Preparations for packaging / distribution

[drammock]

TODOs (probably incomplete)

• add pyproject.toml
• add release workflow
• create deployment environment (workflow expects name pypi)
• set required reviewers for environment?
• create the project on PyPI under MNE org should happen automatically on release
• create PyPI Trusted Publisher token and add to deploy environment
• disable the curryreader workflow? (it's run upstream already)

cc @larsoner @dominikwelke

[dominikwelke]

thanks for pushing this forward @drammock . i had also worked a bit on packaging today, but built manually and with moving files to a src folder so far (see https://github.com/dominikwelke/curry-python-reader/tree/build)

i put my version up on TestPyPI and everything works.
(pip install -i https://test.pypi.org/simple/ curry-python-reader)
but an automatic workflow without touching other files is better of course!

the only relevant file for us is curryreader.py, btw, if you want to exclude more.
also there's only one relevant def in there (curryreader.read)

[drammock]

built manually and with moving files to a src folder so far [...]
but an automatic workflow without touching other files is better of course!

Yeah I was trying to avoid moving/deleting/changing existing files, to simplify the process of merging in upstream changes from compumedics/neuroscan. If you want to play around locally with the changes here, python -m build . in the root directory works, assuming you have relevant packages installed.

[dominikwelke]

hi @drammock -
can i help in any way to push this forward?

alternatively, i could temporarily publish the version i already packaged on PyPI proper to allow testing of the MNE code.
we can take it down again once this version here is up and running..

[drammock]

can i help in any way to push this forward?

the nudge was good enough. Stand by...

[drammock]

here you go! https://pypi.org/project/curryreader/

[dominikwelke]

great! thank you @drammock