FIX: Make GiftiMetaData.data a list proxy, deprecate

[effigies]

This is a follow-up to #1091 with things that should have been done there, including proxying actions on the GiftiMetaData.data internal list and deprecating the property altogether.

I added a comprehensive test that checks that attempting to use GiftiMetaData.data as one might have pre-4.0 works once again.

[pep8speaks]

Hello @effigies, Thank you for updating!

Cheers! There are no style issues detected in this Pull Request. 🍻 To test for issues locally, pip install flake8 and then run flake8 nibabel.

Comment last updated at 2022-07-26 13:27:38 UTC

[effigies]

@htwangtw Could I bug you for a review here?

[codecov]

Codecov Report

Merging #1127 (301e024) into maint/4.0.x (c90b75d) will increase coverage by 0.01%.
The diff coverage is 96.49%.

@@               Coverage Diff               @@
##           maint/4.0.x    #1127      +/-   ##
===============================================
+ Coverage        91.76%   91.77%   +0.01%     
===============================================
  Files              101      101              
  Lines            12380    12433      +53     
  Branches          2423     2429       +6     
===============================================
+ Hits             11360    11411      +51     
- Misses             693      694       +1     
- Partials           327      328       +1     

Impacted Files	Coverage Δ
nibabel/gifti/gifti.py	87.78% <96.49%> (+1.31%)	⬆️

---

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 c90b75d...301e024. Read the comment docs.

[effigies]

I have, somewhat tediously, looked through every instance of GiftiNVPairs I could find on GitHub, and there are many uses cases where people darray.meta.data.insert(0, GiftiNVPairs(...)), but fortunately nobody seems to be slicing or anything like that.

One thing that isn't addressed here is that we no longer offer anything but insertion order, so people who previously put insert(0, ...) will not get the same result out. It doesn't seem like a major crime, though.

[htwangtw]

The test looks comprehensive and I cannot think of anything else. LGTM

[effigies]

Thanks!