int faidx_main(int argc, char *argv[])
{
if (argc == 1) {
fprintf(stderr, "Usage: faidx <in.fasta> [<reg> [...]]\n");
return 1;
} else {
if (argc == 2) fai_build(argv[1]);
else {
int i, j, k, l;
char *s;
faidx_t *fai;
fai = fai_load(argv[1]);
if (fai == 0) return 1;
for (i = 2; i != argc; ++i) {
printf(">%s\n", argv[i]);
s = fai_fetch(fai, argv[i], &l);
for (j = 0; j < l; j += 60) {
for (k = 0; k < 60 && k < l - j; ++k)
putchar(s[j + k]);
putchar('\n');
}
free(s);
}
fai_destroy(fai);
}
}
return 0;
}
faidx_t *fai_load(const char *fn)
{
char *str;
FILE *fp;
faidx_t *fai;
str = (char*)calloc(strlen(fn) + 5, 1);
sprintf(str, "%s.fai", fn);
#ifdef _USE_KNETFILE
if (strstr(fn, "ftp://") == fn || strstr(fn, "http://") == fn)
{
fp = download_and_open(str);
if ( !fp )
{
fprintf(stderr, "[fai_load] failed to open remote FASTA index %s\n", str);
free(str);
return 0;
}
}
else
#endif
fp = fopen(str, "rb");
if (fp == 0) {
fprintf(stderr, "[fai_load] build FASTA index.\n");
fai_build(fn);
fp = fopen(str, "rb");
if (fp == 0) {
fprintf(stderr, "[fai_load] fail to open FASTA index.\n");
free(str);
return 0;
}
}
fai = fai_read(fp);
fclose(fp);
fai->bgzf = bgzf_open(fn, "rb");
free(str);
if (fai->bgzf == 0) {
fprintf(stderr, "[fai_load] fail to open FASTA file.\n");
return 0;
}
if ( fai->bgzf->is_compressed==1 )
{
if ( bgzf_index_load(fai->bgzf, fn, ".gzi") < 0 )
{
fprintf(stderr, "[fai_load] failed to load .gzi index: %s[.gzi]\n", fn);
fai_destroy(fai);
return NULL;
}
}
return fai;
}
void ReadDB::build(const std::string& input_reads_filename)
{
// generate output filename
m_indexed_reads_filename = input_reads_filename + GZIPPED_READS_SUFFIX;
// Populate database with read names and convert the fastq
// input into fasta for faidx
import_reads(input_reads_filename, m_indexed_reads_filename);
// build faidx
int ret = fai_build(m_indexed_reads_filename.c_str());
if(ret != 0) {
fprintf(stderr, "Error running faidx_build on %s\n", m_indexed_reads_filename.c_str());
exit(EXIT_FAILURE);
}
m_fai = NULL;
}
faidx_t *fai_load(const char *fn)
{
char *str;
FILE *fp;
faidx_t *fai;
str = (char*)calloc(strlen(fn) + 5, 1);
sprintf(str, "%s.fai", fn);
#ifdef _USE_KNETFILE
if (strstr(fn, "ftp://") == fn || strstr(fn, "http://") == fn)
{
fp = download_and_open(str);
if ( !fp )
{
fprintf(stderr, "[fai_load] failed to open remote FASTA index %s\n", str);
free(str);
return 0;
}
}
else
#endif
fp = fopen(str, "rb");
if (fp == 0) {
fprintf(stderr, "[fai_load] build FASTA index.\n");
fai_build(fn);
fp = fopen(str, "rb");
if (fp == 0) {
fprintf(stderr, "[fai_load] fail to open FASTA index.\n");
free(str);
return 0;
}
}
fai = fai_read(fp);
fclose(fp);
fai->rz = razf_open(fn, "rb");
free(str);
if (fai->rz == 0) {
fprintf(stderr, "[fai_load] fail to open FASTA file.\n");
return 0;
}
return fai;
}
char *samfaipath(const char *fn_ref)
{
char *fn_list = 0;
if (fn_ref == 0) return 0;
fn_list = calloc(strlen(fn_ref) + 5, 1);
strcat(strcpy(fn_list, fn_ref), ".fai");
if (access(fn_list, R_OK) == -1) { // fn_list is unreadable
if (access(fn_ref, R_OK) == -1) {
fprintf(stderr, "[samfaipath] fail to read file %s.\n", fn_ref);
} else {
if (bam_verbose >= 3) fprintf(stderr, "[samfaipath] build FASTA index...\n");
if (fai_build(fn_ref) == -1) {
fprintf(stderr, "[samfaipath] fail to build FASTA index.\n");
free(fn_list); fn_list = 0;
}
}
}
return fn_list;
}
static void faidx1(const char *filename)
{
int n, n_exp = 0, n_fq_exp = 0;
char tmpfilename[FILENAME_MAX], line[500];
FILE *fin, *fout;
faidx_t *fai;
fin = fopen(filename, "rb");
if (fin == NULL) fail("can't open %s\n", filename);
sprintf(tmpfilename, "%s.tmp", filename);
fout = fopen(tmpfilename, "wb");
if (fout == NULL) fail("can't create temporary %s\n", tmpfilename);
while (fgets(line, sizeof line, fin)) {
if (line[0] == '>') n_exp++;
if (line[0] == '+' && line[1] == '\n') n_fq_exp++;
fputs(line, fout);
}
fclose(fin);
fclose(fout);
if (n_exp == 0 && n_fq_exp != 0) {
// probably a fastq file
n_exp = n_fq_exp;
}
if (fai_build(tmpfilename) < 0) fail("can't index %s", tmpfilename);
fai = fai_load(tmpfilename);
if (fai == NULL) { fail("can't load faidx file %s", tmpfilename); return; }
n = faidx_fetch_nseq(fai);
if (n != n_exp)
fail("%s: faidx_fetch_nseq returned %d, expected %d", filename, n, n_exp);
n = faidx_nseq(fai);
if (n != n_exp)
fail("%s: faidx_nseq returned %d, expected %d", filename, n, n_exp);
fai_destroy(fai);
}
explicit MMappedFastaFile(std::string const & _filename) :
filename(_filename)
{
struct stat st;
stat(_filename.c_str(), &st);
filesize = (size_t) st.st_size;
fd = open(_filename.c_str(), O_RDONLY, 0);
assert(fd != -1);
#if __APPLE__
base = (uint8_t *) mmap(NULL, filesize, PROT_READ, MAP_PRIVATE, fd, 0);
#else
base = (uint8_t *) mmap(NULL, filesize, PROT_READ, MAP_PRIVATE | MAP_POPULATE, fd, 0);
#endif
if(base == MAP_FAILED)
{
const int err = errno;
error("Cannot mmap %s (errno=%i / %s) -- do you have enough memory available?",
_filename.c_str(), err, strerror(err));
}
auto fai_fp = fopen((filename + ".fai").c_str(), "r");
if(!fai_fp)
{
assert(fai_build(filename.c_str()) == 0);
}
else
{
fclose(fai_fp);
}
std::ifstream index_in(filename + ".fai");
while(index_in.good())
{
std::string line;
std::getline(index_in, line);
std::vector<std::string> parts;
stringutil::split(line, parts, "\n\t", false);
if(parts.size() == 5)
{
const std::string contig = parts[0];
index_entry ientry;
sscanf(parts[1].c_str(), "%zu", &ientry.length);
sscanf(parts[2].c_str(), "%zu", &ientry.start_offset);
sscanf(parts[3].c_str(), "%zu", &ientry.chars_per_line);
sscanf(parts[4].c_str(), "%zu", &ientry.bytes_per_line);
fai[contig] = ientry;
// determine non-N length; we don't do this because it's pretty slow.
// size_t offset = ientry.start_offset;
// const size_t line_number = (ientry.length - 1) / ientry.chars_per_line;
// size_t offset_in_line = (ientry.length - 1) % ientry.chars_per_line;
// size_t offset_end = ientry.start_offset + line_number*ientry.bytes_per_line + offset_in_line + 1;
// size_t ns = 0;
//
// while(offset < offset_end)
// {
// if(std::tolower(base[offset - 1]) == 'n')
// {
// ++ns;
// }
// ++offset;
// }
// fai[contig].non_n_length = fai[contig].length - ns;
// length trimming off N's at start and end
size_t pos = 0;
size_t ns_at_start = 0;
bool done = false;
// start of contig
while(pos < ientry.length && !done)
{
const std::string s = get(contig, pos, 10000);
for(size_t j = 0; j < s.size(); ++j)
{
if(std::tolower(s.at(j)) == 'n')
{
++ns_at_start;
}
else
{
done = true;
break;
}
}
pos += s.size();
}
size_t ns_at_end = 0;
// check if we had all Ns
if(ns_at_start < ientry.length)
{
const size_t line_number = (ientry.length - 1) / ientry.chars_per_line;
size_t offset_in_line = (ientry.length - 1) % ientry.chars_per_line;
size_t offset = ientry.start_offset + line_number*ientry.bytes_per_line + offset_in_line + 1;
// end of contig
while(offset > ientry.start_offset)
{
if (base[offset - 1] == '\n' || base[offset - 1] == '\r')
{
// skip newlines
}
else if(std::tolower(base[offset - 1]) == 'n')
{
++ns_at_end;
}
else
{
break;
}
--offset;
}
}
fai[contig].non_n_length = fai[contig].length - (ns_at_start + ns_at_end);
}
else if(parts.size() > 0)
{
error("invalid fai line %s in %s", line.c_str(), (filename + ".fai").c_str());
}
}
}
