you %s.' % name
class MyClass(object):
"""This is a simple docstring"""
def __init__(self, name):
self.name = name
def get_annoying_name(self):
return self.name.upper() + '!!!!111'
Expressions and Statements
General whitespace rules:
• No whitespace for unary operators that are not words (e.g.: -, ~ etc.) as well
on the inner side of parentheses.
• Whitespace is placed between binary operators.
exp =
= (item_value / item_count) * offset / exp
= my_list[index]
= my_dict['key']
exp = value =
value =
value =
value =
( item_value / item_count ) * offset / exp
item_value/item_count ) * offset/exp
my_list[ index ]
my_dict ['key']
Yoda statements are a no-go: Never compare constant with variable, always variable
with constant:
if method == 'md5':
if 'md5' == method:
• against arbitrary types: == and !=
• against singletons with is and is not (eg: foo is not None)
• never compare something with True or False (for example never do foo ==
False, do not foo instead)
Negated containment checks: use foo not in bar instead of not foo in bar
Instance checks: isinstance(a, C) instead of type(A) is C, but try to avoid instance
checks in general. Check for features.
Naming Conventions
• Class names: CamelCase, with acronyms kept uppercase (HTTPWriter and not
• Variable names: lowercase_with_underscores
• Method and function names: lowercase_with_underscores
• precompiled regular expressions: name_re
Protected members are prefixed with a single underscore. Double underscores are
reserved for mixin classes.
On classes with keywords, trailing underscores are appended. Clashes with builtins
are allowed and must not be resolved by appending an underline to the variable name.
If the function needs to access a shadowed builtin, rebind the builtin to a different
name instead.
Function and method arguments:
• class methods: cls as first parameter
• instance methods: self as first parameter
• lambdas for properties might have the first parameter replaced with x like
in display_name = property(lambda x: x.real_name or x.username)
Docstring conventions: All docstrings are formatted with reStructuredText as understood by Sphinx. Depending on the number of lines in the docstring, they are
laid out differently. If it’s just one line, the closing triple quote is on the same line
as the opening, otherwise the text is on the same line as the opening quote and
the triple quote that closes the string on its own line:
def foo():
"""This is a simple docstring"""
def bar():
"""This is a longer docstring with so much information in there
that it spans three lines. In this case the closing triple quote
is on its own line.
Module header: The module header consists of a utf-8 encoding declaration (if non
ASCII letters are used, but it is recommended all the time) and a standard docstring:
# -*- coding: utf-8 -*"""
A brief description goes here.
:copyright: (c) YEAR by AUTHOR.
:license: LICENSE_NAME, see LICENSE_FILE for more details.
Please keep in mind that proper copyrights and license files are a requirement
for approved Flask extensions.
Rules for comments are similar to docstrings. Both are formatted with reStructuredText. If a comment is used to document an attribute, put a colon after the opening
pound sign (#):
class User(object):
#: the name of the user as unicode string
name = Column(String)
#: the sha1 hash of the password + inline salt
pw_hash = Column(String)
Python 3 Support
Flask, its dependencies, and most Flask extensions support Python 3. You should start
using Python 3 for your next project, but there are a few things to be aware of.
You need to use Python 3.3 or higher. 3.2 and older are not supported.
You should use the latest versions of all Flask-related packages. Flask 0.10 and
Werkzeug 0.9 were the first versions to introduce Python 3 support.
Python 3 changed how unicode and bytes are handled, which complicates how low
level code handles HTTP data. This mainly affects WSGI middleware interacting with
the WSGI environ data. Werkzeug wraps that information in high-level helpers, so
encoding issues should not affect you.
The majority of the upgrade work is in the lower-level libraries like Flask and
Werkzeug, not the high-level application code. For example, all of the examples in
the Flask repository work on both Python 2 and 3 and did not require a single line of
code changed.
Upgrading to Newer Releases
Flask itself is changing like any software is changing over time. Most of the changes
are the nice kind, the kind where you don’t have to change anything in your code to
profit from a new release.
However every once in a while there are changes that do require some changes in
your code or there are changes that make it possible for you to improve your own
code quality by taking advantage of new features in Flask.
This section of the documentation enumerates all the changes in Flask from release to
release and how you can change your code to have a painless updating experience.
Use the pip command to upgrade your existing Flask installation by providing the
--upgrade parameter:
$ pip install --upgrade Flask
Version 0.12
Changes to send_file
The filename is no longer automatically inferred from file-like objects. This means
that the following code will no longer automatically have X-Sendfile support, etag
generation or MIME-type guessing:
response = send_file(open('/path/to/file.txt'))
Any of the following is functionally equivalent:
fname = '/path/to/file.txt'
# Just pass the filepath directly
response = send_file(fname)
# Set the MIME-type and ETag explicitly
response = send_file(open(fname), mimetype='text/plain')
# Set `attachment_filename` for MIME-type guessing
# ETag still needs to be manually set
response = send_file(open(fname), attachment_filename=fname)
The reason for this is that some file-like objects have an invalid or even misleading
name attribute. Silently swallowing errors in such cases was not a satisfying solution.
Additionally the default of falling back to application/octet-stream has been restricted. If Flask can’t guess one or the user didn’t provide one, the function fails if
no filename information was provided.
Version 0.11
0.11 is an odd release in the Flask release cycle because it was supposed to be the 1.0
release. However because there was such a long lead time up to the release we decided
to push out a 0.11 release first with some changes removed to make the transition
easier. If you have been tracking the master branch which was 1.0 you might see some
unexpected changes.
In case you did track the master branch you will notice that flask --app is removed
now. You need to use the environment variable to specify an application.
Flask 0.11 removed the debug_log_format attribute from Flask applications. Instead
the new LOGGER_HANDLER_POLICY configuration can be used to disable the default log
handlers and custom log handlers can be set up.
Error handling
The behavior of error handlers was changed. The precedence of handlers used to be
based on the decoration/call order of errorhandler() and register_error_handler(),
respectively. Now the inheritance hierarchy takes precedence and handlers for more
specific exception classes are executed instead of more general ones. See Error handlers
for specifics.
Trying to register a handler on an instance now raises ValueError.
Note: There used to be a logic error allowing you to register handlers only for exception instances. This was unintended and plain wrong, and therefore was replaced with
the intended behavior of registering handlers only using exception classes and HTTP
error codes.
The render_template_string() function has changed to autoescape template variables by default. This better matches the behavior of render_template().
Extension imports
Extension imports of the form flask.ext.foo are deprecated, you should use
The old form still works, but Flask will issue a flask.exthook.ExtDeprecationWarning
for each extension you import the old way. We also provide a migration utility called
flask-ext-migrate that is supposed to automatically rewrite your imports for this.
Version 0.10
The biggest change going from 0.9 to 0.10 is that the cookie serialization format
changed from pickle to a specialized JSON format. This change has been done in order
to avoid the damage an attacker can do if the secret key is leaked. When you upgrade
you will notice two major changes: all sessions that were issued before the upgrade
are invalidated and you can only store a limited amount of types in the session. The
new sessions are by design much more restricted to only allow JSON with a few small
extensions for tuples and strings with HTML markup.
In order to not break people’s sessions it is possible to continue using the old session
system by using the Flask-OldSessions extension.
Flask also started storing the flask.g object on the application context instead of the
request context. This change should be transparent for you but it means that you now
can store things on the g object when there is no request context yet but an application context. The old flask.Flask.request_globals_class attribute was renamed to
Version 0.9
The behavior of returning tuples from a function was simplified. If you return a tuple
it no longer defines the arguments for the response object you’re creating, it’s now
always a tuple in the form (response, status, headers) where at least one item has
to be provided. If you depend on the old behavior, you can add it easily by subclassing
class TraditionalFlask(Flask):
def make_response(self, rv):
if isinstance(rv, tuple):
return self.response_class(*rv)
return Flask.make_response(self, rv)
If you maintain an extension that was using _request_ctx_stack before, please consider changing to _app_ctx_stack if it makes sense for your extension. For instance,
the app context stack makes sense for extensions which connect to databases. Using
the app context stack instead of the request context stack will make extensions more
readily handle use cases outside of requests.
Version 0.8
Flask introduced a new session interface system. We also noticed that there was a
naming collision between flask.session the module that implements sessions and
flask.session which is the global session object. With that introduction we moved
the implementation details for the session system into a new module called flask.
sessions. If you used the previously undocumented session support we urge you to
If invalid JSON data was submitted Flask will now raise a BadRequest exception instead of letting the default ValueError bubble up. This has the advantage that you no
longer have to handle that error to avoid an internal server error showing up for the
user. If you were catching this down explicitly in the past as ValueError you will need
to change this.
Due to a bug in the test client Flask 0.7 did not trigger teardown handlers when the
test client was used in a with statement. This was since fixed but might require some
changes in your test suites if you relied on this behavior.
Version 0.7
In Flask 0.7 we cleaned up the code base internally a lot and did some backwards incompatible changes that make it easier to implement larger applications with Flask.
Because we want to make upgrading as easy as possible we tried to counter the problems arising from these changes by providing a script that can ease the transition.
The script scans your whole application and generates a unified diff with changes it
assumes are safe to apply. However as this is an automated tool it won’t be able to
find all use cases and it might miss some. We internally spread a lot of deprecation
warnings all over the place to make it easy to find pieces of code that it was unable to
We strongly recommend that you hand review the generated patchfile and only apply
the chunks that look good.
If you are using git as version control system for your project we recommend applying
the patch with path -p1 < patchfile.diff and then using the interactive commit
feature to only apply the chunks that look good.
To apply the upgrade script do the following:
1. Download the script: flask-07-upgrade.py
2. Run it in the directory of your application:
python flask-07-upgrade.py > patchfile.diff
3. Review the generated patchfile.
4. Apply the patch:
patch -p1 < patchfile.diff
5. If you were using per-module template folders you need to move some templates
around. Previously if you had a folder named templates next to a blueprint
named admin the implicit template path automatically was admin/index.html for
a template file called templates/index.html. This no longer is the case. Now you
need to name the template templates/admin/index.html. The tool will not detect
this so you will have to do that on your own.
Please note that deprecation warnings are disabled by default starting with Python 2.7.
In order to see the deprecation warnings that might be emitted you have to enabled
them with the warnings module.
If you are working with windows and you lack the patch command line utility you can
get it as part of various Unix runtime environments for windows including cygwin,
msysgit or ming32. Also source control systems like svn, hg or git have builtin support
for applying unified diffs as generated by the tool. Check the manual of your version
control system for more information.
Bug in Request Locals
Due to a bug in earlier implementations the request local proxies now raise a
RuntimeError instead of an AttributeError when they are unbound. If you
caught these exceptions with AttributeError before, you should catch them with
RuntimeError now.
Additionally the send_file() function is now issuing deprecation warnings if you
depend on functionality that will be removed in Flask 0.11. Previously it was possible
to use etags and mimetypes when file objects were passed. This was unreliable and
caused issues for a few setups. If you get a deprecation warning, make sure to update
your application to work with either filenames there or disable etag attaching and
attach them yourself.
Old code:
return send_file(my_file_object)
return send_file(my_file_object)
New code:
return send_file(my_file_object, add_etags=False)
Upgrading to new Teardown Handling
We streamlined the behavior of the callbacks for request handling. For things that
modify the response the after_request() decorators continue to work as expected,
but for things that absolutely must happen at the end of request we introduced the new
teardown_request() decorator. Unfortunately that change also made after-request
work differently under error conditions. It’s not consistently skipped if exceptions
happen whereas previously it might have been called twice to ensure it is executed at
the end of the request.
If you have database connection code that looks like this:
def after_request(response):
return response
You are now encouraged to use this instead:
def after_request(exception):
if hasattr(g, 'db'):
On the upside this change greatly improves the internal code flow and makes it easier
to customize the dispatching and error handling. This makes it now a lot easier to
write unit tests as you can prevent closing down of database connections for a while.
You can take advantage of the fact that the teardown callbacks are called when the response context is removed from the stack so a test can query the database after request
with app.test_client() as client:
resp = client.get('/')
# g.db is still bound if there is such a thing
# and here it's gone
Manual Error Handler Attaching
While it is still possible to attach error handlers to Flask.error_handlers it’s discouraged to do so and in fact deprecated. In general we no longer recommend custom
error handler attaching via assignments to the underlying dictionary due to the more
complex internal handling to support arbitrary exception classes and blueprints. See
Flask.errorhandler() for more information.
The proper upgrade is to change this:
app.error_handlers[403] = handle_error
Into this:
app.register_error_handler(403, handle_error)
Alternatively you should just attach the function with a decorator:
def handle_error(e):
(Note that register_error_handler() is new in Flask 0.7)
Blueprint Support
Blueprints replace the previous concept of “Modules” in Flask. They provide better
semantics for various features and work better with large applications. The update
script provided should be able to upgrade your applications automatically, but there
might be some cases where it fails to upgrade. What changed?
• Blueprints need explicit names. Modules had an automatic name guessing
scheme where the shortname for the module was taken from the last part of
the import module. The upgrade script tries to guess that name but it might fail
as this information could change at runtime.
• Blueprints have an inverse behavior for url_for(). Previously .foo told
url_for() that it should look for the endpoint foo on the application. Now it
means “relative to current module”. The script will inverse all calls to url_for()
automatically for you. It will do this in a very eager way so you might end up
with some unnecessary leading dots in your code if you’re not using modules.
• Blueprints do not automatically provide static folders. They will also no longer
automatically export templates from a folder called templates next to their location however but it can be enabled from the constructor. Same with static files: if
you want to continue serving static files you need to tell the constructor explicitly the path to the static folder (which can be relative to the blueprint’s module
• Rendering templates was simplified. Now the blueprints can provide template
folders which are added to a general template searchpath. This means that you
need to add another subfolder with the blueprint’s name into that folder if you
want blueprintname/template.html as the template name.
If you continue to use the Module object which is deprecated, Flask will restore the
previous behavior as good as possible. However we strongly recommend upgrading
to the new blueprints as they provide a lot of useful improvement such as the ability
to attach a blueprint multiple times, blueprint specific error handlers and a lot more.
Version 0.6
Flask 0.6 comes with a backwards incompatible change which affects the order of afterrequest handlers. Previously they were called in the order of the registration, now
they are called in reverse order. This change was made so that Flask behaves more
like people expected it to work and how other systems handle request pre- and postprocessing. If you depend on the order of execution of post-request functions, be sure
to change the order.
Another change that breaks backwards compatibility is that context processors will
no longer override values passed directly to the template rendering function. If for
example request is as variable passed directly to the template, the default context
processor will not override it with the current request object. This makes it easier to
extend context processors later to inject additional variables without breaking existing
template not expecting them.
Version 0.5
Flask 0.5 is the first release that comes as a Python package instead of a single module. There were a couple of internal refactoring so if you depend on undocumented
internal details you probably have to adapt the imports.
The following changes may be relevant to your application:
• autoescaping no longer happens for all templates. Instead it is configured to only
happen on files ending with .html, .htm, .xml and .xhtml. If you have templates
with different extensions you should override the select_jinja_autoescape()
• Flask no longer supports zipped applications in this release. This functionality
might come back in future releases if there is demand for this feature. Removing
support for this makes the Flask internal code easier to understand and fixes a
couple of small issues that make debugging harder than necessary.
• The create_jinja_loader function is gone. If you want to customize the Jinja
loader now, use the create_jinja_environment() method instead.
Version 0.4
For application developers there are no changes that require changes in your code.
In case you are developing on a Flask extension however, and that extension has a
unittest-mode you might want to link the activation of that mode to the new TESTING
Version 0.3
Flask 0.3 introduces configuration support and logging as well as categories for flashing messages. All these are features that are 100% backwards compatible but you
might want to take advantage of them.
Configuration Support
The configuration support makes it easier to write any kind of application that requires
some sort of configuration. (Which most likely is the case for any application out
If you previously had code like this:
app.debug = DEBUG
app.secret_key = SECRET_KEY
You no longer have to do that, instead you can just load a configuration into the config
object. How this works is outlined in Configuration Handling.
Logging Integration
Flask now configures a logger for you with some basic and useful defaults. If you
run your application in production and want to profit from automatic error logging,
you might be interested in attaching a proper log handler. Also you can start logging
warnings and errors into the logger when appropriately. For more information on
that, read Application Errors.
Categories for Flash Messages
Flash messages can now have categories attached. This makes it possible to render
errors, warnings or regular messages differently for example. This is an opt-in feature
because it requires some rethinking in the code.
Read all about that in the Message Flashing pattern.
Flask Changelog
Here you can see the full list of changes between each Flask release.
Version 0.13
Major release, unreleased
• Make app.run() into a noop if a Flask application is run from the development
server on the command line. This avoids some behavior that was confusing to
debug for newcomers.
• Change default configuration JSONIFY_PRETTYPRINT_REGULAR=False.
jsonify() method returns compressed response by default, and pretty response
in debug mode.
• Change Flask.__init__ to accept two new keyword arguments, host_matching
and static_host. This enables host_matching to be set properly by the time
the constructor adds the static route, and enables the static route to be properly
associated with the required host. (#1559)
• send_file supports Unicode in attachment_filename. (#2223)
• Pass _scheme argument from url_for to handle_build_error. (#2017)
• Add support for provide_automatic_options in add_url_rule to disable adding
OPTIONS method when the view_func argument is not a class. (#1489).
• MethodView can inherit method handlers from base classes. (#1936)
• Errors caused while opening the session at the beginning of the request are handled by the app’s error handlers. (#2254)
• Blueprints gained json_encoder and json_decoder attributes to override the
app’s encoder and decoder. (#1898)
• Flask.make_response raises TypeError instead of ValueError for bad response
types. The error messages have been improved to describe why the type is invalid. (#2256)
• Add routes CLI command to output routes registered on the application. (#2259)
• Show warning when session cookie domain is a bare hostname or an IP address,
as these may not behave properly in some browsers, such as Chrome. (#2282)
• Allow IP address as exact session cookie domain. (#2282)
• SESSION_COOKIE_DOMAIN is set if it is detected through SERVER_NAME. (#2282)
• Auto-detect zero-argument app factory called create_app or make_app from
FLASK_APP. (#2297)
• Factory functions are not required to take a script_info parameter to work with
the flask command. If they take a single parameter or a parameter named
script_info, the ScriptInfo object will be passed. (#2319)
Version 0.12.2
Released on May 16 2017
• Fix a bug in safe_join on Windows.
Version 0.12.1
Bugfix release, released on March 31st 2017
• Prevent flask run from showing a NoAppException when an ImportError occurs
within the imported application module.
• Fix encoding behavior of app.config.from_pyfile for Python 3. Fix #2118.
• Use the SERVER_NAME config if it is present as default values for app.run. #2109,
• Call ctx.auto_pop with the exception object instead of None, in the event that a
BaseException such as KeyboardInterrupt is raised in a request handler.
Version 0.12
Released on December 21st 2016, codename Punsch.
• the cli command now responds to –version.
• Mimetype guessing and ETag generation for file-like objects in send_file has
been removed, as per issue #104. See pull request #1849.
• Mimetype guessing in send_file now fails loudly and doesn’t fall back to
application/octet-stream. See pull request #1988.
• Make flask.safe_join able to join multiple paths like os.path.join (pull request #1730).
• Revert a behavior change that made the dev server crash instead of returning a
Internal Server Error (pull request #2006).
• Correctly invoke response handlers for both regular request dispatching as well
as error handlers.
• Disable logger propagation by default for the app logger.
• Add support for range requests in send_file.
• app.test_client includes preset default environment, which can now be directly set, instead of per client.get.
Version 0.11.2
Bugfix release, unreleased
• Fix crash when running under PyPy3, see pull request #1814.
Version 0.11.1
Bugfix release, released on June 7th 2016.
• Fixed a bug that prevented FLASK_APP=foobar/__init__.py from working. See
pull request #1872.
Version 0.11
Released on May 29th 2016, codename Absinthe.
• Added support to serializing top-level arrays to flask.jsonify(). This introduces a security risk in ancient browsers. See JSON Security for details.
• Added before_render_template signal.
• Added **kwargs to flask.Test.test_client() to support passing additional
keyword arguments to the constructor of flask.Flask.test_client_class.
• Added SESSION_REFRESH_EACH_REQUEST config key that controls the set-cookie
behavior. If set to True a permanent session will be refreshed each request and
get their lifetime extended, if set to False it will only be modified if the session
actually modifies. Non permanent sessions are not affected by this and will always expire if the browser window closes.
• Made Flask support custom JSON mimetypes for incoming data.
• Added support for returning tuples in the form (response, headers) from a
view function.
• Added flask.Config.from_json().
• Added flask.Flask.config_class.
• Added flask.Config.get_namespace().
• Templates are no longer automatically reloaded outside of debug mode. This
can be configured with the new TEMPLATES_AUTO_RELOAD config key.
• Added a workaround for a limitation in Python 3.3’s namespace loader.
• Added support for explicit root paths when using Python 3.3’s namespace packages.
• Added flask and the flask.cli module to start the local debug server through
the click CLI system. This is recommended over the old flask.run() method
as it works faster and more reliable due to a different design and also replaces
• Error handlers that match specific classes are now checked first, thereby allowing catching exceptions that are subclasses of HTTP exceptions (in werkzeug.
exceptions). This makes it possible for an extension author to create exceptions
that will by default result in the HTTP error of their choosing, but may be caught
with a custom error handler if desired.
• Added flask.Config.from_mapping().
• Flask will now log by default even if debug is disabled. The log format
is now hardcoded but the default log handling can be disabled through the
LOGGER_HANDLER_POLICY configuration key.
• Removed deprecated module functionality.
• Added the EXPLAIN_TEMPLATE_LOADING config flag which when enabled will instruct Flask to explain how it locates templates. This should help users debug
when the wrong templates are loaded.
• Enforce blueprint handling in the order they were registered for template loading.
• Ported test suite to py.test.
• Deprecated request.json in favour of request.get_json().
• Add “pretty” and “compressed” separators definitions in jsonify() method. Reduces JSON response size when JSONIFY_PRETTYPRINT_REGULAR=False by
removing unnecessary white space included by default after separators.
• JSON responses are now terminated with a newline character, because it is a
convention that UNIX text files end with a newline and some clients don’t deal
well when this newline is missing. See https://github.com/pallets/flask/pull/
1262 – this came up originally as a part of https://github.com/kennethreitz/
• The automatically provided OPTIONS method is now correctly disabled if the user
registered an overriding rule with the lowercase-version options (issue #1288).
• flask.json.jsonify now supports the datetime.date type (pull request #1326).
• Don’t leak exception info of already catched exceptions to context teardown handlers (pull request #1393).
• Allow custom Jinja environment subclasses (pull request #1422).
• flask.g now has pop() and setdefault methods.
• Turn on autoescape for flask.templating.render_template_string by default
(pull request #1515).
• flask.ext is now deprecated (pull request #1484).
• send_from_directory now raises BadRequest if the filename is invalid on the
server OS (pull request #1763).
• Added the JSONIFY_MIMETYPE configuration variable (pull request #1728).
• Exceptions during teardown handling will no longer leave bad application contexts lingering around.
Version 0.10.2
(bugfix release, release date to be announced)
• Fixed broken test_appcontext_signals() test case.
• Raise an AttributeError in flask.helpers.find_package() with a useful message explaining why it is raised when a PEP 302 import hook is used without an
is_package() method.
• Fixed an issue causing exceptions raised before entering a request or app context
to be passed to teardown handlers.
• Fixed an issue with query parameters getting removed from requests in the test
client when absolute URLs were requested.
• Made @before_first_request into a decorator as intended.
• Fixed an etags bug when sending a file streams with a name.
• Fixed send_from_directory not expanding to the application root path correctly.
• Changed logic of before first request handlers to flip the flag after invoking. This
will allow some uses that are potentially dangerous but should probably be permitted.
• Fixed Python 3 bug when a handler from app.url_build_error_handlers reraises the
Version 0.10.1
(bugfix release, released on June 14th 2013)
• Fixed an issue where |tojson was not quoting single quotes which made the
filter not work properly in HTML attributes. Now it’s possible to use that filter
in single quoted attributes. This should make using that filter with angular.js
• Added support for byte strings back to the session system. This broke compatibility with the common case of people putting binary data for token verification
into the session.
• Fixed an issue where registering the same method twice for the same endpoint
would trigger an exception incorrectly.
Version 0.10
Released on June 13th 2013, codename Limoncello.
• Changed default cookie serialization format from pickle to JSON to limit the
impact an attacker can do if the secret key leaks. See Version 0.10 for more information.
• Added template_test methods
template_filter method family.
• Added template_global methods in addition to the already existing
template_filter method family.
• Set the content-length header for x-sendfile.
• tojson filter now does not escape script blocks in HTML5 parsers.
• tojson used in templates is now safe by default due. This was allowed due to
the different escaping behavior.
• Flask will now raise an error if you attempt to register a new function on an
already used endpoint.
• Added wrapper module around simplejson and added default serialization of
datetime objects. This allows much easier customization of how JSON is handled
by Flask or any Flask extension.
• Removed deprecated internal flask.session module alias. Use flask.sessions
instead to get the session module. This is not to be confused with flask.session
the session proxy.
• Templates can now be rendered without request context. The behavior is
slightly different as the request, session and g objects will not be available and
blueprint’s context processors are not called.
• The config object is now available to the template as a real global and not through
a context processor which makes it available even in imported templates by default.
• Added an option to generate non-ascii encoded JSON which should result in less
bytes being transmitted over the network. It’s disabled by default to not cause
confusion with existing libraries that might expect flask.json.dumps to return
bytestrings by default.
• flask.g is now stored on the app context instead of the request context.
• flask.g now gained a get() method for not erroring out on non existing items.
• flask.g now can be used with the in operator to see what’s defined and it now
is iterable and will yield all attributes stored.
• flask.Flask.request_globals_class
app_ctx_globals_class which is a better name to what it does since 0.10.
• request, session and g are now also added as proxies to the template context which
makes them available in imported templates. One has to be very careful with
those though because usage outside of macros might cause caching.
• Flask will no longer invoke the wrong error handlers if a proxy exception is
passed through.
• Added a workaround for chrome’s cookies in localhost not working as intended
with domain names.
• Changed logic for picking defaults for cookie values from sessions to work better
with Google Chrome.
• Added message_flashed signal that simplifies flashing testing.
• Added support for copying of request contexts for better working with greenlets.
• Removed custom JSON HTTP exception subclasses. If you were relying on
them you can reintroduce them again yourself trivially. Using them however
is strongly discouraged as the interface was flawed.
• Python requirements changed: requiring Python 2.6 or 2.7 now to prepare for
Python 3.3 port.
• Changed how the teardown system is informed about exceptions. This is now
more reliable in case something handles an exception halfway through the error
handling process.
• Request context preservation in debug mode now keeps the exception information around which means that teardown handlers are able to distinguish error
from success cases.
• Added the JSONIFY_PRETTYPRINT_REGULAR configuration variable.
• Flask now orders JSON keys by default to not trash HTTP caches due to different
hash seeds between different workers.
• Added appcontext_pushed and appcontext_popped signals.
• The builtin run method now takes the SERVER_NAME into account when picking
the default port to run on.
• Added flask.request.get_json() as a replacement for the old flask.request.json property.
Version 0.9
Released on July 1st 2012, codename Campari.
• The flask.Request.on_json_loading_failed() now returns a JSON formatted
response by default.
• The flask.url_for() function now can generate anchors to the generated links.
• The flask.url_for() function now can also explicitly generate URL rules specific to a given HTTP method.
• Logger now only returns the debug log setting if it was not set explicitly.
• Unregister a circular dependency between the WSGI environment and the request object when shutting down the request. This means that environ werkzeug.
request will be None after the response was returned to the WSGI server but has
the advantage that the garbage collector is not needed on CPython to tear down
the request unless the user created circular dependencies themselves.
• Session is now stored after callbacks so that if the session payload is stored in the
session you can still modify it in an after request callback.
• The flask.Flask class will avoid importing the provided import name if it can
(the required first parameter), to benefit tools which build Flask instances programmatically. The Flask class will fall back to using import on systems with
custom module hooks, e.g. Google App Engine, or when the import name is
inside a zip archive (usually a .egg) prior to Python 2.7.
• Blueprints now have a decorator to add custom template filters application wide,
• The Flask and Blueprint classes now have a non-decorator method for adding
custom template filters application wide, flask.Flask.add_template_filter()
and flask.Blueprint.add_app_template_filter().
• The flask.get_flashed_messages() function now allows rendering flashed message categories in separate blocks, through a category_filter argument.
• The flask.Flask.run() method now accepts None for host and port arguments,
using default values when None. This allows for calling run using configuration
values, e.g. app.run(app.config.get('MYHOST'), app.config.get('MYPORT')),
with proper behavior whether or not a config file is provided.
• The flask.render_template() method now accepts a either an iterable of template names or a single template name. Previously, it only accepted a single
template name. On an iterable, the first template found is rendered.
• Added flask.Flask.app_context() which works very similar to the request context but only provides access to the current application. This also adds support
for URL generation without an active request context.
• View functions can now return a tuple with the first instance being an instance
of flask.Response. This allows for returning jsonify(error="error msg"), 400
from a view function.
• Flask and Blueprint now provide a get_send_file_max_age() hook for subclasses to override behavior of serving static files from Flask when using
flask.Flask.send_static_file() (used for the default static file handler) and
send_file(). This hook is provided a filename, which for example allows changing cache controls by file extension. The default max-age for send_file and static
files can be configured through a new SEND_FILE_MAX_AGE_DEFAULT configuration
variable, which is used in the default get_send_file_max_age implementation.
• Fixed an assumption in sessions implementation which could break message
flashing on sessions implementations which use external storage.
• Changed the behavior of tuple return values from functions. They are no longer
arguments to the response object, they now have a defined meaning.
• Added flask.Flask.request_globals_class to allow a specific class to be used
on creation of the g instance of each request.
• Added required_methods attribute to view functions to force-add methods on registration.
• Added flask.after_this_request().
• Added flask.stream_with_context() and the ability to push contexts multiple
times without producing unexpected behavior.
Version 0.8.1
Bugfix release, released on July 1st 2012
• Fixed an issue with the undocumented flask.session module to not work properly
on Python 2.5. It should not be used but did cause some problems for package
Version 0.8
Released on September 29th 2011, codename Rakija
• Refactored session support into a session interface so that the implementation of
the sessions can be changed without having to override the Flask class.
• Empty session cookies are now deleted properly automatically.
• View functions can now opt out of getting the automatic OPTIONS implementation.
• HTTP exceptions and Bad Request errors can now be trapped so that they show
up normally in the traceback.
• Flask in debug mode is now detecting some common problems and tries to warn
you about them.
• Flask in debug mode will now complain with an assertion error if a view was
attached after the first request was handled. This gives earlier feedback when
users forget to import view code ahead of time.
• Added the ability to register callbacks that are only triggered once at the beginning of the first request. (Flask.before_first_request())
• Malformed JSON data will now trigger a bad request HTTP exception instead
of a value error which usually would result in a 500 internal server error if not
handled. This is a backwards incompatible change.
• Applications now not only have a root path where the resources and modules are
located but also an instance path which is the designated place to drop files that
are modified at runtime (uploads etc.). Also this is conceptually only instance depending and outside version control so it’s the perfect place to put configuration
files etc. For more information see Instance Folders.
• Added the APPLICATION_ROOT configuration variable.
• Implemented session_transaction() to easily modify sessions from the test environment.
• Refactored test client internally. The APPLICATION_ROOT configuration variable as
well as SERVER_NAME are now properly used by the test client as defaults.
• Added flask.views.View.decorators to support simpler decorating of pluggable (class-based) views.
• Fixed an issue where the test client if used with the “with” statement did not
trigger the execution of the teardown handlers.
• Added finer control over the session cookie parameters.
• HEAD requests to a method view now automatically dispatch to the get method
if no handler was implemented.
• Implemented the virtual flask.ext package to import extensions from.
• The context preservation on exceptions is now an integral component of Flask
itself and no longer of the test client. This cleaned up some internal logic and
lowers the odds of runaway request contexts in unittests.
Version 0.7.3
Bugfix release, release date to be decided
• Fixed the Jinja2 environment’s list_templates method not returning the correct
names when blueprints or modules were involved.
Version 0.7.2
Bugfix release, released on July 6th 2011
• Fixed an issue with URL processors not properly working on blueprints.
Version 0.7.1
Bugfix release, released on June 29th 2011
• Added missing future import that broke 2.5 compatibility.
• Fixed an infinite redirect issue with blueprints.
Version 0.7
Released on June 28th 2011, codename Grappa
• Added make_default_options_response() which can be used by subclasses to
alter the default behavior for OPTIONS responses.
• Unbound locals now raise a proper RuntimeError instead of an AttributeError.
• Mimetype guessing and etag support based on file objects is now deprecated for
flask.send_file() because it was unreliable. Pass filenames instead or attach
your own etags and provide a proper mimetype by hand.
• Static file handling for modules now requires the name of the static folder to
be supplied explicitly. The previous autodetection was not reliable and caused
issues on Google’s App Engine. Until 1.0 the old behavior will continue to work
but issue dependency warnings.
• fixed a problem for Flask to run on jython.
• added a PROPAGATE_EXCEPTIONS configuration variable that can be used to flip
the setting of exception propagation which previously was linked to DEBUG alone
and is now linked to either DEBUG or TESTING.
• Flask no longer internally depends on rules being added through the add_url_rule
function and can now also accept regular werkzeug rules added to the url map.
• Added an endpoint method to the flask application object which allows one to
register a callback to an arbitrary endpoint with a decorator.
• Use Last-Modified for static file sending instead of Date which was incorrectly
introduced in 0.6.
• Added create_jinja_loader to override the loader creation process.
• Implemented a silent flag for config.from_pyfile.
• Added teardown_request decorator, for functions that should run at the end of
a request regardless of whether an exception occurred. Also the behavior for
after_request was changed. It’s now no longer executed when an exception is
raised. See Upgrading to new Teardown Handling
• Implemented flask.has_request_context()
• Deprecated init_jinja_globals. Override the create_jinja_environment() method
instead to achieve the same functionality.
• Added flask.safe_join()
• The automatic JSON request data unpacking now looks at the charset mimetype
• Don’t modify the session on flask.get_flashed_messages() if there are no messages in the session.
• before_request handlers are now able to abort requests with errors.
• it is not possible to define user exception handlers. That way you can provide
custom error messages from a central hub for certain errors that might occur during request processing (for instance database connection errors, timeouts from
remote resources etc.).
• Blueprints can provide blueprint specific error handlers.
• Implemented generic Pluggable Views (class-based views).
Version 0.6.1
Bugfix release, released on December 31st 2010
• Fixed an issue where the default OPTIONS response was not exposing all valid
methods in the Allow header.
• Jinja2 template loading syntax now allows ”./” in front of a template load path.
Previously this caused issues with module setups.
• Fixed an issue where the subdomain setting for modules was ignored for the
static folder.
• Fixed a security problem that allowed clients to download arbitrary files if the
host server was a windows based operating system and the client uses backslashes to escape the directory the files where exposed from.
Version 0.6
Released on July 27th 2010, codename Whisky
• after request functions are now called in reverse order of registration.
• OPTIONS is now automatically implemented by Flask unless the application
explicitly adds ‘OPTIONS’ as method to the URL rule. In this case no automatic
OPTIONS handling kicks in.
• static rules are now even in place if there is no static folder for the module. This
was implemented to aid GAE which will remove the static folder if it’s part of a
mapping in the .yml file.
• the config is now available in the templates as config.
• context processors will no longer override values passed directly to the render
• added the ability to limit the incoming request data with the new
MAX_CONTENT_LENGTH configuration value.
• the endpoint for the flask.Module.add_url_rule() method is now optional to
be consistent with the function of the same name on the application object.
• added a flask.make_response() function that simplifies creating response object
instances in views.
• added signalling support based on blinker. This feature is currently optional and
supposed to be used by extensions and applications. If you want to use it, make
sure to have blinker installed.
• refactored the way URL adapters are created. This process is now fully customizable with the create_url_adapter() method.
• modules can now register for a subdomain instead of just an URL prefix. This
makes it possible to bind a whole module to a configurable subdomain.
Version 0.5.2
Bugfix Release, released on July 15th 2010
• fixed another issue with loading templates from directories when modules were
Version 0.5.1
Bugfix Release, released on July 6th 2010
• fixes an issue with template loading from directories when modules where used.
Version 0.5
Released on July 6th 2010, codename Calvados
• fixed a bug with subdomains that was caused by the inability to specify the
server name. The server name can now be set with the SERVER_NAME config key.
This key is now also used to set the session cookie cross-subdomain wide.
• autoescaping is no longer active for all templates. Instead it is only active for .
html, .htm, .xml and .xhtml. Inside templates this behavior can be changed with
the autoescape tag.
• refactored Flask internally. It now consists of more than a single file.
• flask.send_file() now emits etags and has the ability to do conditional responses builtin.
• (temporarily) dropped support for zipped applications. This was a rarely used
feature and led to some confusing behavior.
• added support for per-package template and static-file directories.
• removed support for create_jinja_loader which is no longer used in 0.5 due to the
improved module support.
• added a helper function to expose files from any directory.
Version 0.4
Released on June 18th 2010, codename Rakia
• added the ability to register application wide error handlers from modules.
• after_request() handlers are now also invoked if the request dies with an exception and an error handling page kicks in.
• test client has not the ability to preserve the request context for a little longer.
This can also be used to trigger custom requests that do not pop the request
stack for testing.
• because the Python standard library caches loggers, the name of the logger is
configurable now to better support unittests.
• added TESTING switch that can activate unittesting helpers.
• the logger switches to DEBUG mode now if debug is enabled.
Version 0.3.1
Bugfix release, released on May 28th 2010
• fixed a error reporting bug with flask.Config.from_envvar()
• removed some unused code from flask
• release does no longer include development leftover files (.git folder for themes,
built documentation in zip and pdf file and some .pyc files)
Version 0.3
Released on May 28th 2010, codename Schnaps
• added support for categories for flashed messages.
• the application now configures a logging.Handler and will log request handling
exceptions to that logger when not in debug mode. This makes it possible to
receive mails on server errors for example.
• added support for context binding that does not require the use of the with statement for playing in the console.
• the request context is now available within the with statement making it possible
to further push the request context or pop it.
• added support for configurations.
Version 0.2
Released on May 12th 2010, codename Jägermeister
• various bugfixes
• integrated JSON support
• added get_template_attribute() helper function.
• add_url_rule() can now also register a view function.
• refactored internal request dispatching.
• server listens on by default now to fix issues with chrome.
• added external URL support.
• added support for send_file()
• module support and internal request handling refactoring to better support
pluggable applications.
• sessions can be set to be permanent now on a per-session basis.
• better error reporting on missing secret keys.
• added support for Google Appengine.
Version 0.1
First public preview release.
Flask is licensed under a three clause BSD License. It basically means: do whatever
you want with it as long as the copyright in Flask sticks around, the conditions are not
modified and the disclaimer is present. Furthermore you must not use the names of
the authors to promote derivatives of the software without written consent.
The full license text can be found below (Flask License). For the documentation and
artwork different licenses apply.
Flask is written and maintained by Armin Ronacher and various contributors:
Development Lead
• Armin Ronacher <[email protected]>
Patches and Suggestions
• Adam Zapletal
• Ali Afshar
• Chris Edgemon
• Chris Grindstaff
• Christopher Grebs
• Daniel Neuhäuser
• Dan Sully
• David Lord @davidism
• Edmond Burnett
• Florent Xicluna
• Georg Brandl
• Jeff Widman @jeffwidman
• Joshua Bronson @jab
• Justin Quick
• Kenneth Reitz
• Keyan Pishdadian
• Marian Sigler
• Martijn Pieters
• Matt Campell
• Matthew Frazier
• Michael van Tellingen
• Ron DuPlain
• Sebastien Estienne
• Simon Sapin
• Stephane Wirtel
• Thomas Schranz
• Zhao Xiaohong
General License Definitions
The following section contains the full license texts for Flask and the documentation.
• “AUTHORS” hereby refers to all the authors listed in the Authors section.
• The “Flask License” applies to all the source code shipped as part of Flask (Flask
itself as well as the examples and the unittests) as well as documentation.
• The “Flask Artwork License” applies to the project’s Horn-Logo.
Flask License
Copyright (c) 2015 by Armin Ronacher and contributors. See AUTHORS for more
Some rights reserved.
Redistribution and use in source and binary forms of the software as well as documentation, with or without modification, are permitted provided that the following
conditions are met:
• Redistributions of source code must retain the above copyright notice, this list of
conditions and the following disclaimer.
• Redistributions in binary form must reproduce the above copyright notice, this
list of conditions and the following disclaimer in the documentation and/or
other materials provided with the distribution.
• The names of the contributors may not be used to endorse or promote products
derived from this software without specific prior written permission.
Flask Artwork License
Copyright (c) 2010 by Armin Ronacher.
Some rights reserved.
This logo or a modified version may be used by anyone to refer to the Flask project,
but does not indicate endorsement by the project.
Redistribution and use in source (the SVG file) and binary forms (rendered PNG files
etc.) of the image, with or without modification, are permitted provided that the following conditions are met:
• Redistributions of source code must retain the above copyright notice and this
list of conditions.
• The names of the contributors to the Flask software (see AUTHORS) may not be
used to endorse or promote products derived from this software without specific
prior written permission.
Note: we would appreciate that you make the image a link to http://flask.pocoo.org/
if you use it on a web page.
add_url_rule() (flask.Flask method), 195
_app_ctx_stack (in module flask), 254
_request_ctx_stack (in module flask), 253
after_request() (flask.Blueprint method),
(flask.Flask method), 196
abort() (in module flask), 238
accept_charsets (flask.Request attribute), after_request_funcs (flask.Flask attribute),
at- after_this_request() (in module flask), 240
app (flask.blueprints.BlueprintSetupState
tribute), 222
attribute), 254
atapp_context() (flask.Flask method), 196
tribute), 222
accept_mimetypes (flask.Request at- app_context_processor() (flask.Blueprint
method), 217
tribute), 222
access_route (flask.Request attribute), 222 app_ctx_globals_class (flask.Flask attribute), 197
accessed (flask.sessions.SessionMixin atapp_errorhandler()
tribute), 234
method), 217
(flask.cli.ScriptInfo at(flask.Blueprint method), 216
tribute), 262
(flask.Blueprint method), 216
method), 217
add_app_template_test() (flask.Blueprint
method), 217
method), 195
method), 195
method), 218
add_template_test() (flask.Flask method),
(flask.Blueprint method), 218
add_url_rule() (flask.Blueprint method),
AppContext (class in flask.ctx), 253
appcontext_popped (in module flask), 257
(flask.blueprints.BlueprintSetupStateappcontext_pushed (in module flask), 256
appcontext_tearing_down (in module
method), 254
flask), 256
AppGroup (class in flask.cli), 262
application() (flask.Request method), 222
args (flask.Request attribute), 222
method), 258
authorization (flask.Request attribute),
method), 197
content_type (flask.Request attribute), 223
method), 218
context_processor() (flask.Flask method),
cookies (flask.Request attribute), 223
copy() (flask.ctx.RequestContext method),
copy_current_request_context() (in module flask), 236
create_app (flask.cli.ScriptInfo attribute),
create_global_jinja_loader() (flask.Flask
method), 198
method), 199
create_url_adapter() (flask.Flask method),
current_app (in module flask), 236
base_url (flask.Request attribute), 221, 222
(flask.Blueprint method), 218
method), 218
method), 197
before_first_request_funcs (flask.Flask attribute), 197
before_request() (flask.Blueprint method), D
data (flask.cli.ScriptInfo attribute), 263
before_request() (flask.Flask method), 197 data (flask.Request attribute), 223
before_request_funcs (flask.Flask at- data (flask.Response attribute), 230
tribute), 198
date (flask.Request attribute), 223
Blueprint (class in flask), 216
debug (flask.Flask attribute), 199
blueprint (flask.blueprints.BlueprintSetupState
decorators (flask.views.View attribute),
attribute), 254
blueprint (flask.Request attribute), 222
blueprints (flask.Flask attribute), 198
method), 247
in default_config (flask.Flask attribute), 199
flask.blueprints), 254
attribute), 223
cache_control (flask.Request attribute),
static method), 233
cli (flask.Flask attribute), 198
dispatch_request() (flask.Flask method),
close() (flask.Request method), 222
command() (flask.cli.AppGroup method), dispatch_request()
method), 259
Config (class in flask), 248
do_teardown_appcontext() (flask.Flask
config (flask.Flask attribute), 198
method), 199
config_class (flask.Flask attribute), 198
atmethod), 199
tribute), 223
dump() (in module flask.json), 246
content_length (flask.Request attribute), dumps() (in module flask.json), 246
content_md5 (flask.Request attribute), 223
endpoint (flask.Request attribute), 223
endpoint() (flask.Blueprint method), 218
endpoint() (flask.Flask method), 200
environ (flask.Request attribute), 221
environment variable
error_handler_spec (flask.Flask attribute),
errorhandler() (flask.Blueprint method),
errorhandler() (flask.Flask method), 200
escape() (flask.Markup class method), 243
escape() (in module flask), 242
extensions (flask.Flask attribute), 201
g (in module flask), 235
method), 232
method), 232
method), 232
method), 232
get_data() (flask.Request method), 224
method), 232
get_flashed_messages() (in module flask),
files (flask.Request attribute), 223
get_json() (flask.Request method), 225
(flask.blueprints.BlueprintSetupStateget_namespace() (flask.Config method),
attribute), 254
flash() (in module flask), 244
get_send_file_max_age() (flask.Blueprint
Flask (class in flask), 193
method), 218
flask (module), 193
flask.ext (in module flask), 251
method), 201
flask.json (module), 245
flask), 248
FlaskClient (class in flask.testing), 235
got_first_request (flask.Flask attribute),
FlaskGroup (class in flask.cli), 262
got_request_exception (in module flask),
form (flask.Request attribute), 224
form_data_parser_class (flask.Request at- group() (flask.cli.AppGroup method), 262
tribute), 224
from_envvar() (flask.Config method), 249 H
from_json() (flask.Config method), 249
handle_exception() (flask.Flask method),
from_mapping() (flask.Config method),
from_object() (flask.Config method), 250
method), 202
from_pyfile() (flask.Config method), 250
from_values() (flask.Request method),
method), 202
method), 202
method), 201
has_app_context() (in module flask), 237
full_path (flask.Request attribute), 221, has_request_context() (in module flask),
has_static_folder (flask.Blueprint attribute), 219
has_static_folder (flask.Flask attribute),
headers (flask.Request attribute), 225
headers (flask.Response attribute), 229
host (flask.Request attribute), 225
host_url (flask.Request attribute), 225
if_match (flask.Request attribute), 225
attribute), 225
if_none_match (flask.Request attribute),
if_range (flask.Request attribute), 225
if_unmodified_since (flask.Request attribute), 226
init_jinja_globals() (flask.Flask method),
inject_url_defaults() (flask.Flask method),
instance_path (flask.Flask attribute), 202
is_json (flask.Request attribute), 226
is_multiprocess (flask.Request attribute),
is_multithread (flask.Request attribute),
method), 232
is_run_once (flask.Request attribute), 226
is_secure (flask.Request attribute), 226
is_xhr (flask.Request attribute), 226
iter_blueprints() (flask.Flask method), 203
json_encoder (flask.Blueprint attribute),
json_encoder (flask.Flask attribute), 203
JSONDecoder (class in flask.json), 247
JSONEncoder (class in flask.json), 247
jsonify() (in module flask.json), 245
attribute), 233
attribute), 226
load() (in module flask.json), 247
load_app() (flask.cli.ScriptInfo method),
loads() (in module flask.json), 246
log_exception() (flask.Flask method), 203
logger (flask.Flask attribute), 203
logger_name (flask.Flask attribute), 203
make_config() (flask.Flask method), 204
(flask.Flask method), 204
make_form_data_parser() (flask.Request
method), 226
make_null_session() (flask.Flask method),
method), 232
make_response() (flask.Flask method),
make_response() (in module flask), 239
jinja_env (flask.Flask attribute), 203
method), 219
jinja_environment (flask.Flask attribute),
method), 205
jinja_loader (flask.Blueprint attribute),
Markup (class in flask), 242
jinja_loader (flask.Flask attribute), 203
jinja_options (flask.Flask attribute), 203
method), 253
json (flask.Request attribute), 226
max_content_length (flask.Request atjson_decoder (flask.Blueprint attribute),
tribute), 226
max_forwards (flask.Request attribute),
json_decoder (flask.Flask attribute), 203
message_flashed (in module flask), 257
method (flask.Request attribute), 227
methods (flask.views.View attribute), 259
MethodView (class in flask.views), 259
mimetype (flask.Request attribute), 227
mimetype (flask.Response attribute), 230
mimetype_params (flask.Request attribute), 227
modified (flask.session attribute), 231
modified (flask.sessions.SessionMixin attribute), 234
module (flask.Request attribute), 227
attribute), 233
pop() (flask.ctx.AppContext method), 253
pop() (flask.ctx.RequestContext method),
pragma (flask.Request attribute), 227
method), 206
(flask.Flask attribute), 206
process_response() (flask.Flask method),
propagate_exceptions (flask.Flask atname (flask.Flask attribute), 205
tribute), 206
new (flask.session attribute), 231
new (flask.sessions.SessionMixin at(flask.views.View
tribute), 234
push() (flask.ctx.AppContext method),
attribute), 232
NullSession (class in flask.sessions), 233
method), 253
Python Enhancement Proposals
PEP 8, 293
on_json_loading_failed() (flask.Request
method), 227
(flask.Flask query_string (flask.Request attribute), 227
method), 205
open_resource() (flask.Blueprint method), R
range (flask.Request attribute), 227
open_resource() (flask.Flask method), 205 record() (flask.Blueprint method), 220
open_session() (flask.Flask method), 205 record_once() (flask.Blueprint method),
redirect() (in module flask), 238
method), 233
referrer (flask.Request attribute), 227
options (flask.blueprints.BlueprintSetupState
register() (flask.Blueprint method), 220
attribute), 254
register_blueprint() (flask.Flask method),
register_error_handler() (flask.Blueprint
parameter_storage_class (flask.Request
method), 220
attribute), 227
pass_script_info() (in module flask.cli),
method), 206
remote_addr (flask.Request attribute), 228
path (flask.Request attribute), 221, 227
remote_user (flask.Request attribute), 228
permanent (flask.session attribute), 231
render_template() (in module flask), 247
permanent (flask.sessions.SessionMixin render_template_string() (in module
attribute), 234
flask), 247
permanent_session_lifetime (flask.Flask Request (class in flask), 221
attribute), 206
request (in module flask), 229
request_class (flask.Flask attribute), 207
request_context() (flask.Flask method),
request_finished (in module flask), 255
request_started (in module flask), 255
request_tearing_down (in module flask),
RequestContext (class in flask.ctx), 252
Response (class in flask), 229
response_class (flask.Flask attribute), 207
RFC 822, 245
route() (flask.Blueprint method), 220
route() (flask.Flask method), 207
attribute), 228
run() (flask.Flask method), 208
run_command (in module flask.cli), 263
serializer (flask.sessions.SecureCookieSessionInterface
attribute), 233
session (class in flask), 230
attribute), 233
session_cookie_name (flask.Flask attribute), 209
session_interface (flask.Flask attribute),
flask.sessions), 234
(flask.testing.FlaskClient method),
SessionInterface (class in flask.sessions),
SessionMixin (class in flask.sessions), 234
set_cookie() (flask.Response method), 230
shell_command (in module flask.cli), 263
safe_join() (in module flask), 242
method), 210
salt (flask.sessions.SecureCookieSessionInterface
shell_context_processors (flask.Flask atattribute), 233
tribute), 210
save_session() (flask.Flask method), 209
method), 210
method), 233
scheme (flask.Request attribute), 228
method), 233
script_root (flask.Request attribute), 221, signal()
method), 257
ScriptInfo (class in flask.cli), 262
signals.Namespace (class in flask), 257
secret_key (flask.Flask attribute), 209
signals.signals_available (in module
flask), 254
flask.sessions), 233
static_folder (flask.Blueprint attribute),
SecureCookieSessionInterface (class in
flask.sessions), 233
static_folder (flask.Flask attribute), 210
(flask.Flask status (flask.Response attribute), 229
method), 209
status_code (flask.Response attribute),
send_file() (in module flask), 240
send_file_max_age_default (flask.Flask stream (flask.Request attribute), 228
attribute), 209
stream_with_context() (in module flask),
send_from_directory() (in module flask),
striptags() (flask.Markup method), 243
(flask.Blueprint subdomain (flask.blueprints.BlueprintSetupState
method), 220
attribute), 254
send_static_file() (flask.Flask method),
teardown_app_request() (flask.Blueprint
method), 220
method), 210
teardown_appcontext_funcs (flask.Flask
attribute), 210
method), 220
teardown_request() (flask.Flask method),
teardown_request_funcs (flask.Flask attribute), 211
template_context_processors (flask.Flask
attribute), 211
template_filter() (flask.Flask method), 211
template_global() (flask.Flask method),
template_rendered (in module flask), 254
template_test() (flask.Flask method), 212
test_client() (flask.Flask method), 212
test_client_class (flask.Flask attribute),
method), 213
testing (flask.Flask attribute), 213
method), 214
url_root (flask.Request attribute), 221, 228
url_rule (flask.Request attribute), 228
url_rule_class (flask.Flask attribute), 215
url_value_preprocessor() (flask.Blueprint
method), 221
method), 215
url_value_preprocessors (flask.Flask attribute), 215
use_x_sendfile (flask.Flask attribute), 215
user_agent (flask.Request attribute), 229
values (flask.Request attribute), 229
View (class in flask.views), 258
view_args (flask.Request attribute), 229
view_functions (flask.Flask attribute), 215
want_form_data_parsed (flask.Request
attribute), 229
with_appcontext() (in module flask.cli),
wsgi_app() (flask.Flask method), 216
unescape() (flask.Markup method), 243
method), 214
url (flask.Request attribute), 221, 228
url_build_error_handlers (flask.Flask attribute), 214
url_charset (flask.Request attribute), 228
url_default_functions (flask.Flask attribute), 214
url_defaults (flask.blueprints.BlueprintSetupState
attribute), 254
url_defaults() (flask.Blueprint method),
url_defaults() (flask.Flask method), 214
url_for() (in module flask), 237
url_map (flask.Flask attribute), 214
url_prefix (flask.blueprints.BlueprintSetupState
attribute), 254
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF