dssg / aequitas

Bias Auditing & Fair ML Toolkit
http://www.datasciencepublicpolicy.org/aequitas/
MIT License
656 stars 110 forks source link

Update GitHub pages #154

Open sgpjesus opened 5 months ago

sgpjesus commented 5 months ago

The current solution for documentation generation for GitHub pages (link here) has some flaws. These are:

  1. New packages, classes, etc. are not included and the process to include them is unclear.
  2. The build of documents is stored in the package, cluttering the docs directory.
  3. The code documentation lacks some aspects (code examples, type hints).
  4. The license is of 2018.
  5. The style is outdated.
  6. There is no landing page.

The current solution is based on sphinx (used in many projects). How do they organize it? Following the example of sklearn, there is a mix of both generated and explicitly written documentation in the rst files.

Sphinx is the basis of documents for many packages (torch is a giant one but super messy, sklearn in between, fairlearn is smaller and pretty clean).

Style is is normaly stored in _static/css or have a dedicated directory for the effect (the case of Sklearn).

Landing pages are a bit weird to see how they are done.

sgpjesus commented 2 months ago

Tasks: