#include <stdio.h>
#include <stdlib.h>
#include <ctype.h>
-#include "bam.h"
+#include <assert.h>
+#include "sam.h"
typedef struct __linkbuf_t {
bam1_t b;
int ret = 1, is_restart = 1;
if (c->flag&BAM_FUNMAP) return 0; // unmapped read
- assert(x <= pos);
+ assert(x <= pos); // otherwise a bug
p->qpos = -1; p->indel = 0; p->is_del = p->is_head = p->is_tail = 0;
for (k = 0; k < c->n_cigar; ++k) {
int op = bam1_cigar(b)[k] & BAM_CIGAR_MASK; // operation
int op_next = cigar&BAM_CIGAR_MASK; // next CIGAR operation
if (op_next == BAM_CDEL) p->indel = -(int32_t)(cigar>>BAM_CIGAR_SHIFT); // del
else if (op_next == BAM_CINS) p->indel = cigar>>BAM_CIGAR_SHIFT; // ins
+ if (op_next == BAM_CDEL || op_next == BAM_CINS) {
+ if (k + 2 < c->n_cigar) op_next = bam1_cigar(b)[k+2]&BAM_CIGAR_MASK;
+ else p->is_tail = 1;
+ }
if (op_next == BAM_CSOFT_CLIP || op_next == BAM_CREF_SKIP || op_next == BAM_CHARD_CLIP)
p->is_tail = 1; // tail
} else p->is_tail = 1; // this is the last operation; set tail
break;
}
}
- assert(x > pos);
+ assert(x > pos); // otherwise a bug
return ret;
}
int32_t tid, pos, max_tid, max_pos;
int max_pu, is_eof;
bam_pileup1_t *pu;
+ int flag_mask;
};
void bam_plbuf_reset(bam_plbuf_t *buf)
buf->head = buf->tail;
}
+void bam_plbuf_set_mask(bam_plbuf_t *buf, int mask)
+{
+ if (mask < 0) buf->flag_mask = BAM_DEF_MASK;
+ else buf->flag_mask = BAM_FUNMAP | mask;
+}
+
bam_plbuf_t *bam_plbuf_init(bam_pileup_f func, void *data)
{
bam_plbuf_t *buf;
buf->head = buf->tail = mp_alloc(buf->mp);
buf->dummy = mp_alloc(buf->mp);
buf->max_tid = buf->max_pos = -1;
+ buf->flag_mask = BAM_DEF_MASK;
return buf;
}
int bam_plbuf_push(const bam1_t *b, bam_plbuf_t *buf)
{
if (b) { // fill buffer
+ if (b->core.tid < 0) return 0;
+ if (b->core.flag & buf->flag_mask) return 0;
bam_copy1(&buf->tail->b, b);
buf->tail->beg = b->core.pos; buf->tail->end = bam_calend(&b->core, bam1_cigar(b));
- if (!(b->core.tid >= buf->max_tid || (b->core.tid == buf->max_tid && buf->tail->beg >= buf->max_pos))) {
- fprintf(stderr, "[bam_pileup_core] the input is not sorted. Abort!\n");
- abort();
+ if (b->core.tid < buf->max_tid) {
+ fprintf(stderr, "[bam_pileup_core] the input is not sorted (chromosomes out of order)\n");
+ return -1;
+ }
+ if ((b->core.tid == buf->max_tid) && (buf->tail->beg < buf->max_pos)) {
+ fprintf(stderr, "[bam_pileup_core] the input is not sorted (reads out of order)\n");
+ return -1;
}
buf->max_tid = b->core.tid; buf->max_pos = buf->tail->beg;
- if (buf->tail->end > buf->pos) {
+ if (buf->tail->end > buf->pos || buf->tail->b.core.tid > buf->tid) {
buf->tail->next = mp_alloc(buf->mp);
buf->tail = buf->tail->next;
}
buf->func(buf->tid, buf->pos, n_pu, buf->pu, buf->func_data);
}
// update tid and pos
- if (buf->head->next) assert(buf->tid <= buf->head->b.core.tid); // otherwise, not sorted
+ if (buf->head->next) {
+ if (buf->tid > buf->head->b.core.tid) {
+ fprintf(stderr, "[bam_plbuf_push] unsorted input. Pileup aborts.\n");
+ return 1;
+ }
+ }
if (buf->tid < buf->head->b.core.tid) { // come to a new reference sequence
buf->tid = buf->head->b.core.tid; buf->pos = buf->head->beg; // jump to the next reference
} else if (buf->pos < buf->head->beg) { // here: tid == head->b.core.tid
return 0;
}
-int bam_pileup_file(bamFile fp, bam_pileup_f func, void *func_data)
+int bam_pileup_file(bamFile fp, int mask, bam_pileup_f func, void *func_data)
{
bam_plbuf_t *buf;
int ret;
bam1_t *b;
- b = (bam1_t*)calloc(1, sizeof(bam1_t));
+ b = bam_init1();
buf = bam_plbuf_init(func, func_data);
+ bam_plbuf_set_mask(buf, mask);
while ((ret = bam_read1(fp, b)) >= 0)
bam_plbuf_push(b, buf);
bam_plbuf_push(0, buf);
bam_plbuf_destroy(buf);
- free(b->data); free(b);
+ bam_destroy1(b);
return 0;
}