Directed labeled multigraphs and directed labeled multigraphs enriched with information about which graph patterns match a graph and how they match a graph

This dataset contains all directed labeled multigraphs that consist of one or two triples and that were generated from a set of 5 terms.

The graphs are represented as RDF data in ntriples format.

The graphs 1-125 consist of a single triple. The graphs 126-4625 consist of two triples.

Each graph also exists in an enriched version, where each graph pattern that matches the graph is made explicit in the graph and it is also made explicit how it matches the graph.

Enrichment means that for each graph pattern that matches a graph, a node is introduced that is labeled with the name of that graph pattern. Each term that occurs in that graph pattern is linked to that graph pattern node with an edge labeled "occurs-in". For each term t in the graph that can be bound to a variable v according to a match of that graph pattern an edge will be created from the node t to the graph pattern node where the edge is labeled with "bound-to-v".

The generation of the dataset and the enrichment procedure is described in more detail in a paper that is currently under double-blind submission. We will link to the paper once it is accepted for publication.

Example: the graph #100 (file 100.nt)

.

Example: the graph #2000 (file 2000.nt)

. .

Example: the enriched graph #100 (file 100-extended.nt)

. . . . . . . . . . . . . . . . . . .

For each graph pattern exists a CSV file that lists the IDs of all graphs that are matched by that pattern.For example, the file target-p-t3-t4-v1=v1-t1-t2-graph_size_2.csv lists all those graphs that are matched by the pattern t3-t4-v1=v1-t1-t2.

Finally, there are four JSON files:

plain-graphs_of_size_1.json, plain-graphs_of_size_2.json, enriched-graphs_of_size_1.json, enriched-graphs_of_size_2.json 

These files contain the same graphs that are also stored in the .nt-files, but allow to conveniently read in a set of graphs at once.