⬆️ UPGRADE: myst-parser to 0.15.2

[mmcky]

This PR upgrades myst-parser to version 0.15.2.

This includes the following improvements and changes

Tasks

• Wait for myst-parser 0.15.2 release
• Update dependency here
• Merge!

[chrisjsewell]

Thanks @mmcky! but I think you are missing the crucial change of actually updating setup.cfg lol

[chrisjsewell]

ah yeh thats the one

[mmcky]

Thanks @mmcky! but I think you are missing the crucial change of actually updating setup.cfg lol

Yeah thanks -- I used git checkout . when working with the test set and it reset the change :-(. All fixed up.

[codecov]

Codecov Report

Merging #353 (be185b2) into master (b132a50) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #353   +/-   ##
=======================================
  Coverage   87.43%   87.43%           
=======================================
  Files          12       12           
  Lines        1337     1337           
=======================================
  Hits         1169     1169           
  Misses        168      168           

Flag	Coverage Δ
pytests	87.43% <ø> (ø)

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

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b132a50...be185b2. Read the comment docs.

[mmcky]

• update readthedocs and fix this warning

WARNING: The config value `myst_url_schemes' has type `tuple', defaults to `list'.

[choldgraf]

ah excellent! @mmcky is this ready for review or do you wanna get tests passing first?

[mmcky]

thanks @choldgraf I think I just fixed the readthedocs issue. So I'll tag this for review once everything is green.

[mmcky]

Thanks @choldgraf and @chrisjsewell I think this is ready to 🚢

[chrisjsewell]

WARNING: The config value myst_url_schemes' has type tuple', defaults to `list'.

Hmm, maybe this should be "fixed" in myst-parser, as really there should be no problem with a tuple (you can specify a list of allowed types https://www.sphinx-doc.org/en/master/extdev/appapi.html#sphinx.application.Sphinx.add_config_value)

[chrisjsewell]

Thanks @choldgraf and @chrisjsewell I think this is ready to 🚢

Maybe just wait for me to release v0.15.2 and pin to that, since it is anyway needed for sphinx 4: #352 (comment). Will try to do in next day or 2

[choldgraf]

sounds good - I have added the next steps in this PR...we should get it merged quickly once 0.15.2 is out so that we can bump up the dependencies in jupyter book and resolve some of these pipfile issues 😬

[mmcky]

WARNING: The config value myst_url_schemes' has type tuple', defaults to `list'.

Hmm, maybe this should be "fixed" in myst-parser, as really there should be no problem with a tuple (you can specify a list of allowed types https://www.sphinx-doc.org/en/master/extdev/appapi.html#sphinx.application.Sphinx.add_config_value)

@chrisjsewell yeah I thought that might be true but it was actually specified in docs/conf.py of this repo so this is fixed now.

Ah I just re-read your comment -- I see. So a tuple should be OK.

[choldgraf]

Note: I've bumped to 0.15.2.

The changes all look good to me - seem pretty straightforward. I had one question about person-specific paths in the regression files, but I don't think that needs to block this since those will be manually updated in future PRs anyway

[choldgraf]

is there any way we can configure tests to not be person-specific? otherwise I think this will be re-written any time somebody has to update the tests who didn't update them in the past.

[mmcky]

@choldgraf I just updated a whole bunch of metadata in the tests in upgrade to sphinx 4 and the tests actually ignore the metadata

MyST-NB/tests/conftest.py

Line 23 in b132a50

set_notebook_diff_targets(metadata=False)

I don't see an easy way to generate these fixtures with generic values.

[chrisjsewell]

as mentioned, the actual tests ignore these values when diffing.
if you wanted to reduce the diff though, it would be possible to add a pre-commit hook, which replaces all the "non-deterministic" content with deterministic ones

[mmcky]

I think this is good to go -- do you agree @chrisjsewell?

Once this is merged I will start on sphinx >= 4 work and will rebase on this starting point.

[chrisjsewell]

cheers!

[mmcky]

thanks @chrisjsewell 👍 once all tests reporting green -- I'll merge.