My take on how to package DENSS using pyproject.toml
There was some refactoring needed: All executable python files were moved to a package denss and their code moved to a function main(). This allowed creation of proper entrypoints. Resulting CLI commands are missing the *.py suffix.
Scipy and Numpy now installs automatically. Matplotlib was kept as "optional", however, personally I would put in the full requirements.
A citation file was added.
In this stage, superdenss and best_enantiomers.sh do not install. Dirty fix would be wrapping them using subprocess for example, ideally they should be rewritten to Python as then it would make DENSS platform agnostic 100% (it mostly runs on Windows already).
My take on how to package DENSS using pyproject.toml
There was some refactoring needed: All executable python files were moved to a package
denssand their code moved to a functionmain(). This allowed creation of proper entrypoints. Resulting CLI commands are missing the*.pysuffix.ScipyandNumpynow installs automatically.Matplotlibwas kept as "optional", however, personally I would put in the full requirements.A citation file was added.
In this stage,
superdenssandbest_enantiomers.shdo not install. Dirty fix would be wrapping them usingsubprocessfor example, ideally they should be rewritten to Python as then it would make DENSS platform agnostic 100% (it mostly runs on Windows already).