static void destroy_data(args_t *args)
{
if ( args->prefix )
{
fclose(args->fh_log);
int i, n = args->isec_op==OP_VENN ? 3 : args->files->nreaders;
for (i=0; i<n; i++)
{
if ( !args->fnames[i] ) continue;
hts_close(args->fh_out[i]);
if ( args->output_type==FT_VCF_GZ )
{
tbx_conf_t conf = tbx_conf_vcf;
tbx_index_build(args->fnames[i], -1, &conf);
}
else if ( args->output_type==FT_BCF_GZ )
{
if ( bcf_index_build(args->fnames[i],14) ) error("Could not index %s\n", args->fnames[i]);
}
free(args->fnames[i]);
}
free(args->fh_out);
free(args->fnames);
if ( args->fh_sites ) fclose(args->fh_sites);
if ( args->write ) free(args->write);
}
}
int main_vcfindex(int argc, char *argv[])
{
int c, force = 0, tbi = 0, stats = 0;
int min_shift = BCF_LIDX_SHIFT;
static struct option loptions[] =
{
{"csi",no_argument,NULL,'c'},
{"tbi",no_argument,NULL,'t'},
{"force",no_argument,NULL,'f'},
{"min-shift",required_argument,NULL,'m'},
{"stats",no_argument,NULL,'s'},
{"nrecords",no_argument,NULL,'n'},
{NULL, 0, NULL, 0}
};
char *tmp;
while ((c = getopt_long(argc, argv, "ctfm:sn", loptions, NULL)) >= 0)
{
switch (c)
{
case 'c': tbi = 0; break;
case 't': tbi = 1; min_shift = 0; break;
case 'f': force = 1; break;
case 'm':
min_shift = strtol(optarg,&tmp,10);
if ( *tmp ) error("Could not parse argument: --min-shift %s\n", optarg);
break;
case 's': stats |= 1; break;
case 'n': stats |= 2; break;
default: usage();
}
}
if ( optind==argc ) usage();
if (stats>2)
{
fprintf(stderr, "[E::%s] expected only one of --stats or --nrecords options\n", __func__);
return 1;
}
if (tbi && min_shift>0)
{
fprintf(stderr, "[E::%s] min-shift option only expected for CSI indices \n", __func__);
return 1;
}
if (min_shift < 0 || min_shift > 30)
{
fprintf(stderr, "[E::%s] expected min_shift in range [0,30] (%d)\n", __func__, min_shift);
return 1;
}
char *fname = argv[optind];
if (stats) return vcf_index_stats(fname, stats);
htsFile *fp = hts_open(fname,"r");
if ( !fp ) error("Failed to read %s\n", fname);
htsFormat type = *hts_get_format(fp);
hts_close(fp);
if ( (type.format!=bcf && type.format!=vcf) || type.compression!=bgzf )
{
fprintf(stderr, "[E::%s] unknown filetype; expected bgzip compressed VCF or BCF\n", __func__);
if ( type.compression!=bgzf )
fprintf(stderr, "[E::%s] was the VCF/BCF compressed with bgzip?\n", __func__);
return 1;
}
if (tbi && type.format==bcf)
{
fprintf(stderr, "[Warning] TBI-index does not work for BCF files. Generating CSI instead.\n");
tbi = 0; min_shift = BCF_LIDX_SHIFT;
}
if (min_shift == 0 && type.format==bcf)
{
fprintf(stderr, "[E::%s] Require min_shift>0 for BCF files.\n", __func__);
return 1;
}
if (!tbi && type.format==vcf && min_shift == 0)
{
fprintf(stderr, "[Warning] min-shift set to 0 for VCF file. Generating TBI file.\n");
tbi = 1;
}
if (!force)
{
// Before complaining about existing index, check if the VCF file isn't newer.
char *idx_fname = (char*)alloca(strlen(fname) + 5);
strcat(strcpy(idx_fname, fname), tbi ? ".tbi" : ".csi");
struct stat stat_tbi, stat_file;
if ( stat(idx_fname, &stat_tbi)==0 )
{
stat(fname, &stat_file);
if ( stat_file.st_mtime <= stat_tbi.st_mtime )
{
fprintf(stderr,"[E::%s] the index file exists. Please use '-f' to overwrite.\n", __func__);
return 1;
}
}
}
if (type.format==bcf)
{
if ( bcf_index_build(fname, min_shift) != 0 )
{
fprintf(stderr,"[E::%s] bcf_index_build failed for %s\n", __func__, fname);
return 1;
}
}
else
{
if ( tbx_index_build(fname, min_shift, &tbx_conf_vcf) != 0 )
{
fprintf(stderr,"[E::%s] tbx_index_build failed for %s\n", __func__, fname);
return 1;
}
}
return 0;
}
int main_vcfindex(int argc, char *argv[])
{
int c, min_shift = 14, force = 0;
static struct option loptions[] =
{
{"help",0,0,'h'},
{"force",0,0,'f'},
{"min-shift",1,0,'m'},
{0,0,0,0}
};
while ((c = getopt_long(argc, argv, "h?fm:", loptions,NULL)) >= 0)
{
switch (c)
{
case 'f': force = 1; break;
case 'm': min_shift = atoi(optarg); break;
default: usage();
}
}
if ( optind==argc ) usage();
if (min_shift < 0 || min_shift > 30)
{
fprintf(stderr, "[E::%s] expected min_shift in range [0,30] (%d)\n", __func__, min_shift);
return 1;
}
char *fname = argv[optind];
int ftype = hts_file_type(fname);
if (!ftype)
{
fprintf(stderr, "[E::%s] unknown filetype; expected .vcf.gz or .bcf\n", __func__);
return 1;
}
if (!force)
{
// Before complaining about existing index, check if the VCF file isn't newer.
char *idx_fname = (char*)alloca(strlen(fname) + 5);
strcat(strcpy(idx_fname, fname), min_shift <= 0 ? ".tbi" : ".csi");
struct stat stat_tbi, stat_file;
if ( stat(idx_fname, &stat_tbi)==0 )
{
stat(fname, &stat_file);
if ( stat_file.st_mtime <= stat_tbi.st_mtime )
{
fprintf(stderr,"[E::%s] the index file exists. Please use '-f' to overwrite.\n", __func__);
return 1;
}
}
}
if (ftype == FT_BCF_GZ)
{
if ( bcf_index_build(fname, min_shift) != 0 )
{
fprintf(stderr,"[E::%s] bcf_index_build failed: %s\n", __func__, fname);
return 1;
}
}
else if (ftype == FT_VCF_GZ)
{
if ( tbx_index_build(fname, min_shift, &tbx_conf_vcf) != 0 )
{
fprintf(stderr,"[E::%s] tbx_index_build failed for %s\n", __func__, fname);
return 1;
}
}
return 0;
}
int main_vcfindex(int argc, char *argv[])
{
int c, force = 0, tbi = 0;
int min_shift = BCF_LIDX_SHIFT;
static struct option loptions[] =
{
{"csi",no_argument,NULL,'c'},
{"tbi",no_argument,NULL,'t'},
{"force",no_argument,NULL,'f'},
{"min-shift",required_argument,NULL,'m'},
{NULL, 0, NULL, 0}
};
while ((c = getopt_long(argc, argv, "ctfm:", loptions, NULL)) >= 0)
{
switch (c)
{
case 'c': tbi = 0; break;
case 't': tbi = 1; min_shift = 0; break;
case 'f': force = 1; break;
case 'm': min_shift = atoi(optarg); break;
default: usage();
}
}
if ( optind==argc ) usage();
if (tbi && min_shift>0)
{
fprintf(stderr, "[E::%s] min-shift option only expected for CSI indices \n", __func__);
return 1;
}
if (min_shift < 0 || min_shift > 30)
{
fprintf(stderr, "[E::%s] expected min_shift in range [0,30] (%d)\n", __func__, min_shift);
return 1;
}
char *fname = argv[optind];
int ftype = hts_file_type(fname);
if (!ftype || (ftype != FT_BCF_GZ && ftype != FT_VCF_GZ))
{
fprintf(stderr, "[E::%s] unknown filetype; expected bgzip compressed VCF or BCF\n", __func__);
if (!(ftype & FT_GZ))
fprintf(stderr, "[E::%s] was the VCF/BCF compressed with bgzip?\n", __func__);
return 1;
}
if (tbi && ftype == FT_BCF_GZ)
{
fprintf(stderr, "[Warning] TBI-index does not work for BCF files. Generating CSI instead.\n");
tbi = 0; min_shift = BCF_LIDX_SHIFT;
}
if (min_shift == 0 && ftype == FT_BCF_GZ)
{
fprintf(stderr, "[E::%s] Require min_shift>0 for BCF files.\n", __func__);
return 1;
}
if (!tbi && ftype == FT_VCF_GZ && min_shift == 0)
{
fprintf(stderr, "[Warning] min-shift set to 0 for VCF file. Generating TBI file.\n");
tbi = 1;
}
if (!force)
{
// Before complaining about existing index, check if the VCF file isn't newer.
char *idx_fname = (char*)alloca(strlen(fname) + 5);
strcat(strcpy(idx_fname, fname), tbi ? ".tbi" : ".csi");
struct stat stat_tbi, stat_file;
if ( stat(idx_fname, &stat_tbi)==0 )
{
stat(fname, &stat_file);
if ( stat_file.st_mtime <= stat_tbi.st_mtime )
{
fprintf(stderr,"[E::%s] the index file exists. Please use '-f' to overwrite.\n", __func__);
return 1;
}
}
}
if (ftype == FT_BCF_GZ)
{
if ( bcf_index_build(fname, min_shift) != 0 )
{
fprintf(stderr,"[E::%s] bcf_index_build failed for %s\n", __func__, fname);
return 1;
}
}
else
{
if ( tbx_index_build(fname, min_shift, &tbx_conf_vcf) != 0 )
{
fprintf(stderr,"[E::%s] tbx_index_build failed for %s\n", __func__, fname);
return 1;
}
}
return 0;
}
int main(int argc, char *argv[])
{
int c, detect = 1, min_shift = 0, is_force = 0, list_chroms = 0, do_csi = 0;
tbx_conf_t conf = tbx_conf_gff;
char *reheader = NULL;
args_t args;
memset(&args,0,sizeof(args_t));
static const struct option loptions[] =
{
{"help", no_argument, NULL, 2},
{"regions", required_argument, NULL, 'R'},
{"targets", required_argument, NULL, 'T'},
{"csi", no_argument, NULL, 'C'},
{"zero-based", no_argument, NULL, '0'},
{"print-header", no_argument, NULL, 'h'},
{"only-header", no_argument, NULL, 'H'},
{"begin", required_argument, NULL, 'b'},
{"comment", required_argument, NULL, 'c'},
{"end", required_argument, NULL, 'e'},
{"force", no_argument, NULL, 'f'},
{"preset", required_argument, NULL, 'p'},
{"sequence", required_argument, NULL, 's'},
{"skip-lines", required_argument, NULL, 'S'},
{"list-chroms", no_argument, NULL, 'l'},
{"reheader", required_argument, NULL, 'r'},
{"version", no_argument, NULL, 1},
{NULL, 0, NULL, 0}
};
char *tmp;
while ((c = getopt_long(argc, argv, "hH?0b:c:e:fm:p:s:S:lr:CR:T:", loptions,NULL)) >= 0)
{
switch (c)
{
case 'R': args.regions_fname = optarg; break;
case 'T': args.targets_fname = optarg; break;
case 'C': do_csi = 1; break;
case 'r': reheader = optarg; break;
case 'h': args.print_header = 1; break;
case 'H': args.print_header = 1; args.header_only = 1; break;
case 'l': list_chroms = 1; break;
case '0': conf.preset |= TBX_UCSC; detect = 0; break;
case 'b':
conf.bc = strtol(optarg,&tmp,10);
if ( *tmp ) error("Could not parse argument: -b %s\n", optarg);
detect = 0;
break;
case 'e':
conf.ec = strtol(optarg,&tmp,10);
if ( *tmp ) error("Could not parse argument: -e %s\n", optarg);
detect = 0;
break;
case 'c': conf.meta_char = *optarg; detect = 0; break;
case 'f': is_force = 1; break;
case 'm':
min_shift = strtol(optarg,&tmp,10);
if ( *tmp ) error("Could not parse argument: -m %s\n", optarg);
break;
case 'p':
detect = 0;
if (strcmp(optarg, "gff") == 0) conf = tbx_conf_gff;
else if (strcmp(optarg, "bed") == 0) conf = tbx_conf_bed;
else if (strcmp(optarg, "sam") == 0) conf = tbx_conf_sam;
else if (strcmp(optarg, "vcf") == 0) conf = tbx_conf_vcf;
else if (strcmp(optarg, "bcf") == 0) detect = 1; // bcf is autodetected, preset is not needed
else if (strcmp(optarg, "bam") == 0) detect = 1; // same as bcf
else error("The preset string not recognised: '%s'\n", optarg);
break;
case 's':
conf.sc = strtol(optarg,&tmp,10);
if ( *tmp ) error("Could not parse argument: -s %s\n", optarg);
detect = 0;
break;
case 'S':
conf.line_skip = strtol(optarg,&tmp,10);
if ( *tmp ) error("Could not parse argument: -S %s\n", optarg);
detect = 0;
break;
case 1:
printf(
"tabix (htslib) %s\n"
"Copyright (C) 2017 Genome Research Ltd.\n", hts_version());
return EXIT_SUCCESS;
default: return usage();
}
}
if ( optind==argc ) return usage();
if ( list_chroms )
return query_chroms(argv[optind]);
if ( argc > optind+1 || args.header_only || args.regions_fname || args.targets_fname )
{
int nregs = 0;
char **regs = NULL;
if ( !args.header_only )
regs = parse_regions(args.regions_fname, argv+optind+1, argc-optind-1, &nregs);
return query_regions(&args, argv[optind], regs, nregs);
}
char *fname = argv[optind];
int ftype = file_type(fname);
if ( detect ) // no preset given
{
if ( ftype==IS_GFF ) conf = tbx_conf_gff;
else if ( ftype==IS_BED ) conf = tbx_conf_bed;
else if ( ftype==IS_SAM ) conf = tbx_conf_sam;
else if ( ftype==IS_VCF )
{
conf = tbx_conf_vcf;
if ( !min_shift && do_csi ) min_shift = 14;
}
else if ( ftype==IS_BCF )
{
if ( !min_shift ) min_shift = 14;
}
else if ( ftype==IS_BAM )
{
if ( !min_shift ) min_shift = 14;
}
}
if ( do_csi )
{
if ( !min_shift ) min_shift = 14;
min_shift *= do_csi; // positive for CSIv2, negative for CSIv1
}
if ( min_shift!=0 && !do_csi ) do_csi = 1;
if ( reheader )
return reheader_file(fname, reheader, ftype, &conf);
char *suffix = ".tbi";
if ( do_csi ) suffix = ".csi";
else if ( ftype==IS_BAM ) suffix = ".bai";
else if ( ftype==IS_CRAM ) suffix = ".crai";
char *idx_fname = calloc(strlen(fname) + 5, 1);
strcat(strcpy(idx_fname, fname), suffix);
struct stat stat_tbi, stat_file;
if ( !is_force && stat(idx_fname, &stat_tbi)==0 )
{
// Before complaining about existing index, check if the VCF file isn't
// newer. This is a common source of errors, people tend not to notice
// that tabix failed
stat(fname, &stat_file);
if ( stat_file.st_mtime <= stat_tbi.st_mtime )
error("[tabix] the index file exists. Please use '-f' to overwrite.\n");
}
free(idx_fname);
if ( ftype==IS_CRAM )
{
if ( bam_index_build(fname, min_shift)!=0 ) error("bam_index_build failed: %s\n", fname);
return 0;
}
else if ( do_csi )
{
if ( ftype==IS_BCF )
{
if ( bcf_index_build(fname, min_shift)!=0 ) error("bcf_index_build failed: %s\n", fname);
return 0;
}
if ( ftype==IS_BAM )
{
if ( bam_index_build(fname, min_shift)!=0 ) error("bam_index_build failed: %s\n", fname);
return 0;
}
if ( tbx_index_build(fname, min_shift, &conf)!=0 ) error("tbx_index_build failed: %s\n", fname);
return 0;
}
else // TBI index
{
if ( tbx_index_build(fname, min_shift, &conf) ) error("tbx_index_build failed: %s\n", fname);
return 0;
}
return 0;
}
int main_tabix(int argc, char *argv[])
{
int c, min_shift = -1, is_force = 0, is_all = 0;
tbx_conf_t conf = tbx_conf_gff, *conf_ptr = NULL;
while ((c = getopt(argc, argv, "0fap:s:b:e:S:c:m:")) >= 0)
if (c == '0') conf.preset |= TBX_UCSC;
else if (c == 'f') is_force = 1;
else if (c == 'a') is_all = 1;
else if (c == 'm') min_shift = atoi(optarg);
else if (c == 's') conf.sc = atoi(optarg);
else if (c == 'b') conf.bc = atoi(optarg);
else if (c == 'e') conf.ec = atoi(optarg);
else if (c == 'c') conf.meta_char = *optarg;
else if (c == 'S') conf.line_skip = atoi(optarg);
else if (c == 'p') {
if (strcmp(optarg, "gff") == 0) conf_ptr = &tbx_conf_gff;
else if (strcmp(optarg, "bed") == 0) conf_ptr = &tbx_conf_bed;
else if (strcmp(optarg, "sam") == 0) conf_ptr = &tbx_conf_sam;
else if (strcmp(optarg, "vcf") == 0) conf_ptr = &tbx_conf_vcf;
else {
fprintf(stderr, "The type '%s' not recognised\n", optarg);
return 1;
}
}
if (optind == argc) {
fprintf(stderr, "\nUsage: bcftools tabix [options] <in.gz> [reg1 [...]]\n\n");
fprintf(stderr, "Options: -p STR preset: gff, bed, sam or vcf [gff]\n");
fprintf(stderr, " -s INT column number for sequence names (suppressed by -p) [1]\n");
fprintf(stderr, " -b INT column number for region start [4]\n");
fprintf(stderr, " -e INT column number for region end (if no end, set INT to -b) [5]\n");
fprintf(stderr, " -0 specify coordinates are zero-based\n");
fprintf(stderr, " -S INT skip first INT lines [0]\n");
fprintf(stderr, " -c CHAR skip lines starting with CHAR [null]\n");
fprintf(stderr, " -a print all records\n");
fprintf(stderr, " -f force to overwrite existing index\n");
fprintf(stderr, " -m INT set the minimal interval size to 1<<INT; 0 for the old tabix index [0]\n");
fprintf(stderr, "\n");
return 1;
}
if (is_all) { // read without random access
kstring_t s;
BGZF *fp;
s.l = s.m = 0; s.s = 0;
fp = bgzf_open(argv[optind], "r");
while (bgzf_getline(fp, '\n', &s) >= 0) puts(s.s);
bgzf_close(fp);
free(s.s);
} else if (optind + 2 > argc) { // create index
if ( !conf_ptr )
{
// auto-detect file type by file name
int l = strlen(argv[optind]);
int strcasecmp(const char *s1, const char *s2);
if (l>=7 && strcasecmp(argv[optind]+l-7, ".gff.gz") == 0) conf_ptr = &tbx_conf_gff;
else if (l>=7 && strcasecmp(argv[optind]+l-7, ".bed.gz") == 0) conf_ptr = &tbx_conf_bed;
else if (l>=7 && strcasecmp(argv[optind]+l-7, ".sam.gz") == 0) conf_ptr = &tbx_conf_sam;
else if (l>=7 && strcasecmp(argv[optind]+l-7, ".vcf.gz") == 0) conf_ptr = &tbx_conf_vcf;
}
if ( conf_ptr ) conf = *conf_ptr;
if (!is_force) {
char *fn;
FILE *fp;
fn = (char*)alloca(strlen(argv[optind]) + 5);
strcat(strcpy(fn, argv[optind]), min_shift <= 0? ".tbi" : ".csi");
if ((fp = fopen(fn, "rb")) != 0) {
fclose(fp);
fprintf(stderr, "[E::%s] the index file exists; use option '-f' to overwrite\n", __func__);
return 1;
}
}
if ( tbx_index_build(argv[optind], min_shift, &conf) )
{
fprintf(stderr,"tbx_index_build failed: Is the file bgzip-compressed? Was wrong -p [type] option used?\n");
return 1;
}
} else { // read with random access
tbx_t *tbx;
BGZF *fp;
kstring_t s;
int i;
if ((tbx = tbx_index_load(argv[optind])) == 0) return 1;
if ((fp = bgzf_open(argv[optind], "r")) == 0) return 1;
s.s = 0; s.l = s.m = 0;
for (i = optind + 1; i < argc; ++i) {
hts_itr_t *itr;
if ((itr = tbx_itr_querys(tbx, argv[i])) == 0) continue;
while (tbx_bgzf_itr_next(fp, tbx, itr, &s) >= 0) puts(s.s);
tbx_itr_destroy(itr);
}
free(s.s);
bgzf_close(fp);
tbx_destroy(tbx);
}
return 0;
}
