WSGI Types for Python

This is an attempt to bring some type safety to WSGI applications using Python's new typing features (TypedDicts, Protocols). It seems to work OK but may still be full of gaps, holes, bugs, missteps, etc. It helped bring a lot of extra safety to a couple of places that really needed it though, and seemed to remain quite stable for a couple of months.

This is implemented as a Python module, rather than MyPy stubs, as it represents a protocol things can satisfy rather than a set of types for something concrete.

This package came together during an exploration documented here: https://github.com/python/mypy/issues/7654

Define a callable application as a class:

import wsgitypes

class MyApplication(wsgitypes.Application):
    def __call__(
        self,
        environ: wsgitypes.Environ,
        start_response: wsgitypes.StartResponse,
    ) -> wsgitypes.Response:
        my_header = environ.get("REQUEST_METHOD", "")
        return []

Environ should be type-safe:

class MyApplication(wsgitypes.Application):
    def __call__(
        self,
        environ: wsgitypes.Environ,
        start_response: wsgitypes.StartResponse,
    ) -> wsgitypes.Response:
        environ["wsgi.input"] # Good
        environ["wsgi.unpot"] # BORK! MyPy will catch this.
        return []

You can define your own extensions to Environ using TypedDict inheritance, like so:

class MyEnviron(wsgitypes.Environ):
    HTTP_X_MY_HEADER: t.Optional[str]

class MyApplication(wsgitypes.Application):
    def __call__(
        self,
        environ: MyEnviron,
        start_response: wsgitypes.StartResponse,
    ) -> wsgitypes.Response:
        environ = typing.cast(MyEnviron, environ)
        environ.get("HTTP_X_MY_HEADER") # Good
        return []

Note that you need to use typing.cast to convert the incoming Environ to your derived version. An attempt was made to use a type param for Environ, but it wasn't viable (even with GVR helping!): https://github.com/python/mypy/issues/7654

MyPy types for WSGI applications

Related tags

Overview

WSGI Types for Python

Owner

Blake Williams

Tool for pinpointing circular imports in Python. Find cyclic imports in any project

Static type checker for Python

flake8 plugin to catch useless `assert` statements

MonkeyType as a pytest plugin.

Mypy stubs, i.e., type information, for numpy, pandas and matplotlib

Enforce the same configuration across multiple projects

Type annotations builder for boto3 compatible with VSCode, PyCharm, Emacs, Sublime Text, pyright and mypy.

Utilities for refactoring imports in python-like syntax.

coala provides a unified command-line interface for linting and fixing all your code, regardless of the programming languages you use.

Flake8 plugin to validate annotations complexity

Pylint plugin for improving code analysis for when using Django

A plugin for Flake8 finding likely bugs and design problems in your program. Contains warnings that don't belong in pyflakes and pycodestyle.

Tools for improving Python imports

Flake8 Type Annotation Checking

A Pylint plugin to analyze Flask applications.

Naming Convention checker for Python

PEP-484 typing stubs for SQLAlchemy 1.4 and SQLAlchemy 2.0

Tool to automatically fix some issues reported by flake8 (forked from autoflake).

flake8 plugin that integrates isort

Run isort, pyupgrade, mypy, pylint, flake8, and more on Jupyter Notebooks