document class methods using sphinx3 recursive doc

[veronicaguo]

• Closes class methods do not have doc pages #1055
• I am familiar with the contributing guidelines
• Adds description and name entries in the appropriate "what's new" file in docs/sphinx/source/whatsnew for all changes. Includes link to the GitHub Issue with :issue:`num` or this Pull Request with :pull:`num`. Includes contributor name and/or GitHub username (link with :ghuser:`user`).
• Pull request is nearly complete and ready for detailed review.

[kandersolar]

Thanks @veronicaguo for opening this PR! I just quickly clicked through some API Reference pages, so I don't have thoughts on #1055 but here are a couple points about bumping the sphinx version:

• The list of references is improved over what I remember from sphinx 2.x, but still different from what we're seeing on 1.8.5. I'm okay with the new formatting and could be convinced it's even an improvement over the current format. Example: https://pvlib-python--1075.org.readthedocs.build/en/1075/generated/pvlib.iam.martin_ruiz.html
• The in-text links are now bare numbers rather than numbers in brackets (e.g. 1 vs [1]). I personally prefer the bracketed version. If others feel the same, I think this custom CSS (see docs/sphinx/source/_static/) should bring back the brackets:

.footnote-reference::before {
  content: '[';
}
.footnote-reference::after {
  content: ']';
}

[veronicaguo]

Thanks @kanderso-nrel for the feedback!

• I agree that bracketed version for reference numbers looks better. will add the custom CSS.

[kandersolar]

Interestingly, the reference list on the clearsky page renders differently from the ones in the docstrings. I wonder if it's because it's using citations instead of footnotes?

https://pvlib-python--1075.org.readthedocs.build/en/1075/clearsky.html#references

Not related to this PR, just something I saw while clicking through the pages: the comparison with the MATLAB version could use a refresh, specifically the details about our continuous integration setup.

[wholmgren]

Thanks @veronicaguo!

Interestingly, the reference list on the clearsky page renders differently from the ones in the docstrings. I wonder if it's because it's using citations instead of footnotes?

Also on the forecasting page: https://pvlib-python--1075.org.readthedocs.build/en/1075/forecasts.html#lar16

I have a mild preference for citation style for the narrative documentation.

[wholmgren]

thanks @veronicaguo. @kanderso-nrel ok to merge?

[kandersolar]

LGTM!