X-Git-Url: http://woldlab.caltech.edu/gitweb/?p=samtools.git;a=blobdiff_plain;f=bam.c;h=5fac17df6ec27d7cbb04b5600c3ce54d5c06b354;hp=94b0aa8bb85bef4ae4a787af3337f1765ed8899e;hb=e582623cf8c4778c7dc792318635821d3c494b0d;hpb=cb12a866906ec4ac644de0e658679261c82ab098 diff --git a/bam.c b/bam.c index 94b0aa8..5fac17d 100644 --- a/bam.c +++ b/bam.c @@ -7,7 +7,7 @@ #include "kstring.h" #include "sam_header.h" -int bam_is_be = 0; +int bam_is_be = 0, bam_verbose = 2; char *bam_flag2char_table = "pPuUrR12sfd\0\0\0\0\0"; /************************** @@ -79,7 +79,7 @@ bam_header_t *bam_header_read(bamFile fp) // with ESPIPE. Suppress the error message in this case. if (errno != ESPIPE) perror("[bam_header_read] bgzf_check_EOF"); } - else if (i == 0) fprintf(stderr, "[bam_header_read] EOF marker is absent.\n"); + else if (i == 0) fprintf(stderr, "[bam_header_read] EOF marker is absent. The input is probably truncated.\n"); // read "BAM1" magic_len = bam_read(fp, buf, 4); if (magic_len != 4 || strncmp(buf, "BAM\001", 4) != 0) { @@ -160,6 +160,19 @@ static void swap_endian_data(const bam1_core_t *c, int data_len, uint8_t *data) else if (type == 'I' || type == 'F') { bam_swap_endian_4p(s); s += 4; } else if (type == 'D') { bam_swap_endian_8p(s); s += 8; } else if (type == 'Z' || type == 'H') { while (*s) ++s; ++s; } + else if (type == 'B') { + int32_t n, Bsize = bam_aux_type2size(*s); + memcpy(&n, s + 1, 4); + if (1 == Bsize) { + } else if (2 == Bsize) { + for (i = 0; i < n; i += 2) + bam_swap_endian_2p(s + 5 + i); + } else if (4 == Bsize) { + for (i = 0; i < n; i += 4) + bam_swap_endian_4p(s + 5 + i); + } + bam_swap_endian_4p(s+1); + } } } @@ -245,7 +258,11 @@ char *bam_format1_core(const bam_header_t *header, const bam1_t *b, int of) kputc('\t', &str); } if (c->tid < 0) kputsn("*\t", 2, &str); - else { kputs(header->target_name[c->tid], &str); kputc('\t', &str); } + else { + if (header) kputs(header->target_name[c->tid] , &str); + else kputw(c->tid, &str); + kputc('\t', &str); + } kputw(c->pos + 1, &str); kputc('\t', &str); kputw(c->qual, &str); kputc('\t', &str); if (c->n_cigar == 0) kputc('*', &str); else { @@ -257,7 +274,11 @@ char *bam_format1_core(const bam_header_t *header, const bam1_t *b, int of) kputc('\t', &str); if (c->mtid < 0) kputsn("*\t", 2, &str); else if (c->mtid == c->tid) kputsn("=\t", 2, &str); - else { kputs(header->target_name[c->mtid], &str); kputc('\t', &str); } + else { + if (header) kputs(header->target_name[c->mtid], &str); + else kputw(c->mtid, &str); + kputc('\t', &str); + } kputw(c->mpos + 1, &str); kputc('\t', &str); kputw(c->isize, &str); kputc('\t', &str); if (c->l_qseq) { for (i = 0; i < c->l_qseq; ++i) kputc(bam_nt16_rev_table[bam1_seqi(s, i)], &str); @@ -281,6 +302,23 @@ char *bam_format1_core(const bam_header_t *header, const bam1_t *b, int of) else if (type == 'f') { ksprintf(&str, "f:%g", *(float*)s); s += 4; } else if (type == 'd') { ksprintf(&str, "d:%lg", *(double*)s); s += 8; } else if (type == 'Z' || type == 'H') { kputc(type, &str); kputc(':', &str); while (*s) kputc(*s++, &str); ++s; } + else if (type == 'B') { + uint8_t sub_type = *(s++); + int32_t n; + memcpy(&n, s, 4); + s += 4; // no point to the start of the array + kputc(type, &str); kputc(':', &str); kputc(sub_type, &str); // write the typing + for (i = 0; i < n; ++i) { + kputc(',', &str); + if ('c' == sub_type || 'c' == sub_type) { kputw(*(int8_t*)s, &str); ++s; } + else if ('C' == sub_type) { kputw(*(uint8_t*)s, &str); ++s; } + else if ('s' == sub_type) { kputw(*(int16_t*)s, &str); s += 2; } + else if ('S' == sub_type) { kputw(*(uint16_t*)s, &str); s += 2; } + else if ('i' == sub_type) { kputw(*(int32_t*)s, &str); s += 4; } + else if ('I' == sub_type) { kputuw(*(uint32_t*)s, &str); s += 4; } + else if ('f' == sub_type) { ksprintf(&str, "%g", *(float*)s); s += 4; } + } + } } return str.s; } @@ -297,6 +335,22 @@ void bam_view1(const bam_header_t *header, const bam1_t *b) free(s); } +int bam_validate1(const bam_header_t *header, const bam1_t *b) +{ + char *s; + + if (b->core.tid < -1 || b->core.mtid < -1) return 0; + if (header && (b->core.tid >= header->n_targets || b->core.mtid >= header->n_targets)) return 0; + + if (b->data_len < b->core.l_qname) return 0; + s = memchr(bam1_qname(b), '\0', b->core.l_qname); + if (s != &bam1_qname(b)[b->core.l_qname-1]) return 0; + + // FIXME: Other fields could also be checked, especially the auxiliary data + + return 1; +} + // FIXME: we should also check the LB tag associated with each alignment const char *bam_get_library(bam_header_t *h, const bam1_t *b) {