Add a few more BFS-based algorithms

[eriknw]

BFS is a workhorse! There are still many BFS-based algorithms still to do.

My main TODO is to benchmark these to give them a proper shakedown. I like to shake down most algorithms we add. Also, I wonder whether we should use A.T or AT in a couple places. I also wonder if we can come up with a heuristic to detect negative cycles.

It may be nice to know whether directed graphs are structurally symmetric. If we knew this, we could improve is_weakly_connected and other algorithms.

Also, there may be a couple NetworkX bugs to fix or behaviors to match.

Implemented:

• Components
  • is_connected
  • is_weakly_connected
  • node_connected_component
• Shortest Paths
  • all_pairs_shortest_path_length
  • negative_edge_cycle
  • single_source_shortest_path_length
  • single_target_shortest_path_length
• Traversal
  • bfs_layers
  • descendants_at_distance

[eriknw]

I'm also not sure if these are the best algorithms for connected component-related functions. Connected components are an important class of algorithms that we should add soon (for example, we have an implementation of FastSV in a Jupyter notebook that we can port over), so perhaps we'll revisit these algorithms then.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 87.71429% with 43 lines in your changes missing coverage. Please review.
✅ Project coverage is 71.62%. Comparing base (6de1fd6) to head (029a639).
⚠️ Report is 24 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #51 +/- ##
==========================================
+ Coverage 69.16% 71.62% +2.46% 
==========================================
Files 77 89 +12 Lines 3152 3461 +309 Branches 602 642 +40 ==========================================
+ Hits 2180 2479 +299 + Misses 786 783 -3 - Partials 186 199 +13 

🚀 New features to boost your workflow:

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

[eriknw]

Merging so we can do a release before PyConUS. May be able to benchmark and shakedown some algorithms later, but they're probably good enough.