This is an MIT licensed flake8 plugin for validating Python docstrings markup as reStructuredText (RST) using the Python library docutils. It is available to install from the Python Package Index (PyPI).

This is based heavily off pydocstyle (which is also MIT licensed), which has a flake8 plugin called flake8-docstrings, see:

• https://github.com/PyCQA/pydocstyle
• https://github.com/PyCQA/flake8-docstrings

The reStructuredText (RST) validation is done by calling docutils via Todd Wolfson's restructuredtext-lint code:

• http://docutils.sourceforge.net/
• https://github.com/twolfson/restructuredtext-lint

I recommend you also install the related flake8-docstrings plugin, which brings the pydocstyle checks into flake8. This checks things like missing docstrings, and other recommendations from PEP 257 Docstring Conventions.

You may also wish to install the related flake8 plugin flake8-rst which can check the Python style of doctest formatted snippets of Python code within your *.rst files or the docstrings within your *.py files.

Early versions of flake8 assumed a single character prefix for the validation codes, which became problematic with collisions in the plugin ecosystem. Since v3.0, flake8 has supported longer prefixes therefore this plugin uses RST as its prefix.

Internally we use docutils for RST validation, which has this to say in PEP258:

• Level-0, "DEBUG": an internal reporting issue. There is no effect on the processing. Level-0 system messages are handled separately from the others.
• Level-1, "INFO": a minor issue that can be ignored. There is little or no effect on the processing. Typically level-1 system messages are not reported.
• Level-2, "WARNING": an issue that should be addressed. If ignored, there may be minor problems with the output. Typically level-2 system messages are reported but do not halt processing
• Level-3, "ERROR": a major issue that should be addressed. If ignored, the output will contain unpredictable errors. Typically level-3 system messages are reported but do not halt processing
• Level-4, "SEVERE": a critical error that must be addressed. Typically level-4 system messages are turned into exceptions which halt processing. If ignored, the output will contain severe errors.

The docutils "DEBUG" level messages are not reported, and the plugin currently ignores the "INFO" level messages.

Within each category, the individual messages are mapped to flake8 codes using one hundred times the level. i.e. Validation codes RST4## are severe or critical errors in RST validation, RST3## are major errors, RST2## are warnings, and while currently not yet used, RST1## would be information only.

Warning codes:

Major error codes:

Severe or critial error codes:

Codes ending 99, for example RST499, indicate a previously unseen validation error for which we have yet to assign a unique validation code in the assocated range, which would be RST4## in this example. If you see one of these codes, please report it on our GitHub issue tracker, ideally with an example we can use for testing.

Codes starting RST9## indicate there was a problem parsing the Python file in order to extract the docstrings, or in processing the contents.

Python 3.6 or later is recommended, but Python 3.3 onwards is also supported. Earlier versions did support Python 2.7, use v0.0.14 if required.

We recommend installing the plugin using pip, which handles the dependencies:

$ pip install flake8-rst-docstrings

Alternatively, if you are using the Anaconda packaging system, the following command will install the plugin with its dependencies:

$ conda install -c conda-forge flake8-rst-docstrings

The new validator should be automatically included when using flake8 which may now report additional validation codes starting with RST (as defined above). For example:

$ flake8 example.py

You can request only the RST codes be shown using:

$ flake8 --select RST example.py

Similarly you might add particular RST validation codes to your flake8 configuration file's select or ignore list.

Note in addition to the RST prefix alone you can use partial codes like RST2 meaning RST200, RST201, ... and so on.

We assume you are familiar with flake8 configuration.

If you are using Sphinx or other extensions to reStructuredText, you will want to define any additional directives or roles you are using to avoid false positive RST303 and RST304 violations. You may also need to ingore RST307 if using Sphinx directives with arguments.

You can set these at the command line if you wish:

$ flake8 --rst-roles class,func,ref --rst-directives envvar,exception ...

We recommend recording these settings in your flake8 configuration, for example in your .flake8, setup.cfg, or tox.ini file, e.g.:

[flake8]
rst-roles =
    class,
    func,
    ref,
rst-directives =
    envvar,
    exception,
extend-ignore =
    RST307,
    # ...

Note that flake8 allows splitting the comma separated lists over multiple lines, and allows including of hash comment lines.

If you are using the Google Python Style you will sometimes get unwanted warnings from this plugin - particularly in the argument descriptions - as it does not use strict RST. We therefore currently suggest ignoring some of the violation codes:

[flake8]
extend-ignore =
    # Google Python style is not RST until after processed by Napoleon
    # See https://github.com/peterjc/flake8-rst-docstrings/issues/17
    RST201,RST203,RST301,

This plugin is on GitHub at https://github.com/peterjc/flake8-rst-docstrings

To make a new release once tested locally and on TravisCI:

$ git tag vX.Y.Z
$ python setup.py sdist --formats=gztar
$ git push origin master --tags
$ twine upload dist/flake8-rst-docstrings-X.Y.Z.tar.gz

The PyPI upload should trigger an automated pull request updating the flake8-rst-docstrings conda-forge recipe.

• Have the "INFO" level RST1## codes available but ignored by default?
• Can we call docutils rather than bundle a copy of their parser code?
• Create a full test suite and use this for continuous integration.
• Test with raw mode docstrings and slash-escaped characters.