Dependency issue (numpy & sklearn)
Opened this issue · 1 comments
Dear all,
I installed ChemDataExtractor2 with plain "pip install chemdataextractor2" (pip version 23.0.1). The numpy version automatically installed was 1.24.2 and the sklearn 0.22.1. However, in sklearn/linear_model/_least_angle.py -file some older version numpy style is used, for example. np.float and np.int, which in version 1.24.2 numpy are just float and int.
I was able to resolve this just by manually changing all np.int and np.float to plain int and float in the _least_angle.py-file, but just thought to let you know that there exists this kind of an issue.
I confirm that after i had to download hundreds versions of Flask, boto3 and botocore, i got this error:
module 'numpy' has no attribute 'int'.
np.int was a deprecated alias for the builtin int.
I had to edit /usr/local/lib/python3.8/site-packages/sklearn/feature_extraction/image.py to make the "cde data where" command to work.
This would not happened with a docker image. We could just build it and run it.