The primary goal of this repo is to demonstrate the new automatic package detection facility for Python API documentation available in Sphinx 3.1+. See this StackOverflow answer for context.
A secondary goal is to showcase the integration of Jupyter Notebooks with Sphinx.
The resulting Sphinx-built HTML doc set is available to view on ReadTheDocs.
From Sphinx 3.1, sphinx.ext.autosummary has a :recursive: option that iterates over a Python
package automatically, so you no longer have to hard code all your module names, or integrate a 3rd party
extension to provide this functionality. Scroll down
this Sphinx documentation
for more information.
This repo demonstrates how to:
-
Point Sphinx at the top of a Python source code tree, and have it automatically find all the modules in the package, however deeply nested.
-
For each module, list the attributes, functions, classes and exceptions in that module in summary tables.
-
For each entry in a summary table, create a hyperlink to a new page displaying the extracted docstrings for that attribute, function, class or exception.
-
For each class, document (my choice of) inheritance, public members, inherited members, and special members such as
__call__. Other choices are available.
For more information, start with docs/README.
Jupyter Notebooks blend live code, text and visualizations. It’s often useful to integrate them alongside API documentation to provide tutorial-style material.
For more information, start with docs/notebooks/README.
Since it’s hosted on Readthedocs, the default theme (stylesheet) for the HTML doc set is sphinx-rtd-theme.
There's a small alteration to this theme’s CSS to make the page width slightly wider; see docs/_static/readthedocs-custom.css.
Other themes are available; settings to switch over to the popular pydata-sphinx-theme are commented out in docs/conf.py.