flake8-annotations is a plugin for Flake8 that detects the absence of PEP 3107-style function annotations and PEP 484-style type comments (see: Caveats).

What this won't do: Check variable annotations (see: PEP 526), respect stub files, or replace mypy.

Install from PyPi with your favorite pip invocation:

$ pip install flake8-annotations

It will then be run automatically as part of flake8.

You can verify it's being picked up by invoking the following in your shell:

$ flake8 --version
3.9.0 (flake8-annotations: 2.6.1, mccabe: 0.6.1, pycodestyle: 2.7.0, pyflakes: 2.3.0) CPython 3.9.1 on Darwin

All warnings are enabled by default.

Notes:

1. See: PEP 484 and PEP 563 for suggestions on annotating self and cls arguments.

Some opinionated flags are provided to tailor the linting errors emitted.

Suppress ANN200-level errors for functions that meet one of the following criteria:

• Contain no return statement, or
• Explicit return statement(s) all return None (explicitly or implicitly).

Default: False

Suppress ANN000-level errors for dummy arguments, defined as _.

Default: False

Suppress all errors for dynamically typed functions. A function is considered dynamically typed if it does not contain any type hints.

Default: False

Suppress all errors for dynamically typed nested functions. A function is considered dynamically typed if it does not contain any type hints.

Default: False

Allow omission of a return type hint for __init__ if at least one argument is annotated. See mypy's documentation for additional details.

Default: False

Comma-separated list of decorators flake8-annotations should consider as dispatch decorators. Linting errors are suppressed for functions decorated with at least one of these functions.

Decorators are matched based on their attribute name. For example, "singledispatch" will match any of the following:

• import functools; @functools.singledispatch
• import functools as fnctls; @fnctls.singledispatch
• from functools import singledispatch; @singledispatch

NOTE: Deeper imports, such as a.b.singledispatch are not supported.

See: Generic Functions for additional information.

Default: "singledispatch, singledispatchmethod"

Comma-separated list of decorators flake8-annotations should consider as typing.overload decorators.

Decorators are matched based on their attribute name. For example, "overload" will match any of the following:

• import typing; @typing.overload
• import typing as t; @t.overload
• from typing import overload; @overload

NOTE: Deeper imports, such as a.b.overload are not supported.

See: The typing.overload Decorator for additional information.

Default: "overload"

Per the Python Glossary, a generic function is defined as:

A function composed of multiple functions implementing the same operation for different types. Which implementation should be used during a call is determined by the dispatch algorithm.

In the standard library we have some examples of decorators for implementing these generic functions: functools.singledispatch and functools.singledispatchmethod. In the spirit of the purpose of these decorators, errors for missing annotations for functions decorated with at least one of these are ignored.

For example, this code:

import functools

@functools.singledispatch
def foo(a):
    print(a)

@foo.register
def _(a: list) -> None:
    for idx, thing in enumerate(a):
        print(idx, thing)

Will not raise any linting errors for foo.

Decorator(s) to treat as defining generic functions may be specified by the --dispatch-decorators configuration option.

Per the typing documentation:

The @overload decorator allows describing functions and methods that support multiple different combinations of argument types. A series of @overload-decorated definitions must be followed by exactly one non-@overload-decorated definition (for the same function/method).

In the spirit of the purpose of this decorator, errors for missing annotations for non-@overload-decorated functions are ignored if they meet this criteria.

For example, this code:

import typing


@typing.overload
def foo(a: int) -> int:
    ...

def foo(a):
    ...

Will not raise linting errors for missing annotations for the arguments & return of the non-decorated foo definition.

Decorator(s) to treat as typing.overload may be specified by the --overload-decorators configuration option.

Support is provided for mixing argument-level and function-level type comments.

def foo(
    arg1,  # type: bool
    arg2,  # type: bool
):  # type: (...) -> bool
    pass

Note: If present, function-level type comments will override any argument-level type comments.

Partially type hinted functions are supported for non-static class methods.

For example:

class Foo:
    def __init__(self):
        # type: () -> None
        ...

    def bar(self, a):
        # type: (int) -> int
        ...

Will consider bar's self argument as unannotated and use the int type hint for a.

Partial type comments utilizing ellipses as placeholders is also supported:

def foo(arg1, arg2):
    # type: (bool) -> bool
    pass

Will show arg2 as missing a type hint.

def foo(arg1, arg2):
    # type: (..., bool) -> bool
    pass

Will show arg1 as missing a type hint.

Deprecation notice: Explicit support for utilization of ellipses as placeholders will be removed in version 3.0. See this issue for more information.

This project uses Poetry to manage dependencies. With your fork cloned to your local machine, you can install the project and its dependencies to create a development environment using:

$ poetry install

Note: An editable installation of flake8-annotations in the developer environment is required in order for the plugin to be registered for Flake8. By default, Poetry includes an editable install of the project itself when poetry install is invoked.

A pre-commit configuration is also provided to create a pre-commit hook so linting errors aren't committed:

$ pre-commit install

A pytest suite is provided, with coverage reporting from pytest-cov. A tox configuration is provided to test across all supported versions of Python. Testing will be skipped for Python versions that cannot be found.

$ tox

Details on missing coverage, including in the test suite, is provided in the report to allow the user to generate additional tests for full coverage.