The fact that the pipeline is written in bash, forces the user to run it on a Linux environment (i.e. this leaves out Windows users). Since it is mostly calling other scripts, creating files/folders, we can consider two options:
Leave the pipeline written in bash. Prepare a docker container to encapsulate everything (it would not be a python package in this case)
Write the pipeline in pure Python. Deploy the pipeline just as a Python package that can be ran in any system just by installing with pip, as long as the necessary dependencies are met
The fact that the pipeline is written in bash, forces the user to run it on a Linux environment (i.e. this leaves out Windows users). Since it is mostly calling other scripts, creating files/folders, we can consider two options: