Migrate alignment functions to Rust

[padix-key]

Continuing #688, this PR migrates the following functions to Rust:

• align_optimal()
• align_banded()
• align_local_gapped()
• align_local_ungapped()
• align_ungapped()

In addition now all these functions support

• match/mismatch scoring schemes without a substitution matrix
• an faster score-only modes, only returning the alignment score instead of the full alignment.

As align_local_ungapped() already runs fast, the Python interface becomes the bottleneck. Hence the function is deprecated and replaced by SeedExtension, which is only created once and run for multiple ungapped local alignments, reducing repetitive input processing in the Python interface

[codspeed-hq]

Merging this PR will improve performance by ×2

⚡ 3 improved benchmarks
✅ 94 untouched benchmarks
🆕 6 new benchmarks
⏩ 4 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	benchmark_align_multiple	19.2 s	4.3 s	×4.5
⚡	benchmark_concatenate	195.4 µs	129.5 µs	+50.87%
⚡	benchmark_superimpose[superimpose_structural_homologs]	29.5 ms	23.8 ms	+23.67%
🆕	benchmark_align_pairwise[align_banded-score_only]	N/A	3.8 ms	N/A
🆕	benchmark_align_pairwise[align_banded-traceback]	N/A	6.1 ms	N/A
🆕	benchmark_align_pairwise[align_local_gapped-score_only]	N/A	45.4 ms	N/A
🆕	benchmark_align_pairwise[align_local_gapped-traceback]	N/A	58.1 ms	N/A
🆕	benchmark_align_pairwise[align_optimal-score_only]	N/A	66.4 ms	N/A
🆕	benchmark_align_pairwise[align_optimal-traceback]	N/A	78.5 ms	N/A

Tip

Curious why this is faster? Comment @codspeedbot explain why this is faster on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing padix-key:rust-alignment (fe1601a) with main (a2cab9a)²}

Footnotes

1. 4 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 main (641afc4) during the generation of this report, so a2cab9a was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩