Codecov Report

Merging #367 into master will increase coverage by 0.11%.
The diff coverage is 87.5%.

@@            Coverage Diff             @@
##           master     #367      +/-   ##
==========================================
+ Coverage   85.32%   85.44%   +0.11%     
==========================================
  Files          34       34              
  Lines        1881     1951      +70     
  Branches      337      344       +7     
==========================================
+ Hits         1605     1667      +62     
- Misses        237      243       +6     
- Partials       39       41       +2

This looks really good so far. Once the pipeline is agreed upon, I would in this batch also introduce other languages. I really like the fact that resources are downloaded upon the selection of a language, which means if I don't ever use, say Coptic, it won't be downloaded locally.

Please check slo-opinion-lexicon document 7 with UDPipe tokenization. I don't understand why this tokenization considers fullstops at the end as a part of the token. Npr. "narediti.". In document 4, for example, it doesn't do that. Could it be that sentence parsing isn't working well?

Before merging it would be nice to:

• add tests (I am sure this was already planned)
• update documentation! (I will do this part)
• check saving the report (tokenizer is not overridden when UDPipe is selected)

@PrimozGodec This works well for me. Would you mind giving a look at the code and if everything looks fine, merge?