This is trickier than it seems, since the tool doesn’t require that we have token spans. There is no basis to directly export to BIO. Integrating a tokenizer also is language-dependent. Have to think about that.
In the meantime: export to UIMA/DKpro, and run one of its tokenizers and the BIO export from there.
This is trickier than it seems, since the tool doesn’t require that we have token spans. There is no basis to directly export to BIO. Integrating a tokenizer also is language-dependent. Have to think about that.
In the meantime: export to UIMA/DKpro, and run one of its tokenizers and the BIO export from there.