nx_align_match.nx

nx_align_match.nx -- minimizer-based seed match (query vs reference). license_tier: INDEPENDENT_REDERIVE genealogy_id: international-research-sources/li-2018-minimap2 G1.2 of NISHI_GENOMICS_SUBSTRATE_ROADMAP.md. Composes the minimizer-extraction primitive (G1.1) into the seeding stage of read alignment. Pipeline: reference: bases -> minimizer_extract -> (r_vals, r_pos) query : bases -> minimizer_extract -> (q_vals, q_pos) seeds : minimizer_match(q, r) -> {(q_pos, r_pos) : q_val == r_val} Each emitted (q_pos, r_pos) pair is a candidate seed: a k-mer of the query that occurs at r_pos in the reference. Downstream (G1.3 chaining + G1.4 SW extension) groups compatible seeds into chains and extends them into full alignments. Naive O(q_count * r_count) impl for G1.2. For real-world use the reference minimizer index is sorted-by-value or hashed so match is O(q_count * log r_count) or O(q_count) -- those live in nx_align_match_indexed.nx (G1.4). API: minimizer_match(q_vals, q_pos, q_count, r_vals, r_pos, r_count, out_q_pos, out_r_pos, max_out) -> i64 pair count nx_safety_envelope: (schema: nishi-library/seeds/safety-critical-standards.toml) intended_use: "Seed match in read-to-reference alignment; composes with nx_align Smith-Waterman to complete the seed-and-extend pipeline" sil_target: SIL2 asil_target: QM dal_target: DAL C iec_62304_class: B evidence: [no_floating_point, deterministic, bit_equal_reproducible, composes_nx_align_minimizer_KAT,