search

oeg-upm / ner4soft

Repository for expriments and corpora for NER for code repositories and readme files
2 stars 2 forks source link

Rules for tagging package version and qualifier #8

Open dgarijo opened 3 years ago

dgarijo commented 3 years ago

Right now something like: Python 3.6 will produce Python as a language and 3 as a package and 6 as a package. What I would like is to recognize "Python 3.6", and have as a sub-annotation the version.

Also, if we have something like: tensorflow==1.5.0 we should produce something like:

{"text":"tensorflow",
"version":"1.5.0",
"qualifier":"equal",
"type":"PythonPackage",
"init":46,
"end":48}

So I can transform it to a KG later.

dgarijo commented 3 years ago

In addition, some POS tagging may be needed. If you have: You need Python 3.6 or higher, we should recognize this!

dgarijo commented 3 years ago

@SerPablo can you share here the doc on adding this as a Jape rule? I will be happy to incorporate it.