cython-lint

Everything flake8 used to do (by accident), plus much more.

A tool and pre-commit hook to lint Cython files.

Used by

Here's some major projects using cython-lint - is yours missing? Feel free to open a pull request!

• arrow-adbc
• ansible-pylibssh
• arrow
• CuPy
• dpctl
• frozenlist
• mkl_fft
• mkl-service
• NumPy
• open library
• pandas
• pymatgen
• PyRFC
• RAPIDS cuspatial
• RAPIDS cudf
• RAPIDS Memory Manager
• sagemath/sage
• scikit-image
• scikit-learn
• scipy
• spaCy
• yarl
• yt

Installation

$ pip install cython-lint

Usage as a pre-commit hook

See pre-commit for instructions

Sample .pre-commit-config.yaml:

-   repo: https://github.com/MarcoGorelli/cython-lint
    rev: v0.20.0
    hooks:
    -   id: cython-lint
    -   id: double-quote-cython-strings

Command-line example

$ cython-lint my_file_1.pyx my_file_2.pyx
my_file_1.pyx:54:5: 'get_conversion_factor' imported but unused
my_file_2.pyx:1112:38: 'mod' defined but unused (try prefixing with underscore?)
my_file_3.pyx:4:9: dangerous default value!
my_file_3.pyx:5:9: comma after base type in definition

Configuration

The following configuration options are available:

• exclude lines by including a # no-cython-lint comment (analogous to # noqa in flake8);

as well as the following command-line arguments:

• --max-line-length to control the maximum line length used by pycodestyle;
• --no-pycodestyle if you don't want the pycodestyle checks;
• --ban-relative-imports if you want to ban relative imports;
• --ignore if you want to ignore some specific pycodestyle checks;
• --files to pass a Regex pattern with which to match files to include;
• --exclude to pass a Regex pattern with which to match files to exclude.

Configuration can be set project-wide in a .cython-lint.toml or pyproject.toml file at the root of the project. Here's an example:

[tool.cython-lint]
max-line-length = 88
ignore = ['E503', 'E504']
exclude = 'my_project/excluded_cython_file.pyx'

Which checks are implemented?

• assert statement with tuple condition (always true...)
• comma after base type definition (e.g. cdef ndarray, arr)
• comparison between constants
• dangerous default value
• dict key repeated
• dict key variable repeated
• f-string without placeholders
• if-statement with tuple condition (always true...)
• late-binding closures https://docs.python-guide.org/writing/gotchas/#late-binding-closures
• outer loop variable name overwritten by inner loop variable name
• pointless string statement
• pycodestyle nitpicks (which you can turn off with --no-pycodestyle)
• repeated element in set
• .strip, .rstrip, or .lstrip used with repeated characters
• unnecessary list index lookup
• unnecessary import alias
• variable defined but unused
• variable imported but unused
• unnecessary list/set/dict + generator (just use a comprehension)

In addition, the following automated fixers are implemented:

• double-quote-cython-strings (replace single quotes with double quotes, like the black formatter does)

More to come! Requests welcome!