Make package available for pip install
https://betterscientificsoftware.github.io/python-for-hpc/tutorials/python-pypi-packaging/
This is a goal down the line. Milestones before then:
- Implement more IoU scores
- Go through more heavy documentation (read_me and docstrings)
- Enable user optimization of moment-to-moment labelling algorithm.
- Make compatible with Windows and MacOS.
Will also have to figure out how to incorporate Microfaune in a pip install.
Python package publishing tutorial: https://www.youtube.com/watch?v=GIF3LaRqgXo
We can also configure it for pip install on github instead of on PyPi. Or do that first since it has a lower barrier of entry.
Microfaune can be installed through pip using
-e "PyHa/microfaune_package"
The install from GitHub has been solved with the use of poetry.
Remaining is whether you want bdist builds (wheels) from GitHub, or if you just want to deploy to PyPi