Details
In graph theory, a cubic graph is a graph in which every vertex has degree three—equivalently, a 3-regular graph. Such graphs are also known as trivalent graphs.
This dataset employs differential encoding to compress graphs by storing only their differences. The compressed format uses 77.5% less storage space than storing full adjacency list representations for all 41,301 graphs.
Shortcode representation: Each graph is encoded as a sequence of vertex numbers representing edges (only higher-numbered neighbors per vertex). For instance, decoding (2 3 4 3 4 4) for a 4-vertex, 3-regular graph:
Vertex 1: reads 2 3 4 -> Edge: 1-2, Edge: 1-3, Edge: 1-4
Vertex 2: reads 3 4 -> Edge: 2-3, Edge: 2-4
Vertex 3: reads 4 -> Edge: 3-4
Vertex 4: reads nothing (no more entries)
Prefix compression: each graph stores one byte indicating how many leading bytes match the previous graph, and only the remaining differing bytes.Example:

The the two adjacency lists are (
2 3 4 5 3 4 5 6 7 6 7 6 7 7 ) and (
2 3 4 5 3 4 6 5 7 6 7 6 7 7). This compression is very effective when consecutive graphs share long common prefixes, which happens frequently in the systematic generation process.
The conversion algorithm leverages
NumericArray for efficient data storage and faster processing.