search

danieldk / conllx-utils

CoNLL-X utilities
7 stars 1 forks source link

conllx-partition: CoNLL-U format not supported #1

Open timmli opened 6 years ago

timmli commented 6 years ago

When I try to partition a CoNLL-U file like this 

../conllx-utils/target/debug/conllx-partition 15 PREFIX .conllu INPUT.conllu

the following error is thrown:

Error: cannot parse as integer field: '# sent_id = train-s1501'

The same with a CoNLL-X-formatted file works fine.

danieldk commented 5 years ago

Supporting CoNLL-U would require a good crate for CoNLL-U and a good abstraction of both the conllx crate and this conllu crate. This has been on my todo list for a long time, but there are always more urgent issues ;). (Such as converting conllx-utils to use the new petgraph-based conllx API.)