faidx_t *fai_read(FILE *fp)
{
faidx_t *fai;
char *buf, *p;
int len, line_len, line_blen;
#ifdef _WIN32
long offset;
#else
long long offset;
#endif
fai = (faidx_t*)calloc(1, sizeof(faidx_t));
fai->hash = kh_init(s);
buf = (char*)calloc(0x10000, 1);
while (!feof(fp) && fgets(buf, 0x10000, fp)) {
for (p = buf; *p && isgraph(*p); ++p);
*p = 0; ++p;
#ifdef _WIN32
sscanf(p, "%d%ld%d%d", &len, &offset, &line_blen, &line_len);
#else
sscanf(p, "%d%lld%d%d", &len, &offset, &line_blen, &line_len);
#endif
fai_insert_index(fai, buf, len, line_len, line_blen, offset);
}
free(buf);
return fai;
}
faidx_t *fai_build_core(RAZF *rz)
{
char c, *name;
int l_name, m_name, ret;
int len, line_len, line_blen, state;
int l1, l2;
faidx_t *idx;
uint64_t offset;
idx = (faidx_t*)calloc(1, sizeof(faidx_t));
idx->hash = kh_init(s);
name = 0; l_name = m_name = 0;
len = line_len = line_blen = -1; state = 0; l1 = l2 = -1; offset = 0;
while (razf_read(rz, &c, 1)) {
if (c == '\n') { // an empty line
if (state == 1) {
offset = razf_tell(rz);
continue;
} else if ((state == 0 && len < 0) || state == 2) continue;
}
if (c == '>') { // fasta header
if (len >= 0)
fai_insert_index(idx, name, len, line_len, line_blen, offset);
l_name = 0;
while ((ret = razf_read(rz, &c, 1)) != 0 && !isspace(c)) {
if (m_name < l_name + 2) {
m_name = l_name + 2;
kroundup32(m_name);
name = (char*)realloc(name, m_name);
}
name[l_name++] = c;
}
name[l_name] = '\0';
if (ret == 0) {
fprintf(stderr, "[fai_build_core] the last entry has no sequence\n");
free(name); fai_destroy(idx);
return 0;
}
if (c != '\n') while (razf_read(rz, &c, 1) && c != '\n');
state = 1; len = 0;
offset = razf_tell(rz);
} else {
if (state == 3) {
fprintf(stderr, "[fai_build_core] inlined empty line is not allowed in sequence '%s'.\n", name);
free(name); fai_destroy(idx);
return 0;
}
if (state == 2) state = 3;
l1 = l2 = 0;
do {
++l1;
if (isgraph(c)) ++l2;
} while ((ret = razf_read(rz, &c, 1)) && c != '\n');
if (state == 3 && l2) {
fprintf(stderr, "[fai_build_core] different line length in sequence '%s'.\n", name);
free(name); fai_destroy(idx);
return 0;
}
++l1; len += l2;
if (l2 >= 0x10000) {
fprintf(stderr, "[fai_build_core] line length exceeds 65535 in sequence '%s'.\n", name);
free(name); fai_destroy(idx);
return 0;
}
if (state == 1) line_len = l1, line_blen = l2, state = 0;
else if (state == 0) {
if (l1 != line_len || l2 != line_blen) state = 2;
}
}
}
fai_insert_index(idx, name, len, line_len, line_blen, offset);
free(name);
return idx;
}
faidx_t *fai_build_core(BGZF *bgzf)
{
char *name;
int c;
int l_name, m_name;
int line_len, line_blen, state;
int l1, l2;
faidx_t *idx;
uint64_t offset;
int64_t len;
idx = (faidx_t*)calloc(1, sizeof(faidx_t));
name = (char*)calloc(1, sizeof(char)); /* at least 1 byte, for '\0' */
idx->hash = kh_init(s);
l_name = m_name = 0;
len = line_len = line_blen = -1; state = 0; l1 = l2 = -1; offset = 0;
while ( (c=bgzf_getc(bgzf))>=0 ) {
if (c == '\n') { // an empty line
if (state == 1) {
offset = bgzf_utell(bgzf);
continue;
} else if ((state == 0 && len < 0) || state == 2) continue;
}
if (c == '>') { // fasta header
if (len >= 0)
fai_insert_index(idx, name, len, line_len, line_blen, offset);
l_name = 0;
while ( (c=bgzf_getc(bgzf))>=0 && !isspace(c)) {
if (m_name < l_name + 2) {
m_name = l_name + 2;
kroundup32(m_name);
name = (char*)realloc(name, m_name);
}
name[l_name++] = c;
}
name[l_name] = '\0';
if ( c<0 ) {
fprintf(stderr, "[fai_build_core] the last entry has no sequence\n");
free(name); fai_destroy(idx);
return 0;
}
if (c != '\n') while ( (c=bgzf_getc(bgzf))>=0 && c != '\n');
state = 1; len = 0;
offset = bgzf_utell(bgzf);
} else {
if (state == 3) {
fprintf(stderr, "[fai_build_core] inlined empty line is not allowed in sequence '%s'.\n", name);
free(name); fai_destroy(idx);
return 0;
}
if (state == 2) state = 3;
l1 = l2 = 0;
do {
++l1;
if (isgraph(c)) ++l2;
} while ( (c=bgzf_getc(bgzf))>=0 && c != '\n');
if (state == 3 && l2) {
fprintf(stderr, "[fai_build_core] different line length in sequence '%s'.\n", name);
free(name); fai_destroy(idx);
return 0;
}
++l1; len += l2;
if (state == 1) line_len = l1, line_blen = l2, state = 0;
else if (state == 0) {
if (l1 != line_len || l2 != line_blen) state = 2;
}
}
}
if ( name )
fai_insert_index(idx, name, len, line_len, line_blen, offset);
else
{
free(idx);
return NULL;
}
free(name);
return idx;
}