perf: improve alignment::distance::levenshtein and alignment::distance::bounded_levenshtein on strrings where distance is small

[nkkarpov]

Probably, it is better to use the implementation of the Ukkonen algorithm from editdistancek.
Before:

test bench_levenshtein_dist_diverse_str              ... bench: 113,214,418 ns/iter (+/- 781,509)
test bench_levenshtein_dist_equal_str                ... bench:  91,783,885 ns/iter (+/- 183,007)
test bench_levenshtein_dist_worst_str                ... bench:  91,866,787 ns/iter (+/- 1,736,995)
test bench_simd_bounded_levenshtein_dist_diverse_str ... bench:  41,695,427 ns/iter (+/- 1,917,682)
test bench_simd_bounded_levenshtein_dist_equal_str   ... bench:  41,640,232 ns/iter (+/- 2,047,975)
test bench_simd_bounded_levenshtein_dist_worst_str   ... bench:  41,474,403 ns/iter (+/- 3,183,653)
test bench_simd_levenshtein_dist_diverse_str         ... bench:  54,980,802 ns/iter (+/- 2,732,004)
test bench_simd_levenshtein_dist_equal_str           ... bench:      75,170 ns/iter (+/- 1,705)
test bench_simd_levenshtein_dist_worst_str           ... bench: 100,214,002 ns/iter (+/- 21,694,930)

After:

test bench_levenshtein_dist_diverse_str              ... bench:  39,735,136 ns/iter (+/- 194,398)
test bench_levenshtein_dist_equal_str                ... bench:       6,711 ns/iter (+/- 324)
test bench_levenshtein_dist_worst_str                ... bench: 137,810,484 ns/iter (+/- 340,456)
test bench_simd_bounded_levenshtein_dist_diverse_str ... bench:  39,740,574 ns/iter (+/- 159,197)
test bench_simd_bounded_levenshtein_dist_equal_str   ... bench:       6,572 ns/iter (+/- 511)
test bench_simd_bounded_levenshtein_dist_worst_str   ... bench: 137,862,613 ns/iter (+/- 724,980)
test bench_simd_levenshtein_dist_diverse_str         ... bench:  54,768,509 ns/iter (+/- 2,092,364)
test bench_simd_levenshtein_dist_equal_str           ... bench:      75,626 ns/iter (+/- 3,295)
test bench_simd_levenshtein_dist_worst_str           ... bench:  96,541,484 ns/iter (+/- 2,901,857)

The implementation works in time O(nd) where d is the edit distance between strings (+simd optimization). The implementation outperforms the previous version except when the edit distance is close to the maximum possible value. You can check more benchmarks at https://github.com/nkkarpov/editdistancek . On long random strings (1k) with random edits the version from editdistancek can achieve a 24x speedup compared to triple_accell when the percent of edits is less than 1. When the percent of edits is at most 20, and the upper bound is equal to 2 times the real distance, the new implementation has at least a 2.5x speedup. The algorithm is threshold oblivious, and the upper bound on the distance is used only for memory allocation (the algorithm requires O(k) auxiliary memory where k is the upper bound on the distance).

[coveralls]

coverage: 81.001% (-0.07%) from 81.067% when pulling 0ac8365 on nkkarpov:master into 640caa0 on rust-bio:master.

[RagnarGrootKoerkamp]

This doesn't compile on macbook M1.
RagnarGrootKoerkamp/astar-pairwise-aligner#17
nkkarpov/editdistancek#2

Especially the SIMD instructions seem broken.

[nkkarpov]

It should work now.