environment.yml ERROR

[MoonlightFansty]

File configuration error """ dependencies: ... h5py=2.9.0l.c ... """ should be modify. I try to change it: """ dependencies: ... h5py ... """ the code is working. However, I am not sure whether it will affect the subsequent analysis. We look forward to your reply or solution !!

Thank you for your time

[tayebeazimi]

Hi, I have python 3.8 and I can't install h5py 2.9 and as a result I have problem with environment.yml. I will appreciate if you let me know how to solve this problem.

[MoonlightFansty]

Hi, I have python 3.8 and I can't install h5py 2.9 and as a result I have problem with environment.yml. I will appreciate if you let me know how to solve this problem.

Maybe you can try another version. My configuration does not add a version number.

[jasonleongbio]

Hi guys, I also encountered the configuration error with the h5py==2.9.0l.c package.

May I know if anybody has any update on this - whether modifying this to h5py (without the package version) still works fine with the subsequent analysis?

Thanks a lot

[akshdeepsandhu]

Hi all,

I did some digging as I was running into the same issue. Overall it seems that the environment.yml can be a bit unstable across platforms and perhaps in the future the authors can release a docker image for cross-platform compatibility. I have managed to get things running on both a MacOS and Linux system.

The steps outlined are mostly following along from this thread posted on the conda github that led to a nice clean environment.

1. Try modifying the conda config conda config --set restore_free_channel true
2. Updated the environment.yml file adding a few more channels:

channels:
  - plotly
  - defaults
  - pytorch
  - conda-forge

3. Changing the h5py to a stable version: - h5py==2.9.0

After some reading, the h5py package is used to read HDF5 binary data. The version listed in step 3 should work with the python version required here.

@tayebeazimi The reason you are unable to install the package might be because of the version conflict you identified. Since this environment setup requires an older python version you should try specifying this when creating your conda environment as such: conda env create --name pnet_env python=2.7 --file=environment.yml

Hopefully this was helpful - let me know if you still have trouble.

Below is the full enviroment.yml that I used for a clean setup:

name: pnet
channels:
  - plotly
  - defaults
  - pytorch
  - conda-forge
dependencies:
  - pip=18.1
  - enum34=1.1.6
  - futures=3.3.0
  - hdf5=1.10.4
  - imageio=2.6.1
  - keras=2.2.4
  - keras-applications=1.0.8
  - keras-base=2.2.4
  - keras-preprocessing=1.1.0
  - markdown=3.1.1
  - numpy=1.16.6
  - pandas=0.23.4
  - plotly=4.5.4
  - protobuf=3.11.2
  - python=2.7.15
  - pyyaml=5.1.1
  - scikit-image=0.14.2
  - scikit-learn=0.20.1
  - scipy=1.1.0
  - subprocess32=3.5.4
  - tensorboard=1.12.2
  - tensorflow=1.12.0
  - yaml=0.1.7
  - plotly-orca=1.3.1
  - pip:
      - adjusttext==0.7.3
      - lifelines==0.19.5
      - matplotlib==2.2.4
      - psutil==5.8.0
      - networkx==2.2
      - pyvis==0.1.7.0
      - requests==2.23.0
      - rope==0.18.0
      - seaborn==0.9.1
      - toml==0.10.2
      - upsetplot==0.4.0
      - urllib3==1.25.8
      - xlrd==0.9.0
      - h5py==2.9.0

[siddhartherat1997]

Hi I am trying to set this up on a windows machine, but I haven"t been successful so far. Given the fact that the older versions of tensorflow is not available on the website itself. If anyone knows a workaround , it would be really helpful. Thanks