Replace sage.misc.sphinxify with docrepr

[mkoeppe]

https://pypi.org/project/docrepr/

"Docrepr renders Python docstrings to HTML with Sphinx. It can generate rich and plain representations of docstrings, alongside additional metadata about the object to which the docstring belongs. It is based on the sphinxify module developed by Tim Dumol for the Sage Notebook..."

See also https://blog.jupyter.org/inspector-jupyterlab-404cce3e1df6
and https://github.com/ipython/ipython/blob/master/docs/source/whatsnew/version5.rst#provisional-changes on IPython's sphinxify_docstring, enable_html_pager options

See also spyder-ide/docrepr#20 on history

CC: @kwankyu @jhpalmieri @egourgoulhon @antonio-rojas

Component: user interface

Branch/Commit: u/mkoeppe/replace_sage_misc_sphinxify_with_docrepr @ 6275755

Issue created by migration from https://trac.sagemath.org/ticket/33682

[mkoeppe]

Branch: u/mkoeppe/replace_sage_misc_sphinxify_with_docrepr

[mkoeppe]

Commit: 6275755

[mkoeppe]

New commits:

6275755

build/pkgs/docrepr: New

[tornaria]

Doesn't seem to be widely available: https://repology.org/project/python:docrepr/versions

Also: last commit 2 years ago (https://github.com/spyder-ide/docrepr).

Sphinx seems well available and afaik it's used by python docs. Maybe sphinxify can be implemented using just docutils, but is it worth it?

OTOH, sage runs just fine without sphinx, the only difference afaict is that you get help as source rst instead of processed text, but it's not really a problem. As pointed out in #35493, the help for something like CliffordAlgebra is ugly regardless of sphinx or not sphinx.

[mkoeppe]

Thanks for taking a look at this and giving this update.