Clean up vortex-tensor

[connortsui20]

Summary

Tracking issue: #7297

Given our fast velocity on this crate, quite a few things slipped through the cracks.

This change cleans up the vortex-tensor crate by clearly defining the abstraction points, fixing a few bugs (the only real bug was a Tensor Display bug), cleaning up some TODOs, and generally raising the quality.

API Changes

The only relevant change is new helper functions

Testing

Just fixed up existing tests.

[codspeed-hq]

Merging this PR will degrade performance by 18.79%

❌ 6 regressed benchmarks
✅ 1152 untouched benchmarks
🆕 5 new benchmarks
⏩ 1462 skipped benchmarks¹

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	old_alp_prim_test_between[f64, 32768]	284.8 µs	350.7 µs	-18.79%
❌	Simulation	patched_take_10k_contiguous_patches	228.1 µs	258.2 µs	-11.64%
❌	Simulation	patched_take_10k_random	240.4 µs	270.4 µs	-11.12%
❌	Simulation	take_10k_first_chunk_only	225.9 µs	270.9 µs	-16.59%
❌	Simulation	patched_take_10k_contiguous_not_patches	228.5 µs	258.6 µs	-11.62%
❌	Simulation	take_10k_dispersed	239.8 µs	284.7 µs	-15.77%
🆕	Simulation	turboquant_encode_dim768_4bit	N/A	53 ms	N/A
🆕	Simulation	turboquant_encode_dim128_4bit	N/A	6.7 ms	N/A
🆕	Simulation	turboquant_encode_dim1024_4bit	N/A	53.5 ms	N/A
🆕	Simulation	turboquant_encode_dim1024_2bit	N/A	48.3 ms	N/A
🆕	Simulation	turboquant_encode_dim1024_8bit	N/A	63.7 ms	N/A

---

_{Comparing ct/cleanup-tensor (d7f386c) with develop (4135209)²}

Footnotes

1. 1462 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

2. No successful run was found on develop (a6dd776) during the generation of this report, so 4135209 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

[connortsui20]

This is just moved down

[connortsui20]

I forgot that this is a completely separate label than the array IDs, and this is confined to just ID equality checking in the compressor