LanguageMachines/frog

Frog output on pretokenised FoLiA input do not make it into the FoLiA output

proycon opened this issue · 2 comments

Something is wrong still when processing pre-tokenised FoLiA, the annotations are not making it into the FoLiA output (but they DO make it to the stdout columned output!)

Input: https://lst.science.ru.nl/~proycon/issue72_a.xml

Command: $ frog --skip=tmncpa --language=nld issue72_a.xml -X test.xml

The same thing also occurs when not skipping the tokeniser explicitly, and when not adding --language=nld

hmm, interesting....
so the work is done and than forgotten. Will look into it

Ok, so the problem was, that when skipping the tokenizer, the text was not assigned a language at all. Even not "default". And because of the mismatch, a part of the processing was skipped.
This should be fixed now. Please test