search

CogComp / cogcomp-nlpy

CogComp's light-weight Python NLP annotators
http://nlp.cogcomp.org/
Other
116 stars 26 forks source link

Add a minimum model version #111

Open danieldeutsch opened 6 years ago

danieldeutsch commented 6 years ago

The local pipeline hello world code was failing with the error

AttributeError: type object 'edu.illinois.cs.cogcomp.annotation.BasicTextAnnota' has no attribute 'createTextAnnotationFromListofListofTokens'

It turns out it was because the models I had in ~/.ccg_nlpy were outdated and the createTextAnnotationFromListofListofTokens method did not exist in the jar files that I had. After I redownloaded the models, the code ran successfully.

We should add a minimum required model version and check at runtime whether or not the available models are valid. If not, throw an error with instructions to download the new models.

This might have solved #106

danyaljj commented 6 years ago

Yeah, that's a good point.

If no one is doing this, I can give it a try.

nitishgupta commented 6 years ago

@danyaljj: We can do this during the object creation for the LocalPipeline class. Ideally, there should be a static field in the download.py file that has the least-supported version that LocalPipeline checks from. Also, if you're looking at this, could you add a version-number flag to download.py with possibly a least-supported flag.