GraphLab Algorithms

Repository containing various graph algorithm implementations using GraphLab.

Installation

To install GraphLab follow steps as shown in installation guide.

Running

Activate the conda environment and run any of the python scripts.

Activating the Virtual Environment

source activate gl-env

Running the Algorithms

• pagerank, connected-components, triangle-counting

  python script.py W /path/to/edgelist
  

  • where W is the number of workers (cores)
  • e.g. python pagerank.py 8 ~/data/soc-LiveJournal1.txt (to execute on 8 cores)

• single-source-shortest-path, reachability (sp version)

  python script.py W /path/to/edgelist S T
  

  • where S the source vertex id and T the target vertex id
  • the DFS and BFS implementations of reachability also receive an additional argument M (max_depth) to limit the search depth

• neighbors

  python neighbors.py W /path/to/edgelist S D
  

  • where S the source vertex id and D the neighborhood degree

SW Versions

• Python 2.7.1
• Anaconda 4.3.30
• pip 9.0.1 (Do not upgrade to latest, otherwise you might be facing this error)

License

As stated in LICENSE.