❇️ Read configuration for execution from myst.yml

[agoose77]

Fixes #1214

[stevejpurves]

I am wondering if it is the best thing for the myst.yml#thebe config to be used as the first source of the execute configuration?

My thinking is that if I have a manuscript or book ready for publication, or perhap I already have v0.0.666 published and I am making a change to a notebook to add or fix something. Then my myst.yml is set up for production... I do want to re-execute locally before I publish v0.0.667 but I don't want to have to do that via say a public binder (slow) and I just want to hit my local server. I don't really want to have to change my myst.yml as that is all set up for prod, and maybe i'll forget to change it back...

So TLDR, should JUPYTER_BASE_URL take precedence over configServer allowing me to point to a local config, without messing with production myst.yml?

that seems to make sense to me, but maybe I am missing a counter use case that you have in mind @agoose77

[agoose77]

So TLDR, should JUPYTER_BASE_URL take precedence over configServer allowing me to point to a local config, without messing with production myst.yml?

Yes, and that was my original intention. I must've forgotten to invert the order!

The more I think on it, there are a lot of ways we could write this. Perhaps we hold this PR for now, and --execute becomes a multi-modal flag e.g.

• --execute (try jupyter, allow env)
• --execute=config (use jupyter.server)
• --execute=http://(...):8888?token=(...) (use URL)
• --execute=env (use env)

I'm not sure on the best solution, so I'll close this for now