Textract plugin fails to install

[timalamenciak]

Textract 1.6.5 has some problems in its build file which causes the install to fail - documented here: deanmalmgren/textract#476

The only workaround that I have found to work has been installing textract-py3 - a minimally maintained version without this build vulnerability: https://github.com/KyleKing/textract-py3

However, when I try to point OntoGPT at a PDF, it hangs. Presumably it's not trying to use textract-py3 and is looking for vanilla textract.

[caufieldjh]

Strange - I'm unable to reproduce this issue.
Installing with pip (as pip install ontogpt[textract]) or with poetry (as poetry install --extras textract) both appear to work without issue. Of course, in both cases it's installing textract 1.5.0, which is pretty old, on Python 3.10.
I'll enable using textract-py3 as an optional dependency.

[caufieldjh]

OK, please let me know if PDF processing works after removing textract and installing the textract-py3 extra (with the repo version - haven't included this in a release yet)

[timalamenciak]

Installing the repo version seemed to work off the bat - it installs textract 1.5.0, which may not have this problem.

Of course now textract throws an error with the PDFs because of some undefined characters but I think that is beyond the scope of this issue request. Thanks for the quick response.