This repository contains some changes/fixes to FinnWordNet.

The data directory contains the FiWN data files, and the WNgrind-3.0-FiWN directory contains the FiWN version of WNgrind.

There is a script which can either create a false/en based synset id => true fi synset id mapping tsv, or apply the mapping to the tsvs in data. It needs pipenv.

Assuming you put the original data in data rather than the already mapped data included here, you can make a map tsv like so:

$ pipenv run python adjust-fiwn-offsets.py dump data synset_map.tsv

And you can also modify the original data with the new offsets (i.e. the following is the command which has been run to change the data in data to its current state):

$ pipenv run python adjust-fiwn-offsets.py fix data

You can create count data based on the counts in the English data like so:

$ pipenv run python mk-cntlist.py > data/dict/cntlist.rev