Index of /public/dataset/graph/latest/compressed

	Name	Last modified	Size	Description
	Parent Directory		-
	all.pid2node.csv	2019-08-27 16:47	783G
	all.node2pid.csv	2019-08-27 13:45	555G
	all.pid2node.bin	2019-09-05 12:29	326G
	all.node2pid.bin	2019-09-16 18:47	239G
	all.graph	2019-08-03 13:15	91G
	all.order	2019-08-03 13:40	87G
	all-transposed.graph	2019-08-03 16:50	83G
	all-transposed.obl	2019-08-03 16:53	12G
	all.obl	2019-08-03 13:19	11G
	all-transposed.offsets	2019-08-03 16:55	11G
	all.offsets	2019-08-03 13:21	10G
	all.node2type.map	2019-08-27 13:46	4.1G
	all.mph	2019-08-03 13:16	3.1G
	all.indegree	2019-08-03 13:16	357M
	stdout	2019-08-03 12:56	19M
	all.outdegree	2019-08-03 13:40	2.0M
	stderr	2019-08-03 12:56	10K
	all-transposed.properties	2019-08-03 16:55	1.5K
	all.properties	2019-08-03 13:40	1.5K
	all.stats	2019-08-03 13:40	828
	README.md	2019-05-30 07:42	792

Graph compression output

These are the output directories from experiments running the WebGraph framework to compress the Software Heritage graph datasets. Each directory is the output for the corresponding dataset as input.

• Subdirectories: bv/ BVGraph, bv_llp/ BVGraph with LLP ordering, bv_bfs/ BVGraph with BFS ordering, bv_sym/ BVGraph symmetrized.
• A graph is stored as a .graph, .offsets and .properties files (with a .obl file to load the graph faster).
• Statistics about the graph are stored in .properties, .indegree, .indegrees, .outdegree, .outdegrees, and .stats files.

You can find the compression script and Dockerfile here: https://forge.softwareheritage.org/source/graph-compression/browse/master/compression/