Implement k-mer distance metrics and constrained generation

Adds support for Hamming distance, constrained k-mer generation, and optimized Longest Common Subsequence (LCS) calculations using scalar, batch, and SIMD methods.

This includes new methods for k-mer distance calculation, structures for k-mer constraints, and SIMD-based batch processing for LCS to improve performance and correctness.

pkg/obikmer/kmer_lcs_obialign_test.go

@@ -0,0 +1,29 @@
+package obikmer_test
+
+import (
+	"math/rand"
+	"testing"
+
+	"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obialign"
+	"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obikmer"
+	"git.metabarcoding.org/obitools/obitools4/obitools4/pkg/obiseq"
+)
+
+// KmerLCS must reproduce obialign.FastLCSScore on the decoded sequences.
+func TestKmerLCSMatchesFastLCSScore(t *testing.T) {
+	rng := rand.New(rand.NewSource(4))
+	for _, k := range []int{1, 2, 4, 9, 15, 21, 31} {
+		mask := uint64(1)<<(2*uint(k)) - 1
+		for n := 0; n < 300; n++ {
+			a, b := rng.Uint64()&mask, rng.Uint64()&mask
+			sa := obiseq.NewBioSequence("a", obikmer.DecodeKmer(a, k, nil), "")
+			sb := obiseq.NewBioSequence("b", obikmer.DecodeKmer(b, k, nil), "")
+			lcs, alen := obialign.FastLCSScore(sa, sb, -1, nil)
+			got := obikmer.KmerLCS(a, b, k)
+			if got.LCS != lcs || got.AlignLen != alen {
+				t.Fatalf("k=%d %s/%s: KmerLCS=%+v, FastLCSScore=(%d,%d)",
+					k, sa.String(), sb.String(), got, lcs, alen)
+			}
+		}
+	}
+}