Codon Likelihihoods Learned Against Genome Evolution (CoLLAGE): a deep learning framework for identifying naturally selected patterns of codon preference within a species
Move the package down a level. The python package will no longer the root of the repository, it will be the collage folder. This means that all imports can just be collage.something instead of collage.src.something. It also means that scripts in the root of the repository (e,g, the training script) will run even if collage is not in the PYTHONPATH. Further, it sets the stage for future Dockerization in a nicer way.
Future improvements will involve moving the logic from the training script into the package and just having a thin wrapper main at the top level of the repository.
Notebooks that were previously in other directories may need to have their imports adjusted after this change.
Move the package down a level. The python package will no longer the root of the repository, it will be the
collagefolder. This means that all imports can just becollage.somethinginstead ofcollage.src.something. It also means that scripts in the root of the repository (e,g, the training script) will run even ifcollageis not in thePYTHONPATH. Further, it sets the stage for future Dockerization in a nicer way.Future improvements will involve moving the logic from the training script into the package and just having a thin wrapper main at the top level of the repository.
Notebooks that were previously in other directories may need to have their imports adjusted after this change.