#include "sam_header.h"
#include "sam.h"
#include "faidx.h"
+#include "kstring.h"
#include "khash.h"
KHASH_SET_INIT_STR(rg)
+// When counting records instead of printing them,
+// data passed to the bam_fetch callback is encapsulated in this struct.
+typedef struct {
+ bam_header_t *header;
+ int *count;
+} count_func_data_t;
+
typedef khash_t(rg) *rghash_t;
-rghash_t g_rghash = 0;
+static rghash_t g_rghash = 0;
static int g_min_mapQ = 0, g_flag_on = 0, g_flag_off = 0;
static char *g_library, *g_rg;
static int g_sol2sanger_tbl[128];
+static void *g_bed;
+
+void *bed_read(const char *fn);
+void bed_destroy(void *_h);
+int bed_overlap(const void *_h, const char *chr, int beg, int end);
static void sol2sanger(bam1_t *b)
{
{
if (b->core.qual < g_min_mapQ || ((b->core.flag & g_flag_on) != g_flag_on) || (b->core.flag & g_flag_off))
return 1;
+ if (g_bed && b->core.tid >= 0 && !bed_overlap(g_bed, h->target_name[b->core.tid], b->core.pos, bam_calend(&b->core, bam1_cigar(b))))
+ return 1;
if (g_rg || g_rghash) {
uint8_t *s = bam_aux_get(b, "RG");
if (s) {
return 0;
}
-// callback function for bam_fetch()
+static char *drop_rg(char *hdtxt, rghash_t h, int *len)
+{
+ char *p = hdtxt, *q, *r, *s;
+ kstring_t str;
+ memset(&str, 0, sizeof(kstring_t));
+ while (1) {
+ int toprint = 0;
+ q = strchr(p, '\n');
+ if (q == 0) q = p + strlen(p);
+ if (q - p < 3) break; // the line is too short; then stop
+ if (strncmp(p, "@RG\t", 4) == 0) {
+ int c;
+ khint_t k;
+ if ((r = strstr(p, "\tID:")) != 0) {
+ r += 4;
+ for (s = r; *s != '\0' && *s != '\n' && *s != '\t'; ++s);
+ c = *s; *s = '\0';
+ k = kh_get(rg, h, r);
+ *s = c;
+ if (k != kh_end(h)) toprint = 1;
+ }
+ } else toprint = 1;
+ if (toprint) {
+ kputsn(p, q - p, &str); kputc('\n', &str);
+ }
+ p = q + 1;
+ }
+ *len = str.l;
+ return str.s;
+}
+
+// callback function for bam_fetch() that prints nonskipped records
static int view_func(const bam1_t *b, void *data)
{
if (!__g_skip_aln(((samfile_t*)data)->header, b))
return 0;
}
+// callback function for bam_fetch() that counts nonskipped records
+static int count_func(const bam1_t *b, void *data)
+{
+ if (!__g_skip_aln(((count_func_data_t*)data)->header, b)) {
+ (*((count_func_data_t*)data)->count)++;
+ }
+ return 0;
+}
+
static int usage(int is_long_help);
int main_samview(int argc, char *argv[])
{
- int c, is_header = 0, is_header_only = 0, is_bamin = 1, ret = 0, is_uncompressed = 0, is_bamout = 0, slx2sngr = 0;
+ int c, is_header = 0, is_header_only = 0, is_bamin = 1, ret = 0, compress_level = -1, is_bamout = 0, slx2sngr = 0, is_count = 0;
int of_type = BAM_OFDEC, is_long_help = 0;
+ int count = 0;
samfile_t *in = 0, *out = 0;
char in_mode[5], out_mode[5], *fn_out = 0, *fn_list = 0, *fn_ref = 0, *fn_rg = 0;
/* parse command-line options */
strcpy(in_mode, "r"); strcpy(out_mode, "w");
- while ((c = getopt(argc, argv, "Sbt:hHo:q:f:F:ul:r:xX?T:CR:")) >= 0) {
+ while ((c = getopt(argc, argv, "Sbct:h1Ho:q:f:F:ul:r:xX?T:CR:L:")) >= 0) {
switch (c) {
+ case 'c': is_count = 1; break;
case 'C': slx2sngr = 1; break;
case 'S': is_bamin = 0; break;
case 'b': is_bamout = 1; break;
case 'f': g_flag_on = strtol(optarg, 0, 0); break;
case 'F': g_flag_off = strtol(optarg, 0, 0); break;
case 'q': g_min_mapQ = atoi(optarg); break;
- case 'u': is_uncompressed = 1; break;
+ case 'u': compress_level = 0; break;
+ case '1': compress_level = 1; break;
case 'l': g_library = strdup(optarg); break;
+ case 'L': g_bed = bed_read(optarg); break;
case 'r': g_rg = strdup(optarg); break;
case 'R': fn_rg = strdup(optarg); break;
case 'x': of_type = BAM_OFHEX; break;
default: return usage(is_long_help);
}
}
- if (is_uncompressed) is_bamout = 1;
+ if (compress_level >= 0) is_bamout = 1;
if (is_header_only) is_header = 1;
if (is_bamout) strcat(out_mode, "b");
else {
}
if (is_bamin) strcat(in_mode, "b");
if (is_header) strcat(out_mode, "h");
- if (is_uncompressed) strcat(out_mode, "u");
+ if (compress_level >= 0) {
+ char tmp[2];
+ tmp[0] = compress_level + '0'; tmp[1] = '\0';
+ strcat(out_mode, tmp);
+ }
if (argc == optind) return usage(is_long_help); // potential memory leak...
// read the list of read groups
if (fn_list == 0 && fn_ref) fn_list = samfaipath(fn_ref);
// open file handlers
if ((in = samopen(argv[optind], in_mode, fn_list)) == 0) {
- fprintf(stderr, "[main_samview] fail to open file for reading.\n");
+ fprintf(stderr, "[main_samview] fail to open \"%s\" for reading.\n", argv[optind]);
+ ret = 1;
goto view_end;
}
if (in->header == 0) {
- fprintf(stderr, "[main_samview] fail to read the header.\n");
+ fprintf(stderr, "[main_samview] fail to read the header from \"%s\".\n", argv[optind]);
+ ret = 1;
goto view_end;
}
- if ((out = samopen(fn_out? fn_out : "-", out_mode, in->header)) == 0) {
- fprintf(stderr, "[main_samview] fail to open file for writing.\n");
+ if (g_rghash) { // FIXME: I do not know what "bam_header_t::n_text" is for...
+ char *tmp;
+ int l;
+ tmp = drop_rg(in->header->text, g_rghash, &l);
+ free(in->header->text);
+ in->header->text = tmp;
+ in->header->l_text = l;
+ }
+ if (!is_count && (out = samopen(fn_out? fn_out : "-", out_mode, in->header)) == 0) {
+ fprintf(stderr, "[main_samview] fail to open \"%s\" for writing.\n", fn_out? fn_out : "standard output");
+ ret = 1;
goto view_end;
}
if (is_header_only) goto view_end; // no need to print alignments
while ((r = samread(in, b)) >= 0) { // read one alignment from `in'
if (!__g_skip_aln(in->header, b)) {
if (slx2sngr) sol2sanger(b);
- samwrite(out, b); // write the alignment to `out'
+ if (!is_count) samwrite(out, b); // write the alignment to `out'
+ count++;
}
}
- if (r < -1) fprintf(stderr, "[main_samview] truncated file.\n");
+ if (r < -1) {
+ fprintf(stderr, "[main_samview] truncated file.\n");
+ ret = 1;
+ }
bam_destroy1(b);
} else { // retrieve alignments in specified regions
int i;
goto view_end;
}
for (i = optind + 1; i < argc; ++i) {
- int tid, beg, end;
+ int tid, beg, end, result;
bam_parse_region(in->header, argv[i], &tid, &beg, &end); // parse a region in the format like `chr2:100-200'
if (tid < 0) { // reference name is not found
- fprintf(stderr, "[main_samview] fail to get the reference name. Continue anyway.\n");
+ fprintf(stderr, "[main_samview] region \"%s\" specifies an unknown reference name. Continue anyway.\n", argv[i]);
continue;
}
// fetch alignments
- if (bam_fetch(in->x.bam, idx, tid, beg, end, out, view_func) < 0) {
- fprintf(stderr, "[main_samview] retrieval failed due to truncated file or corrupt BAM index file\n");
+ if (is_count) {
+ count_func_data_t count_data = { in->header, &count };
+ result = bam_fetch(in->x.bam, idx, tid, beg, end, &count_data, count_func);
+ } else
+ result = bam_fetch(in->x.bam, idx, tid, beg, end, out, view_func);
+ if (result < 0) {
+ fprintf(stderr, "[main_samview] retrieval of region \"%s\" failed due to truncated file or corrupt BAM index file\n", argv[i]);
ret = 1;
break;
}
}
view_end:
+ if (is_count && ret == 0) {
+ printf("%d\n", count);
+ }
// close files, free and return
free(fn_list); free(fn_ref); free(fn_out); free(g_library); free(g_rg); free(fn_rg);
+ if (g_bed) bed_destroy(g_bed);
if (g_rghash) {
khint_t k;
for (k = 0; k < kh_end(g_rghash); ++k)
kh_destroy(rg, g_rghash);
}
samclose(in);
- samclose(out);
+ if (!is_count)
+ samclose(out);
return ret;
}
fprintf(stderr, " -H print header only (no alignments)\n");
fprintf(stderr, " -S input is SAM\n");
fprintf(stderr, " -u uncompressed BAM output (force -b)\n");
+ fprintf(stderr, " -1 fast compression (force -b)\n");
fprintf(stderr, " -x output FLAG in HEX (samtools-C specific)\n");
fprintf(stderr, " -X output FLAG in string (samtools-C specific)\n");
+ fprintf(stderr, " -c print only the count of matching records\n");
+ fprintf(stderr, " -L FILE output alignments overlapping the input BED FILE [null]\n");
fprintf(stderr, " -t FILE list of reference names and lengths (force -S) [null]\n");
fprintf(stderr, " -T FILE reference sequence file (force -S) [null]\n");
fprintf(stderr, " -o FILE output file name [stdout]\n");