BUG: Get eval results from metadata

[flying-sheep]

Fixes #596

No idea how to elegantly pass the SyntaxNode all the way down to the role, and I’m not super interested in figuring that out, it’s your call how to implement that.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 89.47368% with 2 lines in your changes missing coverage. Please review.

Project coverage is 80.98%. Comparing base (af2d197) to head (6d8447d).
Report is 14 commits behind head on master.

❗ Current head 6d8447d differs from pull request most recent head 498cad5

Please upload reports for the commit 498cad5 to get more accurate results.

Files	Patch %	Lines
myst_nb/core/execute/base.py	87.50%	1 Missing ⚠️
myst_nb/core/execute/inline.py	75.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #597      +/-   ##
==========================================
+ Coverage   80.97%   80.98%   +0.01%     
==========================================
  Files          30       30              
  Lines        2702     2714      +12     
==========================================
+ Hits         2188     2198      +10     
- Misses        514      516       +2     

Flag	Coverage Δ
pytests	80.98% <89.47%> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[flying-sheep]

@agoose77 is there a way this can make it into this week’s release?

[flying-sheep]

Hi @agoose77, would be great if you’d take a look!

The current doc build failures are not mine, but #610

[agoose77]

I will hold this for now; we need to rectify the different ways that myst-nb and jupyterlab-myst perform caching of output expressions.

[flying-sheep]

This just uses the user expression result that’s already stored in the metadata. It doesn’t touch how it’s being stored.

That doesn’t seem to have a potential for problems, or am I missing something?

[agoose77]

This just uses the user expression result that’s already stored in the metadata. It doesn’t touch how it’s being stored.

That doesn’t seem to have a potential for problems, or am I missing something?

My initial concerns were around the way that execution should be handled. But your choice of wording has helped me to disambiguate the two cases; reading from a notebook when execution is disabled is something that we can (and should) to. I'll revisit this PR shortly (probably next week).

[flying-sheep]

Hi, it’s been a month! Would be great if you could take a look here

[agoose77]

Thanks @flying-sheep for your continued patience. I want to add support for inline-expressions in jupyter-cache as well, if we're going to get this one in. That way, all of our execution modes will support expressions. I'll circle back once the myst-nb release / JB cycle is done

[flying-sheep]

Thanks for the update. I’d say merging this as-is would probably already add value, as it’s a tiny change with big effects.

Do you know what adding support for jupyter-cache would entail?

[flying-sheep]

ping