nx_fasta.nx

nx_fasta.nx -- FASTA single-record stream parser + strip-whitespace. license_tier: INDEPENDENT_REDERIVE genealogy_id: international-research-sources/lipman-1985-fasta-format G2.0c of NISHI_GENOMICS_SUBSTRATE_ROADMAP.md. Reference-genome ingestion bridge: most reference assemblies (GRCh38, T2T-CHM13, every UCSC + Ensembl + NCBI download) ship as gzipped FASTA where each chromosome / contig is a record with multi-line wrapped sequence (canonical wrap = 60 or 80 chars per line). FASTA canonical format (Lipman + Pearson 1985): line 1: '>' + identifier [+ optional description] line 2..N: nucleotide (or amino-acid) sequence, line-wrapped record ends at next '>' or EOF Differs from FASTQ: - '>' marker (not '@') - Sequence WRAPS across multiple lines -- any line not starting with '>' is sequence content (no '+' / qual lines) - No quality scores API: fasta_parse_one_record(buf, buf_len, offset, out_meta) -> i64 out_meta layout (5 i64s): [0] header_start (offset after the '>', start of identifier) [1] header_len (bytes from header_start to end of line 1) [2] seq_start (first byte of seq region) [3] seq_end_excl (one past last byte of seq region; == offset of next '>' or buf_len at EOF) [4] next_offset (where the next record begins; same as seq_end_excl) Return: 0 success / 1 EOF / -1 malformed fasta_strip_whitespace(src, src_start, src_end_excl, dst, max_dst) -> i64 Copies src[src_start..src_end_excl) into dst, skipping whitespace bytes (LF 0x0A, CR 0x0D, space 0x20, tab 0x09). Returns count written or -1 if max_dst exceeded. Typical caller flow: