use rustworkx instead of networkx except the plotting

[selmanozleyen]

This is part of my larger goal to replace process based parallelization. I found rustworkx easy to install and fast. That's why I made a PR long ago.

This pull request migrates the group centrality computation in centrality_scores from NetworkX to the high-performance rustworkx library, and introduces a custom, Numba-accelerated clustering coefficient kernel. It also ensures that the results remain numerically consistent with the previous implementation by adding parity tests. The main changes are as follows:

Dependency and Documentation Updates:

• Added rustworkx>=0.18.0 as a dependency in pyproject.toml and included a corresponding citation in references.bib. [1] [2]
• Updated docstrings to reference rustworkx for group centrality measures and clarified citations for both rustworkx and networkx. [1] [2]

Core Implementation Changes:

• Replaced NetworkX-based group centrality calculations in centrality_scores with rustworkx equivalents, and replaced the clustering coefficient computation with a custom, parallelized Numba kernel for improved performance. [1] [2]
• Added a new _build_graph utility to convert connectivity matrices into both rustworkx.PyGraph and a CSR matrix for use in centrality and clustering calculations.

Testing:

• Introduced a test to verify numerical parity between the new rustworkx/Numba implementation and the original NetworkX-based results for all group centrality measures.

[codecov]

Codecov Report

❌ Patch coverage is 50.00000% with 23 lines in your changes missing coverage. Please review.
✅ Project coverage is 77.11%. Comparing base (ca07e64) to head (74f6c34).

Files with missing lines	Patch %	Lines
src/squidpy/gr/_nhood.py	50.00%	23 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1252      +/-   ##
==========================================
- Coverage   77.24%   77.11%   -0.13%     
==========================================
  Files          63       63              
  Lines        9378     9418      +40     
  Branches     1579     1585       +6     
==========================================
+ Hits         7244     7263      +19     
- Misses       1532     1554      +22     
+ Partials      602      601       -1     

Files with missing lines	Coverage Δ
src/squidpy/_docs.py	94.80% <ø> (ø)
src/squidpy/gr/_nhood.py	75.00% <50.00%> (-6.71%)	⬇️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.