Matplotlib Website Feedback

This is a space to let us know what works for you on the site, what’s frustrating, and what you’d like to see. While we can’t accommodate every request, constructive feedback is much appreciated.

Doubt this should be top level, but an easier way to find the What’s New section would probably be a good way to make new features more discoverable.

I would love it, if the “Release” field here: would be a dropdown menu to select docs for the different releases of matplotlib

1 Like

@maxnoe Thanks for the feedback. I’m not sure we will go with that exact solution, but making it easy to both tell what version of the docs you are looking at and to switch versions is something that we are aware we need to do!

One slightly interesting problem that comes with this is that we sometimes rename examples which changes their url. Some automated way of sorting out the “correct” redirect (and applying it retroactively) would be an interesting data science problem that someone should work on :wink: .

The documentation search page fails to load results when viewed from the Jupyterlab Matplotlib reference pane (a cross-site iframe).

The issue is that matplotlib[.]org responds to the CORS preflight OPTIONS request with a 405 Method Not Allowed, even though a GET request would return the correct Access-Control-Allow-Origin: * header.

To fix this, the web server should be configured to respond to an OPTIONS request with 200 OK (or similar) and the headers Access-Control-Allow-Origin: * and Access-Control-Allow-Methods: POST, GET, OPTIONS (or similar). See also

The issue was also exacerbated by a bug in the search page, that prevented the searchindex.js from loading using an alternative method with a <script> tag. I have submitted a Pull Request for that. It does however not fix loading the snippets under each search result, which also uses cross-site XHR.

See also PR #15649: “Fix searchindex.js loading when ajax fails (because e.g. CORS in embedded iframes)”

I would like to see some showcase figures directly on main landing page.

What do you mean by “showcase”? On one hand that may mean an expanded gallery of the “basic” plots or on the other may mean some plots the exploit the full flexibility / power of Matplotlib.