Google and Numpy

Pydoctor now supports numpydoc and google style docstrings!

Docstrings will be first converted to reStructuredText and then parsed with docutils. Any supported reST markup can be use to supplement google-style or numpy-style markup.

The main difference between the two styles is that Google uses indentation to separate sections, whereas NumPy uses underlines. This means that 2 blank lines are needed to end a NumPy section that is followed by a regular paragraph (i.e. not another section header)


We have forked and enhanced the napoleon Sphinx extension.

For more information, refer to pydoctor.napoleon documentation.

For complete markup details, refer to the Google style or NumpyDoc style reference documentation.


List of supported sections:
  • Args, Arguments, Parameters

  • Keyword Args, Keyword Arguments

  • Return(s), Yield(s) (if you use type annotations a Returns section will always be present)

  • Raise(s), Warn(s)

  • See Also, See

  • Example(s)

  • Note(s), Warning(s) and other admonitions

  • Attributes (Items will be translated into ivar fields.)

Sections supported on a “best effort” basis:
  • Methods: Items will be included into a generic “Methods” section.

  • References: Rendered as a generic section.

  • Other Parameters, Receive(s): Parameters described in those sections will be merged with regular parameters.