Closed migueltorrescosta closed 7 months ago
Dear migueltorrescosta,
thank you for the note. Indeed it would be great to sharpen the installation in case there were some missing libraries. I'll think about what is the best way of implementing this. In the meantime, it can be great that you make a clone of the repository on your side so that interested researchers can use your version.
Best regards, Jose
Thank you for the original code @joselado :raised_hands: For anyone looking for this functionality, I have a working clone on https://github.com/migueltorrescosta/dmrgpy/tree/add_pyproject, and it can be installed via
pip install git+https://github.com/migueltorrescosta/dmrgpy.git@add_pyproject
Updates from @joselado do not automatically propagate to my fork of the code.
Problem
We are not able to install this remote github repo with a single command. This consumes human time for researchers who could be running a single command, rather than clonning the repo, setup their own environment and running custom shell scripts to get started.
Solution
Add a
pyproject.toml
file with the required dependencies:numpy scipy numba statsmodels statistics matplotlib
Testing
I have ran the command with the created fork:
:warning: This install procedure assumes that the C++ compiler is already setup. This can be added as a
build.py
file, as described in this StackOverflow answer.