flask/docs/design.rst

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 :class:`~flask.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__)

    @app.route('/')
    def index():
        return 'Hello World!'

Would look like this instead::

    from hypothetical_flask import route

    @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 unit testing.  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
(:class:`~flask.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 :meth:`~flask.Flask.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 :meth:`~flask.Flask.wsgi_app`).

Furthermore this design makes it possible to use a factory function to
create the application which is very helpful for unit testing and similar
things (:doc:`/patterns/appfactories`).

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: Jinja.  Why doesn't Flask have a
pluggable template engine interface?  You can obviously use a different
template engine, but Flask will still configure Jinja for you.  While
that limitation that Jinja is *always* configured will probably go away,
the decision to bundle one template engine and use that will not.

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. Jinja 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, 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 Jinja's extensive autoescaping support.  Also it provides
ways to access macros from Jinja 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.


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.

Why does Flask call itself a microframework and yet it depends on two
libraries (namely Werkzeug and Jinja).  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.


Context Locals
--------------

Flask uses special context locals and proxies to provide access to the
current app and request data to any code running during a request, CLI command,
etc. Context locals are specific to the worker handling the activity, such as a
thread, process, coroutine, or greenlet.

The context and proxies help solve two development issues: circular imports, and
passing around global data. :data:`.current_app: can be used to access the
application object without needing to import the app object directly, avoiding
circular import issues. :data:`.request`, :data:`.session`, and :data`.g` can be
imported to access the current data for the request, rather than needing to
pass them as arguments through every single function in your project.


Async/await and ASGI support
----------------------------

Flask supports ``async`` coroutines for view functions by executing the
coroutine on a separate thread instead of using an event loop on the
main thread as an async-first (ASGI) framework would. This is necessary
for Flask to remain backwards compatible with extensions and code built
before ``async`` was introduced into Python. This compromise introduces
a performance cost compared with the ASGI frameworks, due to the
overhead of the threads.

Due to how tied to WSGI Flask's code is, it's not clear if it's possible
to make the ``Flask`` class support ASGI and WSGI at the same time. Work
is currently being done in Werkzeug to work with ASGI, which may
eventually enable support in Flask as well.

See :doc:`/async-await` for more discussion.


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 Jinja 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.

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.

The idea of Flask is to build a good foundation for all applications.
Everything else is up to you or extensions.