feat: make scalar index training configurable

[dsgibbons]

Closes #2661.

This PR adds a scan_unordered_chunks method to the TrainingSource trait and provides basic implementations where required. train_bitmap_index has been updated to take advantage of this new method. Please advise if any other indexing algorithms do not require sorted columns and I'll update accordingly.

[dsgibbons]

@westonpace looks like UnnestTrainingSource also needs scan_unordered_chunks to be implemented as train_label_list_index calls train_bitmap_index.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 76.19048% with 10 lines in your changes missing coverage. Please review.

Project coverage is 79.45%. Comparing base (712405e) to head (5105a01).
Report is 5 commits behind head on main.

Files	Patch %	Lines
rust/lance-index/src/scalar/btree.rs	0.00%	4 Missing ⚠️
rust/lance/src/index/scalar.rs	81.25%	0 Missing and 3 partials ⚠️
rust/lance-index/src/scalar/label_list.rs	88.88%	1 Missing and 1 partial ⚠️
rust/lance-index/src/scalar/bitmap.rs	0.00%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2686      +/-   ##
==========================================
- Coverage   79.61%   79.45%   -0.17%     
==========================================
  Files         226      227       +1     
  Lines       66516    66920     +404     
  Branches    66516    66920     +404     
==========================================
+ Hits        52955    53168     +213     
- Misses      10465    10626     +161     
- Partials     3096     3126      +30     

Flag	Coverage Δ
unittests	79.45% <76.19%> (-0.17%)	⬇️

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.

[BubbleCal]

Great work, thanks!

[westonpace]

Agreed. This looks great, just a few minor naming suggestions.

[westonpace]

FYI: I filed the original issue because I was doing some testing on a dataset with 1 billion rows and after 90 minutes it still had not created a bitmap index on a boolean column. I just tested with this PR and was able to create a bitmap index in 90 seconds 🎉

[dsgibbons]

Agreed. This looks great, just a few minor naming suggestions.

Done!