bernt-matthias
commented
3 years ago Was reading a bit more, the linked repo already allows for custom extension of the docker files:
Open bernt-matthias opened 3 years ago
bernt-matthias
commented
3 years ago Was reading a bit more, the linked repo already allows for custom extension of the docker files:
The tensorflow image (which is the base image for our container) https://github.com/tensorflow/tensorflow/blob/master/tensorflow/tools/dockerfiles/dockerfiles/gpu.Dockerfile already contains python 3 https://github.com/tensorflow/tensorflow/blob/e0c681c5c738d8898fd698fc9d27e31c1471b993/tensorflow/tools/dockerfiles/dockerfiles/gpu.Dockerfile#L78 and tensorflow https://github.com/tensorflow/tensorflow/blob/e0c681c5c738d8898fd698fc9d27e31c1471b993/tensorflow/tools/dockerfiles/dockerfiles/gpu.Dockerfile#L98
In our docker file we install python3 https://github.com/yigbt/deepFPlearn/blob/e8f60c4251a4e2252ec094be948742f5523b4224/singularity_container/environment.yml#L7 and tensorflow https://github.com/yigbt/deepFPlearn/blob/e8f60c4251a4e2252ec094be948742f5523b4224/singularity_container/environment.yml#L12 including all their dependencies.
We should avoid this.
I guess easiest would be use pip instead of conda to get the additional dependencies. For instance by defining the requirements in https://github.com/yigbt/deepFPlearn/blob/master/setup.py and just pip install deepFPlearn.