WNcsv-3.1 is a CSV-formatted version of WordNet 3.1, derived from WNprolog-3.1 by Eric Kafe (https://github.com/ekaf/wordnet-prolog).
The Prolog databases were generated from the original WordNet 3.1 databases, and bundled with a copy of the WNprolog-3.0 documentation (c) 2012 Princeton University.
The present release contains the following numbers of unique database records:
- wn_ant.csv: 7988
- wn_at.csv: 1278
- wn_cls.csv: 9559
- wn_cs.csv: 221
- wn_der.csv: 74781
- wn_ent.csv: 408
- wn_exc.csv: 6053
- wn_fr.csv: 21684
- wn_g.csv: 117791
- wn_hyp.csv: 89172
- wn_ins.csv: 8589
- wn_mm.csv: 12288
- wn_mp.csv: 9111
- wn_ms.csv: 797
- wn_per.csv: 8074
- wn_ppl.csv: 73
- wn_sa.csv: 4054
- wn_sim.csv: 21434
- wn_sk.csv: 207272
- wn_s.csv: 207272
- wn_syntax.csv: 1054
- wn_vgp.csv: 1744
- total: 810697
This repository also includes alternative branches, with CSV versions of Princeton WordNet 3.0 and 3.1, or Open English Wordnet Editions 2022 and 2025+.
For convenient inter-operation with other projects, the included csv2tab script converts the CSV databases to tab-separated files (TSV), which can be easily imported into many database systems.
- "make tab" produces a ".tab" file for every ".csv" file in the "csv" directory.
- "make clean" deletes the ".tab" files.
- fix double-quotes in CSV strings.
- Separate db records with CRLF, as required by RFC 4180.
- Output ".tab" files to a separate "tab" directory.