Python Docstring: raise vs. raises

Question

I use the PyCharm IDE which assists with crafting PEP0257-compliant docstrings. It provides two attributes I don't entirely understand the distinction/use between:

• :raise Exception: exception explanation here
• :raises Exception: exception explanation here

When would I use raise as opposes to raises in my docstring? Specifically, if a class required an argument that was not provided and raises a TypeError, which should be used to document that?

Answer 1

TL;DR

raises is used to describe the possible exceptions being raised. raise is recognized by Sphinx when running autodoc and is the same as raises.

Full Explanation

PyCharm helps in using a few different styles of docstring comments.

Three which I often use are:

1. NumPy Format
2. Google Format
3. Sphinx (much more than a format)

In all of these there is a special section for Raises which you can see in an older version of the PyCharm code tests:

1. Simple NumPy
2. Simple Google

The implementation for SphinxDocString we can see here there there are numerous keywords which can be recognized. Those tags then link to the list of RAISES_TAGS which can be found here.

I hope this information is useful.

Answer 2

You must use raises to describe exceptions raised by your method/class.

:raises:     Exception: Explanation here. 

For example, for a ValueError exception:

:raises:     ValueError: if fft_data is empty.