parasj/checkmate

Library versions of Python Dependencies

Opened this issue · 1 comments

I am trying to run the test cases, but ran into errors like:

pandas.util.version.InvalidVersion: Invalid version: '0.14.0.RAY'

Reinstalling an older version of pandas probably solved this error but lead to another error:

pyarrow.lib.ArrowTypeError: Did not pass numpy.dtype object

I am wondering if it is possible to provide the versions of the Python libraries in the setup.py script to solve these dependency issues:

checkmate/setup.py

Lines 9 to 25 in 1889c57

install_requires=[
"numpy",
"pandas",
"redis",
"matplotlib",
"seaborn",
"tqdm",
"ray>=0.7.5",
"graphviz",
"python-dotenv",
"pytest",
"pillow",
"gym",
"psutil",
"pydot",
"keras_segmentation @ https://github.com/ajayjain/image-segmentation-keras/archive/master.zip#egg=keras_segmentation-0.2.0remat"
]

After playing around with different library versions, these combinations seem to work:

  • numpy==1.19.5
  • pandas==0.25.2
  • ray==0.7.5
  • tensorflow==2.1.0
  • tensorflow-cpu==2.1.0

The full log of my pip freeze command is here.