filter_t *region_exact_filter_new(char *region_descriptor, int use_region_file, char *type,
const char *url, const char *species, const char *version) {
assert(region_descriptor);
assert(url);
assert(species);
assert(version);
filter_t *filter = (filter_t*) malloc (sizeof(filter_t));
filter->type = REGION;
filter->filter_func = region_filter;
filter->free_func = region_filter_free;
filter->priority = 2;
region_filter_args *filter_args = (region_filter_args*) malloc (sizeof(region_filter_args));
if (use_region_file) {
snprintf(filter->name, 11, "RegionFile");
snprintf(filter->description, 64, "Regions read from '%s'", region_descriptor);
if (ends_with(region_descriptor, ".gff")) {
filter_args->regions = parse_regions_from_gff_file(region_descriptor, url, species, version);
} else if (ends_with(region_descriptor, ".bed")) {
filter_args->regions = parse_regions_from_bed_file(region_descriptor, url, species, version);
} else {
LOG_FATAL_F("Region file %s format not supported! Please use BED or GFF formats\n", region_descriptor);
}
} else {
snprintf(filter->name, 11, "RegionList");
snprintf(filter->description, 64, "Regions (could be more) %s", region_descriptor);
filter_args->regions = parse_regions(region_descriptor, 1, url, species, version);
}
filter_args->type = type;
filter->args = filter_args;
return filter;
}
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(int argc, char* argv[])
{
try
{
if (argc < 13 || argc % 2 != 1)
{
std::cerr << "usage: " << argv[0] << " number_of_threads region_file gff_or_bed_format extension bam_file bam_file_key hdf5_file "
<< "log_file write_warnings_to_stderr strand chr1 length1 ...\n"
<< "\ne.g. " << argv[0] << " /grail/annotations/HG19_SUM159_BRD4_-0_+0.gff gff"
<< "\n /ifs/labs/bradner/bam/hg18/mm1s/04032013_D1L57ACXX_4.TTAGGC.hg18.bwt.sorted.bam 137 counts.hdf5 "
<< "\n output/log.txt 1 _ chr1 247249719 chr2 242951149 chr3 199501827\n"
<< "\nstrand value of _ means use strand that is specified in region file (and use . if strand not specified in region file)."
<< "\nnumber of threads <= 0 means use a number of threads equal to the number of logical cpus."
<< "\nnote that this application is intended to be run from bamliquidator_batch.py -- see"
<< "\nhttps://github.com/BradnerLab/pipeline/wiki for more information"
<< std::endl;
return 1;
}
const int number_of_threads = boost::lexical_cast<int>(argv[1]);
const std::string region_file_path = argv[2];
const std::string region_format = argv[3];
const unsigned int extension = boost::lexical_cast<unsigned int>(argv[4]);
const std::string bam_file_path = argv[5];
const unsigned int bam_file_key = boost::lexical_cast<unsigned int>(argv[6]);
const std::string hdf5_file_path = argv[7];
const std::string log_file_path = argv[8];
const bool write_warnings_to_stderr = boost::lexical_cast<bool>(argv[9]);
const char strand = boost::lexical_cast<char>(argv[10]);
const std::vector<std::pair<std::string, size_t>> chromosome_lengths = extract_chromosome_lengths(argc, argv, 11);
tbb::task_scheduler_init init( number_of_threads <= 0
? tbb::task_scheduler_init::automatic
: number_of_threads);
Logger::configure(log_file_path, write_warnings_to_stderr);
hid_t h5file = H5Fopen(hdf5_file_path.c_str(), H5F_ACC_RDWR, H5P_DEFAULT);
if (h5file < 0)
{
Logger::error() << "Failed to open H5 file " << hdf5_file_path;
return 3;
}
#ifdef time_region_parsing
boost::timer::cpu_timer timer;
#endif
std::map<std::string, size_t> chromosome_to_length;
for (auto& chr_length : chromosome_lengths)
{
chromosome_to_length[chr_length.first] = chr_length.second;
}
std::vector<Region> regions = parse_regions(region_file_path,
region_format,
bam_file_key,
chromosome_to_length,
strand);
#ifdef time_region_parsing
timer.stop();
std::cout << "parsing regions took" << timer.format() << std::endl;
#endif
if (regions.size() == 0)
{
Logger::warn() << "No valid regions detected in " << region_file_path;
return 0;
}
liquidate_and_write(h5file, regions, extension, bam_file_path);
H5Fclose(h5file);
return 0;
}
catch(const std::exception& e)
{
Logger::error() << "Unhandled exception: " << e.what();
return 4;
}
}
