v0.5.3.1: Minor bug fix

• Update POT api to the latest version.
• Networkx's Karate club graph now comes with edge weight, update tutorial and test for that.
• Forman computation's log can be mute now.

Faster OTD computation (up to 10x).

• Switch OTD computation from cvxpy to ot.emd2 since it's way faster (up to 10x) for OTD cases within neighbor counts 2000 x 2000.
• Add a new computation method OTDSinkhornMix that uses ot.emd2 to compute OTD for smaller neighbors (less than _OTDSinkhorn_threshold=2000 in default setting). This is because Sinkhorn works faster for larger cases, for smaller cases, the iterating process makes it slower than directly compute the ot.emd2 written in C.
• Update the default nbr_topk to be 3000 because of the faster computations.
• Suppressed edge weight to zero message from warning to trace since the Ricci flow "converges" faster with exact Wasserstein distance.

v0.5.2.1: Bug fix for network neighbor iterator support

Merge pull request #22 from saibalmars/nk_fix

Networkit fix: update to support the nk's new neighbor iterator

v0.5.2: Support weighted Forman-Ricci Curvature

• Reimplement Forman-Ricci curvature for the weighted graph follows the definition in Comparative analysis of two discretizations of Ricci curvature for complex networks.
• Add an example for Forman-Ricci curvature in the tutorial.

v0.5.1: Merge pull request #18 from saibalmars/update_tutorial

• Update to support examples in the tutorial.
• Reformat code.

v0.5.0

Add ricci_community for auto community detection.

v0.4.5

• Performance tune, the Ricci curvature computation speed is up 33% compare with v0.4.4.
• The pairwise shortest path is still not that idea, now it's an option only.

v0.4.4

• Option for shortest path computation

The Bidirectional Dijkstra pairwise shortest path by Networkit having some problem on a larger graph, add an option for the user to select the shortest path engine, default is by all pairs shortest path, but may require more memory for larger graph.

v0.4.3.1

• Some minor fix to mute the directed graph from networkit.

v0.4.3

• Now optimal transportation distance and Sinkhorn distance can be limited to the top k most important neighbors (by edge weights) to speed up the distance computation.