NF: register_config()

[mih]

This PR requires and sits on top of #6594. The first commit is not unique to this PR.

This also replaces the previously very flexible data structure for such items with a more rigid set of classes that implement lazy evaluation of defaults (avoiding their possibly expensive computation, even when they are not actually needed). They otherwise implement all elements of the previously used API, such that they work as a drop-in replacement.

Only register_config() (and its friend has_config() is intended to be used publicly. Like so:

from datalad.support.extensions import register_config
register_config(
    'datalad.mihextras.blows',
    title="Does it?",
    default=True,
    scope='global',
    type=bool,
    description="I could tell you, but I won't!",
)

Changelog

💫 Enhancements and new features

• A new module datalad.support.extensions offers the utility functions register_config() and has_config() that allow extension developers to announce additional configuration items to the central configuration management. Fixes formalize extending common_cfg.definitions (and other things) in extensions? #6314

[yarikoptic]

+1 on register_config since in-line with what I was looking after in #6314 . I would only also "expose" it within some centralized datalad.support.extensions module.

This PR requires and sits on top of #6594.

but in principle this functionality is not needing #6594, right? i.e. extension can only define config variable it uses, thus would need to be imported by then one way (as via datalad.api or CLI) or another (the #6594), most likely using that register_config within that extension's __init__.py, correct?

[bpoldrack]

This PR is currently in draft mode to gather feedback on the API, before tests are added.

Feedback: I like it.

[mih]

Can be seen in action here: https://github.com/datalad/datalad-next/blob/master/datalad_next/__init__.py (running across platforms)

[mih]

This PR requires and sits on top of #6594.

but in principle this functionality is not needing #6594, right? i.e. extension can only define config variable it uses, thus would need to be imported by then one way (as via datalad.api or CLI) or another (the #6594), most likely using that register_config within that extension's __init__.py, correct?

#6594 is already needed to give the configuration command the chance to see such definitions.

[mslw]

I read the docstring to figure out the proposed mechanism. Looks clear to me.

[codeclimate]

Code Climate has analyzed commit dcb749e and detected 2 issues on this pull request.

Here's the issue category breakdown:

Category	Count
Complexity	2

View more on Code Climate.

[mih]

I took this out of draft mode. Test coverage is now 100%, and missing functionality for the doc builder was added. I also added has_config() to mitigate the concern expressed in #6601 (comment)

[bpoldrack]

Restarted CI, since unrelated failures should be gone (except for snapshot build)

[codecov]

Codecov Report

Merging #6601 (dcb749e) into master (8bf53dd) will increase coverage by 0.17%.
The diff coverage is 91.26%.

@@            Coverage Diff             @@
##           master    #6601      +/-   ##
==========================================
+ Coverage   91.13%   91.31%   +0.17%     
==========================================
  Files         350      353       +3     
  Lines       44232    45580    +1348     
==========================================
+ Hits        40312    41621    +1309     
- Misses       3920     3959      +39     

Impacted Files	Coverage Δ
datalad/__init__.py	85.63% <ø> (-0.33%)	⬇️
datalad/support/entrypoints.py	47.36% <26.66%> (-13.51%)	⬇️
datalad/cli/main.py	76.00% <100.00%> (+1.26%)	⬆️
datalad/coreapi.py	100.00% <100.00%> (ø)
datalad/interface/common_cfg.py	100.00% <100.00%> (ø)
datalad/support/extensions.py	100.00% <100.00%> (ø)
datalad/support/tests/test_extensions.py	100.00% <100.00%> (ø)
datalad/core/local/tests/test_status.py	97.97% <0.00%> (-0.51%)	⬇️
datalad/support/annex_utils.py	100.00% <0.00%> (ø)
datalad/core/local/tests/test_save.py	98.12% <0.00%> (+<0.01%)	⬆️
... and 8 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 8bf53dd...dcb749e. Read the comment docs.

[bpoldrack]

Everyone in agreement it seems. Merging.