Library of graph algorithms we use with Apache Flink. Documentation can be found in the specific algorithm packages. The licence is GPLv3, so please feel free to use the algorithms in your own project.
Extracts Business Transactions Graphs (BTGs) from an integrated instance graph (IIG) as described in BIIIG: Enabling Business Intelligence with Integrated Instance Graphs.
Finds communities inside networks by propagating vertex labels (communities). Vertices migrate to the community represented by the majority of labels sent by its neighbours. Our implementation adds a stabilization mechanism and avoids oscillating states. See Wikipedia (Label propagation).
Partitions a graph using label propagation as described in Adaptive Partitioning of Large-Scale Dynamic Graphs. The implementation uses flink (gelly) aggregators to ensure a balanced partition load.