TurboQuant cleanup part 2

[connortsui20]

Summary

Tracking issue: #7297

Adds named limit constants, removes the old double-quantized similarity path, and updates the docs to better match the current Stage 1 implementation.

It also replaces all remaining to_canonical usage in vortex-tensor with explicit execute(...) calls and tightens cast_possible_truncation handling by using targeted expect annotations or checked conversions instead of broad allows.

API Changes

Added 2 constants.

Testing

NA

[codspeed-hq]

Merging this PR will degrade performance by 17.72%

❌ 1 regressed benchmark
✅ 1121 untouched benchmarks
⏩ 1530 skipped benchmarks¹

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

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	chunked_bool_canonical_into[(1000, 10)]	50.1 µs	60.8 µs	-17.72%

---

_{Comparing ct/tq-cleanup-2 (ce476ff) with develop (be65147)}

Footnotes

1. 1530 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. ↩