Convert event-model to use the dls python skeleton

[evalott100]

As per a discussion with @coretl, we'd like to look into using the dls skeleton with bluesky repos. NSLS are amenable to this, but we'll probably end up cherry picking desired features.

We should first convert event-model.

Expected Behavior

• All the features in the current dev environment should be functional.
• There are features we'll want to exclude from the pip skeleton e.g .vscode.

[danielballan]

I am open to this. It might make sense to use the dls skeleton as a starting point and build a Bluesky skeleton so that Diamond and bluesky do not have to align on skeleton decisions 100% of the time.

We have used https://github.com/NSLS-II/scientific-python-cookiecutter in the past. That has become dated, but I suspect that some of it is worth keeping.

[coretl]

Shall we start with adopting dls skeleton, then we can fork from it and make the changes there when we decide how these are different?

[danielballan]

Sounds good.

[coretl]

Please do the black + isort in a single separate commit like the ophyd adoption of black

[evalott100]

The skeleton setup is on this branch. The main features to pick and choose from are

• using a single pyproject.toml instead of requirements.txt and requirements-dev.txt
• black
• isort
• having the git version generated by setup.py instead of versioneer
• mypy (the CI of which currently fails but will pass once we sort out the type hinting in __init__.py)
• documentation building is largely the same, though I will need to adapt sphinx to not be picky when trying to retrieve docstrings
• .vscode (which I deleted because I assumed you definitely wouldn't want, but can always bring back)
• pre-commit