If I understood the submission process the right way - the test run will pull the requirements from the requirements.txt
My whole is code is based on rdkit, i.e. in order to predict a sentence from SMILES functions from rdkit are needed. But rdkit doesnt have pip support and adding the library to req.txt doesnt make sense.
Is there any work-around for this?
Other ideas?
Yes, conda environment has higher priority than requirements.txt, so you would need to use either of them. But in case you are willing to add PyPI packages, you can simply list them in your environment.yml directly.