Adopt the Xanadu Sphinx Theme

[Mandrenkov]

Context:

Currently, all Xanadu OSS projects include their own version of the Xanadu Sphinx theme; however, the Xanadu Sphinx Theme repository is now publicly available and is the preferred way to access the Xanadu CSS theme and Sphinx directives.

Description of the Change:

• Deleted the docs/xanadu_theme directory.
• Updated docs/requirements.txt and docs/conf.py to reference and use the (centralized) Xanadu Sphinx Theme.
• Added a Python 3.10 classifier to setup.py.

Benefits:

• The size of this repository is smaller, meaning git clone will run faster.
• Updates to the Xanadu Sphinx theme no longer have to be manually ported to this repository.

Possible Drawbacks:
None.

Related GitHub Issues:
Depends on XanaduAI/xanadu-sphinx-theme#2.

[github-actions]

Test Report (C++) on Ubuntu

    1 files  ±0      1 suites  ±0   0s ⏱️ ±0s
524 tests ±0  524 ✔️ ±0  0 💤 ±0  0 ❌ ±0 
873 runs  ±0  873 ✔️ ±0  0 💤 ±0  0 ❌ ±0 

Results for commit 84d75aa. ± Comparison against base commit c74fd61.

♻️ This comment has been updated with latest results.

[github-actions]

Test Report (Python) on Ubuntu

    1 files  ±0      1 suites  ±0   6s ⏱️ -1s
300 tests ±0  300 ✔️ ±0  0 💤 ±0  0 ❌ ±0 

Results for commit 84d75aa. ± Comparison against base commit c74fd61.

♻️ This comment has been updated with latest results.

[doctorperceptron]

Code looks 🍾. Haven't tested.

[thisac]

Looks good! Attempted to build/test it, but wasn't able to due to the makefile. It seems to be attempting to install requirements.txt (using pip) when simply running make docs. This might not be such a good idea, since, at least in my case, it will break my environment. 🤔

[Mandrenkov]

Thanks for the reviews, @doctorperceptron and @thisac!

Looks good! Attempted to build/test it, but wasn't able to due to the makefile. It seems to be attempting to install requirements.txt (using pip) when simply running make docs. This might not be such a good idea, since, at least in my case, it will break my environment. 🤔

Yeah, it also took a few tries for me to build Jet locally. I think we made a mistake by trying to automate too many things in the Makefile since it introduced some assumptions which don't necessarily hold in each environment. That said, I think we should also acknowledge that make docs is a complicated process:

1. First, we need to build the Python bindings for the C++ part of Jet.
2. Then, we need to install the Python part of Jet.
3. Finally, we need to generate the Sphinx documentation for both the C++ and Python parts of Jet.

We can usually skip steps (1) and (2) for pure-Python projects but that is not the case here.

[josh146]

Flyby comment, and I think this is not high priority at all, but you can set up sphinx to mock out the C++ libraries (I think this is what we did with the walrus back when it included C++) to avoid having to build the bindings.

The C++ part is still documented via doxygen/exhale/breathe

[Mandrenkov]

Flyby comment, and I think this is not high priority at all, but you can set up sphinx to mock out the C++ libraries (I think this is what we did with the walrus back when it included C++) to avoid having to build the bindings.

The C++ part is still documented via doxygen/exhale/breathe

Thanks for the comment, @josh146! Does the documented C++ part still include the bindings though? It would be a shame to discard that part of the documentation.

[josh146]

Does the documented C++ part still include the bindings though? It would be a shame to discard that part of the documentation.

Ah The Walrus used Cython (not pybind11), so the C++ bindings were read statically by Sphinx... in retrospect not sure if that has implications here.