This is a toy single-machine implementation of Pregel, Google's system for running large-scale graph algorithms on a large cluster of machines.

For the original paper, see http://dl.acm.org/citation.cfm?id=1807184

For an illustration of how to use this implementation of Pregel, see the example code in pagerank.py.

This repo is forked from https://github.com/mnielsen/Pregel, and a few additions have been made. A tutorial.ipynb Jupyter notebook has been added.

For information on the original toy implementation, see http://michaelnielsen.org/ddi/pregel/

Examples:

• clustering
• partition, as in https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/Partition.pdf (they say "pure label propagation can be implement- ed using the vertex-centric computation model in 2 lines of code")
• semi-clustering

Features:

• aggregators
• combiners
• edge weights

Fixes:

• bug mentioned in PageRank