Thank you so much for looking at this! I've made a couple of suggestions but, in answer to your main point, in general we'd like the sphinx documentation to look much like our current doxygen-based documentation. Things to consider are:

1. As well as auto-generated documentation, we also have a bunch of dox pages we have written - it looks like doxygen includes them in its xml output but I'm not at all clear how to reference them
2. Currently we're using doxypypy to generate documentation for GeNN's python interface and these are somewhat intermingled with the other code documentation. It might make sense to instead build these directly using sphinx.

I also spotted https://github.com/vovkos/doxyrest which might be worth trying as an alternative to breathe

Hi, as you suggested, I used doxyrest as a bridge between sphinx and doxygen. All the doxyrest related files are located in doxyrest/ directory. And I have added doxyrest/build/ dir for your reference to the generated html files (which I'm worried about). I've integrated doxyrest pipeline to the project and finally when sphinx generate html files from .rsts (which is populated by doxygen's XML file [in dir: doxyrest/xml] with the help of doxyrest), I wouldn't able to see any document elements 😞 . (However when I run doxygen/genn-doxygen.conf, I get html, xml and pdf files, in which when I try to open index.html, I end up with same empty documentation page). I'm pretty sure that, I am missing a small piece somewhere but hard to figure out.

And regarding the doxyrest setup I followed the steps mentioned in:

1. https://github.com/vovkos/doxyrest
2. https://vovkos.github.io/doxyrest/build-guide/

Ok, after a lot of fiddling with readthedocs/Jenkins, this is now basically working on the readthedocs2 branch (the best thing may be to update your fork and work from there). Outstanding things that need investigating are:

• It would be nice if the version of GeNN was included in the sphinx documentation - this is specified globally in version.txt so you just need to read it in with Python
• Static images aren't getting copied from doxygen/images/ into wherever sphinx wants them
• Release notes are somehow turned into top-level menu items - I imagine this needs fixing in the doxygen
• There are a lot of warnings being issued at various stages - are these causing any issues?
• The order of pages in doxygen was intentional but they are now ordered alphabetically - is there anything that can be done about this?
• The bibleography seems broken - is this something it's possible to fix? It's possibly a problem with the latex install on our build machine

Oh! I didn't get a notification at all. I randomly surfed through here to check the updates on ReadTheDocs and surprised to see many things that I did not follow for many days. Apologies for that. I'll look into the things mentioned 👍

No worries, github notifications can be a bit flaky at times!

Codecov Report

Merging #250 into master will decrease coverage by 0.72%.
The diff coverage is 86.64%.

@@            Coverage Diff             @@
##           master     #250      +/-   ##
==========================================
- Coverage   87.64%   86.92%   -0.73%     
==========================================
  Files          62      121      +59     
  Lines        8363    23474   +15111     
==========================================
+ Hits         7330    20405   +13075     
- Misses       1033     3069    +2036     

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d1c3d36...2e3783d. Read the comment docs.