This software matches WordNet and ConceptNet 5.3 and loads them into neo4j 2.2.2. This is a sister project of Neo_Concept and Neo_Wordnet.
If the "match" option is used, Neo_Merger matches ConceptNet and Wordnet nodes with a simple algorithm. Matched nodes are connected with relationships of type 'KBmatch'. A 'KB' source property is added to each relationship to make it easier to calculate paths that stay within specific KBs. It has values in 'C', 'W', and 'M' for ConceptNet, Wordnet, and matched relationships, respectively.
- neo4j version >= 2.2.2 (needed for the new import tool)
- Neo_Concept
- Neo_Wordnet
- regular Python, no dependencies
Tested with neo4j-community-2.2.2 and 3.0.0-M02.
mkdir neo-kbs
cd neo-kbs
git clone https://github.com/redsk/neo_merger.git
# Follow Neo_Concept how-to but do not import the two CSV files
# Follow Neo_Wordnet how-to but do not import the two CSV files
mkdir merged-kbs
# To match the two KBs use:
python neo_merger/neo_merger.py wordnet/WNnodes.csv wordnet/WNedges.csv conceptnet/nodes.csv conceptnet/edgesPOS.csv merged-kbs/nodes.csv merged-kbs/edges.csv match
# To avoid the (probably to simplistic) match of the two KBs:
python neo_merger/neo_merger.py wordnet/WNnodes.csv wordnet/WNedges.csv conceptnet/nodes.csv conceptnet/edgesPOS.csv merged-kbs/nodes.csv merged-kbs/edges.csv nomatch
# import the merged nodes.csv and edges.csv using the new import tool
neo4j-community-2.2.2/bin/neo4j-import --into neo4j-community-2.2.2/data/graph.db --nodes merged-kbs/nodes.csv --relationships merged-kbs/edges.csv --delimiter "TAB"
# start neo4j
neo4j-community-2.2.2/bin/neo4j start
Go to localhost:7474 to see the graph. Create indexes on labels for performance reasons:
CREATE INDEX ON :`lemon#LexicalEntry`(id)
CREATE INDEX ON :`lemon#LexicalSense`(id)
CREATE INDEX ON :`wdo#Synset`(id)
CREATE INDEX ON :Concept(id)
You can now query the database. Example:
MATCH (c0:`lemon#LexicalEntry` { id:"wn/bus+driver-n"}),(c1:`lemon#LexicalEntry` { id:"wn/drive-v"}),path=allShortestPaths((c0)-[*..10]-(c1))
RETURN c0,c1,path