perf: make VectorStorage and DistCalculator static to generate better code

[BubbleCal]

4.3647% faster

[BubbleCal]

blocked by #2353

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 73.17073% with 11 lines in your changes are missing coverage. Please review.

Project coverage is 80.69%. Comparing base (21d6a25) to head (fea075a).
Report is 4 commits behind head on main.

Files	Patch %	Lines
rust/lance-index/src/vector/hnsw/builder.rs	81.25%	0 Missing and 3 partials ⚠️
rust/lance-index/src/vector/sq/storage.rs	40.00%	3 Missing ⚠️
rust/lance-index/src/vector/pq/storage.rs	75.00%	1 Missing ⚠️
rust/lance/src/index/vector/fixture_test.rs	0.00%	1 Missing ⚠️
rust/lance/src/index/vector/ivf.rs	0.00%	1 Missing ⚠️
rust/lance/src/index/vector/pq.rs	0.00%	1 Missing ⚠️
rust/lance/src/session/index_extension.rs	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2355      +/-   ##
==========================================
- Coverage   80.72%   80.69%   -0.04%     
==========================================
  Files         192      191       -1     
  Lines       56230    56355     +125     
  Branches    56230    56355     +125     
==========================================
+ Hits        45394    45478      +84     
- Misses       8213     8248      +35     
- Partials     2623     2629       +6     

Flag	Coverage Δ
unittests	80.69% <73.17%> (-0.04%)	⬇️

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.

[wjones127]

By switching to impl, that makes this generic and increases generated code size. Do we think the speedup is worth it?

[BubbleCal]

this leads up to making distance calculator generic,
let me finish it and bench it first.

[jacketsj]

We only have a handful of vector storage types anyway I believe. 4.3% definitely seems worth it to me

[chebbyChefNEQ]

let's measure the binary size before and after? but yeah, 4.3% for single point will turn into more when we enable batch query. If size increase is still reasonable, say 1-2 MB, I think we should go forward with this.