Parallelization of Degree Centrality

[SKADE2303]

1. Degree Centrality Implementation

Distributes degree centrality computation across multiple workers:

• Function Signature: Maintains NetworkX compatibility with added parallelization parameter
• Normalization: Correctly implements degree normalization by dividing by (n-1)
• Special Cases: Handles single-node graphs and empty graphs appropriately
• Technical Details
• Directed Graphs: Computes in-degree centrality (matching NetworkX behavior)
• Multigraphs: Correctly handles parallel edges during degree calculation
• Worker Distribution: Balances node processing across available workers

2. Testing Framework

Tests verify correctness and performance across multiple graph configurations:

• Graph Types: Tests directed, undirected, weighted, and multigraphs
• Size Variations: From empty graphs to large networks (1000+ nodes)
• Topology Tests: Path graphs, random graphs, disconnected components
• Special Case Tests: Self-loops, isolated nodes, single-node graphs
• Chunking Tests: Both default and custom chunking strategies

Future Improvements

Another idea would be to add an approximate version that samples paths instead of calculating every single shortest path - would be way faster for those huge graphs I was testing with. I'm also thinking it'd be really handy to implement an incremental update feature, so you don't waste time recalculating everything when your graph only changes slightly.

[Schefflera-Arboricola]

clossness_centrality is already implemented in PR #72 . Please don't duplicate the effort. Also, could you please clean up this PR and remove the changes outside the scope of this PR(i.e. parallel implementation of degree_centrality)? Please mark the PR "Ready for review" once it's ready.

[SKADE2303]

Thanks for the clarification, @Schefflera-Arboricola. I apologize for the oversight. I wasn’t aware that closeness_centrality was already implemented in PR #72, so I’ll go ahead and remove those changes to avoid any duplication. I’ll also clean up this PR to ensure it strictly focuses on the parallel implementation of degree_centrality and make sure the tests reflect that. Once everything is updated, I’ll mark the PR as "Ready for review." Appreciate the clarification!

[akshitasure12]

Hey @SKADE2303, is this PR being worked on currently? Let me know, thanks!

[Schefflera-Arboricola]

closing due to inactivity-- feel free to re-open if/when you get the bandwidth to work on this in the future. Thanks!