From d1fcb739b58e050328dcfc47c3af235bc0e41455 Mon Sep 17 00:00:00 2001 From: Charles Plessy Date: Thu, 19 Aug 2010 15:49:40 +0900 Subject: [PATCH] Imported Upstream version 0.1.6 --- ChangeLog | 253 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ NEWS | 32 ++++++- index.c | 159 ++++++++++++++++++++++------------ main.c | 45 +++++++--- tabix.h | 46 ++++++++-- 5 files changed, 457 insertions(+), 78 deletions(-) create mode 100644 ChangeLog diff --git a/ChangeLog b/ChangeLog new file mode 100644 index 0000000..19d4f3e --- /dev/null +++ b/ChangeLog @@ -0,0 +1,253 @@ +------------------------------------------------------------------------ +r562 | lh3lh3 | 2010-05-09 19:43:56 -0400 (Sun, 09 May 2010) | 2 lines +Changed paths: + M /trunk/tabix/index.c + +Fixed a bug + +------------------------------------------------------------------------ +r560 | lh3lh3 | 2010-05-05 10:59:09 -0400 (Wed, 05 May 2010) | 3 lines +Changed paths: + A /trunk/tabix/NEWS + M /trunk/tabix/TabixReader.java + M /trunk/tabix/index.c + M /trunk/tabix/main.c + M /trunk/tabix/tabix.1 + M /trunk/tabix/tabix.h + + * Release tabix-0.1.5 (r560) + * Improve seeking efficiency. Index file needs to be rebuilt. + +------------------------------------------------------------------------ +r559 | lh3lh3 | 2010-05-04 23:11:42 -0400 (Tue, 04 May 2010) | 2 lines +Changed paths: + M /trunk/tabix/main.c + +Release tabix-0.1.4 (r559) + +------------------------------------------------------------------------ +r558 | lh3lh3 | 2010-05-01 12:48:01 -0400 (Sat, 01 May 2010) | 2 lines +Changed paths: + M /trunk/tabix/TabixReader.java + +implement SAM/VCF support; NOT tested yet + +------------------------------------------------------------------------ +r557 | lh3lh3 | 2010-05-01 00:42:34 -0400 (Sat, 01 May 2010) | 2 lines +Changed paths: + A /trunk/tabix/TabixReader.java + +The Java implementation of tabix. + +------------------------------------------------------------------------ +r556 | lh3lh3 | 2010-04-30 22:34:07 -0400 (Fri, 30 Apr 2010) | 4 lines +Changed paths: + M /trunk/tabix/index.c + M /trunk/tabix/knetfile.c + M /trunk/tabix/main.c + + * tabix-0.1.3-3 (r556) + * fixed a small memory leak in knetfile + * fixed a minor bug for remote downloading + +------------------------------------------------------------------------ +r555 | lh3lh3 | 2010-04-30 22:15:12 -0400 (Fri, 30 Apr 2010) | 4 lines +Changed paths: + M /trunk/tabix/Makefile + M /trunk/tabix/index.c + M /trunk/tabix/main.c + + * tabix-0.1.3-2 (r555) + * do not overwrite index file by default + * a little code cleanup + +------------------------------------------------------------------------ +r554 | lh3lh3 | 2010-04-30 21:44:31 -0400 (Fri, 30 Apr 2010) | 2 lines +Changed paths: + M /trunk/tabix/index.c + +fixed a potential bug for UCSC-like coordinate + +------------------------------------------------------------------------ +r553 | lh3lh3 | 2010-04-28 17:43:41 -0400 (Wed, 28 Apr 2010) | 2 lines +Changed paths: + M /trunk/tabix/tabix.tex + +minor clarification to the format spec + +------------------------------------------------------------------------ +r552 | lh3lh3 | 2010-04-28 16:33:07 -0400 (Wed, 28 Apr 2010) | 3 lines +Changed paths: + M /trunk/tabix/Makefile + M /trunk/tabix/bgzip.c + A /trunk/tabix/tabix.tex + + * added the format specification + * fixed a typo in bgzip + +------------------------------------------------------------------------ +r550 | petulda | 2010-04-22 11:03:24 -0400 (Thu, 22 Apr 2010) | 1 line +Changed paths: + M /trunk/tabix/bgzip.c + +The behaviour changed slightly to mimic gzip. Detect if std descriptors are connected to the terminal. +------------------------------------------------------------------------ +r549 | petulda | 2010-04-22 09:46:10 -0400 (Thu, 22 Apr 2010) | 1 line +Changed paths: + M /trunk/tabix/bgzip.c + +Fix in src/dst file detection and slight change of behaviour +------------------------------------------------------------------------ +r548 | petulda | 2010-04-19 04:39:46 -0400 (Mon, 19 Apr 2010) | 1 line +Changed paths: + M /trunk/tabix/index.c + +Close file descriptor in ti_list_chromosomes +------------------------------------------------------------------------ +r547 | petulda | 2010-04-16 09:27:11 -0400 (Fri, 16 Apr 2010) | 1 line +Changed paths: + M /trunk/tabix/index.c + M /trunk/tabix/main.c + M /trunk/tabix/tabix.h + +Added the -l option for listing chromosomes +------------------------------------------------------------------------ +r544 | lh3lh3 | 2010-03-29 10:58:48 -0400 (Mon, 29 Mar 2010) | 2 lines +Changed paths: + M /trunk/tabix/main.c + +removed a line of debugging code + +------------------------------------------------------------------------ +r543 | lh3lh3 | 2010-03-19 12:29:16 -0400 (Fri, 19 Mar 2010) | 3 lines +Changed paths: + M /trunk/tabix/index.c + M /trunk/tabix/main.c + M /trunk/tabix/tabix.1 + + * tabix-0.1.3 (r543) + * fixed another off-by-one bug + +------------------------------------------------------------------------ +r542 | lh3lh3 | 2010-03-16 22:35:52 -0400 (Tue, 16 Mar 2010) | 2 lines +Changed paths: + M /trunk/tabix/index.c + M /trunk/tabix/main.c + M /trunk/tabix/tabix.1 + +Release tabix-0.1.1 + +------------------------------------------------------------------------ +r506 | lh3lh3 | 2009-11-02 23:20:12 -0500 (Mon, 02 Nov 2009) | 2 lines +Changed paths: + M /trunk/tabix/main.c + +Release tabix-0.1.0 + +------------------------------------------------------------------------ +r505 | lh3lh3 | 2009-11-02 23:15:49 -0500 (Mon, 02 Nov 2009) | 2 lines +Changed paths: + A /trunk/tabix/tabix.1 + +documentation + +------------------------------------------------------------------------ +r504 | lh3lh3 | 2009-11-02 11:08:18 -0500 (Mon, 02 Nov 2009) | 5 lines +Changed paths: + M /trunk/tabix/Makefile + M /trunk/tabix/bgzip.c + M /trunk/tabix/index.c + M /trunk/tabix/main.c + M /trunk/tabix/tabix.h + + * tabix-0.0.0-5 (r504) + * fixed a critical bug in fetching data (a typo in fact) + * support SAM (tested on ex1.sam) and VCF (not tested) + * improve the command-line interface + +------------------------------------------------------------------------ +r503 | lh3lh3 | 2009-11-02 10:04:43 -0500 (Mon, 02 Nov 2009) | 3 lines +Changed paths: + M /trunk/tabix/Makefile + M /trunk/tabix/index.c + M /trunk/tabix/main.c + + * tabix-0.0.0-4 (r503) + * index files are bgzf compressed + +------------------------------------------------------------------------ +r502 | lh3lh3 | 2009-11-02 09:47:25 -0500 (Mon, 02 Nov 2009) | 4 lines +Changed paths: + M /trunk/tabix/index.c + M /trunk/tabix/main.c + M /trunk/tabix/tabix.h + + * tabix-0.0.0-3 (r502) + * support meta lines (not tested) + * I am going to make the index file in the BGZF format + +------------------------------------------------------------------------ +r501 | lh3lh3 | 2009-11-01 22:03:07 -0500 (Sun, 01 Nov 2009) | 3 lines +Changed paths: + M /trunk/tabix/Makefile + M /trunk/tabix/bgzf.h + M /trunk/tabix/index.c + M /trunk/tabix/main.c + + * tabix-0.0.0-2 (r501) + * accelerate ti_readline() + +------------------------------------------------------------------------ +r500 | lh3lh3 | 2009-11-01 20:49:52 -0500 (Sun, 01 Nov 2009) | 3 lines +Changed paths: + M /trunk/tabix/Makefile + M /trunk/tabix/bgzip.c + M /trunk/tabix/index.c + M /trunk/tabix/main.c + + * tabix-0.0.0-1 (r500) + * apparently working + +------------------------------------------------------------------------ +r499 | lh3lh3 | 2009-11-01 14:04:52 -0500 (Sun, 01 Nov 2009) | 2 lines +Changed paths: + D /trunk/tabix/parser.c + +obsolete file + +------------------------------------------------------------------------ +r498 | lh3lh3 | 2009-11-01 14:04:08 -0500 (Sun, 01 Nov 2009) | 2 lines +Changed paths: + M /trunk/tabix/bgzip.c + +bgzip is more like gzip in its command-line interface + +------------------------------------------------------------------------ +r497 | lh3lh3 | 2009-11-01 13:43:35 -0500 (Sun, 01 Nov 2009) | 2 lines +Changed paths: + A /trunk/tabix/Makefile + A /trunk/tabix/bam_endian.h + A /trunk/tabix/bgzf.c + A /trunk/tabix/bgzf.h + A /trunk/tabix/bgzip.c + A /trunk/tabix/index.c + A /trunk/tabix/khash.h + A /trunk/tabix/knetfile.c + A /trunk/tabix/knetfile.h + A /trunk/tabix/ksort.h + A /trunk/tabix/kstring.c + A /trunk/tabix/kstring.h + A /trunk/tabix/main.c + A /trunk/tabix/parser.c + A /trunk/tabix/tabix.h + +initial source code. It is BUGGY! + +------------------------------------------------------------------------ +r496 | lh3lh3 | 2009-11-01 13:42:39 -0500 (Sun, 01 Nov 2009) | 2 lines +Changed paths: + A /trunk/tabix + +A generic indexer for TAB-delimited genome position files + +------------------------------------------------------------------------ diff --git a/NEWS b/NEWS index aaa95e0..8cb7f0d 100644 --- a/NEWS +++ b/NEWS @@ -1,14 +1,38 @@ +Beta Release 0.1.6 (9 May, 2010) +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Notable changes: + + * Improved backward compatibility. Release 0.1.5 does not work with the + buggy index file generated by 0.1.2. + + * Fixed a bug in building linear index. The bug does not affect the + results, only affects efficiency in rare cases. + + * Reduced the number of seek calls given an index generated by old + version of tabix. + + * Added new APIs for retrieving data via an iterator. The old callback + APIs are not changed, although internally it uses iterator to + retrieve data. + +I am trying to freeze tabix. I just hope I am committing new bugs. + +(0.1.6: 9 May 2010, r563) + + + Beta Release 0.1.5 (5 May, 2010) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Notable changes: - * Clarify that tabix is released under MIT/X11. + * Clarified that tabix is released under MIT/X11. - * Improve the robustness of indexing and retrieval. + * Improved the robustness of indexing and retrieval. - * Reduce the number of seek calls when the specified region starts from - a 16kb block with no data. The index format is the same, but the + * Reduced the number of seek calls when the specified region starts + from a 16kb block with no data. The index format is the same, but the content is changed a little. (0.1.5: 5 May 2010, r560) diff --git a/index.c b/index.c index cfb2938..61d7d6d 100644 --- a/index.c +++ b/index.c @@ -42,6 +42,16 @@ struct __ti_index_t { ti_lidx_t *index2; }; +struct __ti_iter_t { + int from_first; // read from the first record; no random access + int tid, beg, end, n_off, i, finished; + uint64_t curr_off; + kstring_t str; + BGZF *fp; + const ti_index_t *idx; + pair64_t *off; +}; + typedef struct { int tid, beg, end, bin; } ti_intv_t; @@ -241,9 +251,13 @@ static inline void insert_offset2(ti_lidx_t *index2, int _beg, int _end, uint64_ index2->offset = (uint64_t*)realloc(index2->offset, index2->m * 8); memset(index2->offset + old_m, 0, 8 * (index2->m - old_m)); } - for (i = beg + 1; i <= end; ++i) - if (index2->offset[i] == 0) index2->offset[i] = offset; - index2->n = end + 1; + if (beg == end) { + if (index2->offset[beg] == 0) index2->offset[beg] = offset; + } else { + for (i = beg; i <= end; ++i) + if (index2->offset[i] == 0) index2->offset[i] = offset; + } + if (index2->n < end + 1) index2->n = end + 1; } static void merge_chunks(ti_index_t *idx) @@ -310,10 +324,10 @@ ti_index_t *ti_index_core(BGZF *fp, const ti_conf_t *conf) last_tid = intv.tid; last_bin = 0xffffffffu; } else if (last_coor > intv.beg) { - fprintf(stderr, "[ti_index_core] the file is not sorted at line %llu\n", (unsigned long long)lineno); + fprintf(stderr, "[ti_index_core] the file out of order at line %llu\n", (unsigned long long)lineno); exit(1); } - if (intv.bin < 4681) insert_offset2(&idx->index2[intv.tid], intv.beg, intv.end, last_off); + insert_offset2(&idx->index2[intv.tid], intv.beg, intv.end, last_off); if (intv.bin != last_bin) { // then possibly write the binning index if (save_bin != 0xffffffffu) // save_bin==0xffffffffu only happens to the first record insert_offset(idx->index[save_tid], save_bin, save_off, last_off); @@ -612,21 +626,16 @@ static char *get_local_version(const char *fn) free(fnidx); return 0; } -int ti_list_chromosomes(const char *fn) +const char **ti_seqname(const ti_index_t *idx, int *n) { - ti_index_t *idx; - char **names; - int i; + const char **names; khint_t k; - idx = ti_index_load(fn); + *n = idx->n; names = calloc(idx->n, sizeof(void*)); for (k = kh_begin(idx->tname); k < kh_end(idx->tname); ++k) if (kh_exist(idx->tname, k)) - names[kh_val(idx->tname, k)] = (char*)kh_key(idx->tname, k); - for (i = 0; i < idx->n; ++i) printf("%s\n", names[i]); - free(names); - ti_index_destroy(idx); - return 0; + names[kh_val(idx->tname, k)] = kh_key(idx->tname, k); + return names; } ti_index_t *ti_index_load(const char *fn) @@ -732,13 +741,15 @@ static inline int reg2bins(uint32_t beg, uint32_t end, uint16_t list[MAX_BIN]) return i; } -static inline int is_overlap(int32_t beg, int32_t end, int rbeg, int rend) +ti_iter_t ti_first(BGZF *fp) { - return (rend > beg && rbeg < end); + ti_iter_t iter; + iter = calloc(1, sizeof(struct __ti_iter_t)); + iter->fp = fp; iter->from_first = 1; + return iter; } -// ti_fetch helper function retrieves -pair64_t *get_chunk_coordinates(const ti_index_t *idx, int tid, int beg, int end, int* cnt_off) +ti_iter_t ti_query(BGZF *fp, const ti_index_t *idx, int tid, int beg, int end) { uint16_t *bins; int i, n_bins, n_off; @@ -746,18 +757,33 @@ pair64_t *get_chunk_coordinates(const ti_index_t *idx, int tid, int beg, int end khint_t k; khash_t(i) *index; uint64_t min_off; + ti_iter_t iter = 0; + // initialize the iterator + iter = calloc(1, sizeof(struct __ti_iter_t)); + iter->fp = fp; iter->idx = idx; + iter->tid = tid; iter->beg = beg; iter->end = end; iter->i = -1; + // random access bins = (uint16_t*)calloc(MAX_BIN, 2); n_bins = reg2bins(beg, end, bins); index = idx->index[tid]; - min_off = (beg>>TAD_LIDX_SHIFT >= idx->index2[tid].n)? idx->index2[tid].offset[idx->index2[tid].n-1] - : idx->index2[tid].offset[beg>>TAD_LIDX_SHIFT]; + if (idx->index2[tid].n > 0) { + min_off = (beg>>TAD_LIDX_SHIFT >= idx->index2[tid].n)? idx->index2[tid].offset[idx->index2[tid].n-1] + : idx->index2[tid].offset[beg>>TAD_LIDX_SHIFT]; + if (min_off == 0) { // improvement for index files built by tabix prior to 0.1.4 + int n = beg>>TAD_LIDX_SHIFT; + if (n > idx->index2[tid].n) n = idx->index2[tid].n; + for (i = n - 1; i >= 0; --i) + if (idx->index2[tid].offset[i] != 0) break; + if (i >= 0) min_off = idx->index2[tid].offset[i]; + } + } else min_off = 0; // tabix 0.1.2 may produce such index files for (i = n_off = 0; i < n_bins; ++i) { if ((k = kh_get(i, index, bins[i])) != kh_end(index)) n_off += kh_value(index, k).n; } if (n_off == 0) { - free(bins); return 0; + free(bins); return iter; } off = (pair64_t*)calloc(n_off, 16); for (i = n_off = 0; i < n_bins; ++i) { @@ -788,44 +814,67 @@ pair64_t *get_chunk_coordinates(const ti_index_t *idx, int tid, int beg, int end n_off = l + 1; } } - *cnt_off = n_off; - return off; + iter->n_off = n_off; iter->off = off; + return iter; } -int ti_fetch(BGZF *fp, const ti_index_t *idx, int tid, int beg, int end, void *data, ti_fetch_f func) +const char *ti_iter_read(ti_iter_t iter, int *len) { - int n_off; - pair64_t *off = get_chunk_coordinates(idx, tid, beg, end, &n_off); - if (off == 0) return 0; - { - // retrive alignments - uint64_t curr_off; - int i, ret, n_seeks; - kstring_t *str = calloc(1, sizeof(kstring_t)); - n_seeks = 0; i = -1; curr_off = 0; - for (;;) { - if (curr_off == 0 || curr_off >= off[i].v) { // then jump to the next chunk - if (i == n_off - 1) break; // no more chunks - if (i >= 0) assert(curr_off == off[i].v); // otherwise bug - if (i < 0 || off[i].v != off[i+1].u) { // not adjacent chunks; then seek - bgzf_seek(fp, off[i+1].u, SEEK_SET); - curr_off = bgzf_tell(fp); - ++n_seeks; - } - ++i; + if (iter->finished) return 0; + if (iter->from_first) { + int ret; + if ((ret = ti_readline(iter->fp, &iter->str)) < 0) { + iter->finished = 1; + return 0; + } else { + if (len) *len = iter->str.l; + return iter->str.s; + } + } + if (iter->n_off == 0) return 0; + while (1) { + int ret; + if (iter->curr_off == 0 || iter->curr_off >= iter->off[iter->i].v) { // then jump to the next chunk + if (iter->i == iter->n_off - 1) break; // no more chunks + if (iter->i >= 0) assert(iter->curr_off == iter->off[iter->i].v); // otherwise bug + if (iter->i < 0 || iter->off[iter->i].v != iter->off[iter->i+1].u) { // not adjacent chunks; then seek + bgzf_seek(iter->fp, iter->off[iter->i+1].u, SEEK_SET); + iter->curr_off = bgzf_tell(iter->fp); } - if ((ret = ti_readline(fp, str)) >= 0) { - ti_intv_t intv; - curr_off = bgzf_tell(fp); - if (str->s[0] == idx->conf.meta_char) continue; - get_intv((ti_index_t*)idx, str, &intv); - if (intv.tid != tid || intv.beg >= end) break; // no need to proceed - else if (is_overlap(beg, end, intv.beg, intv.end)) func(str->l, str->s, data); - } else break; // end of file + ++iter->i; } -// fprintf(stderr, "[ti_fetch] # seek calls: %d\n", n_seeks); - free(str->s); free(str); + if ((ret = ti_readline(iter->fp, &iter->str)) >= 0) { + ti_intv_t intv; + iter->curr_off = bgzf_tell(iter->fp); + if (iter->str.s[0] == iter->idx->conf.meta_char) continue; + get_intv((ti_index_t*)iter->idx, &iter->str, &intv); + if (intv.tid != iter->tid || intv.beg >= iter->end) break; // no need to proceed + else if (intv.end > iter->beg && iter->end > intv.beg) { + if (len) *len = iter->str.l; + return iter->str.s; + } + } else break; // end of file + } + iter->finished = 1; + return 0; +} + +void ti_iter_destroy(ti_iter_t iter) +{ + if (iter) { + free(iter->str.s); free(iter->off); + free(iter); } - free(off); +} + +int ti_fetch(BGZF *fp, const ti_index_t *idx, int tid, int beg, int end, void *data, ti_fetch_f func) +{ + ti_iter_t iter; + const char *s; + int len; + iter = ti_query(fp, idx, tid, beg, end); + while ((s = ti_iter_read(iter, &len)) != 0) + func(len, s, data); + ti_iter_destroy(iter); return 0; } diff --git a/main.c b/main.c index 520d663..a05939d 100644 --- a/main.c +++ b/main.c @@ -6,13 +6,7 @@ #include "bgzf.h" #include "tabix.h" -#define PACKAGE_VERSION "0.1.5 (r560)" - -static int fetch_func(int l, const char *s, void *data) -{ - printf("%s\n", s); - return 0; -} +#define PACKAGE_VERSION "0.1.6 (r565)" int main(int argc, char *argv[]) { @@ -60,8 +54,21 @@ int main(int argc, char *argv[]) fprintf(stderr, "\n"); return 1; } - if (list_chrms) - return ti_list_chromosomes(argv[optind]); + if (list_chrms) { + ti_index_t *idx; + int i, n; + const char **names; + idx = ti_index_load(argv[optind]); + if (idx == 0) { + fprintf(stderr, "[main] fail to load the index file.\n"); + return 1; + } + names = ti_seqname(idx, &n); + for (i = 0; i < n; ++i) printf("%s\n", names[i]); + free(names); + ti_index_destroy(idx); + return 0; + } if (optind + 1 == argc) { if (force == 0) { struct stat buf; @@ -84,11 +91,14 @@ int main(int argc, char *argv[]) return 1; } if (strcmp(argv[optind+1], ".") == 0) { // retrieve all - kstring_t *str = calloc(1, sizeof(kstring_t)); - while (ti_readline(fp, str) >= 0) { // FIXME: check return code for error - fputs(str->s, stdout); fputc('\n', stdout); + ti_iter_t iter; + const char *s; + int len; + iter = ti_first(fp); + while ((s = ti_iter_read(iter, &len)) != 0) { + fputs(s, stdout); fputc('\n', stdout); } - free(str->s); free(str); + ti_iter_destroy(iter); } else { // retrieve from specified regions ti_index_t *idx; int i; @@ -101,7 +111,14 @@ int main(int argc, char *argv[]) for (i = optind + 1; i < argc; ++i) { int tid, beg, end; if (ti_parse_region(idx, argv[i], &tid, &beg, &end) == 0) { - ti_fetch(fp, idx, tid, beg, end, 0, fetch_func); + ti_iter_t iter; + const char *s; + int len; + iter = ti_query(fp, idx, tid, beg, end); + while ((s = ti_iter_read(iter, &len)) != 0) { + fputs(s, stdout); fputc('\n', stdout); + } + ti_iter_destroy(iter); } else fprintf(stderr, "[main] invalid region: unknown target name or minus interval.\n"); } ti_index_destroy(idx); diff --git a/tabix.h b/tabix.h index 0df29a0..a002d8a 100644 --- a/tabix.h +++ b/tabix.h @@ -43,27 +43,63 @@ typedef int (*ti_fetch_f)(int l, const char *s, void *data); struct __ti_index_t; typedef struct __ti_index_t ti_index_t; +struct __ti_iter_t; +typedef struct __ti_iter_t *ti_iter_t; + typedef struct { int32_t preset; - int32_t sc, bc, ec; + int32_t sc, bc, ec; // seq col., beg col. and end col. int32_t meta_char, line_skip; } ti_conf_t; -extern ti_conf_t ti_conf_gff, ti_conf_bed, ti_conf_psltbl, ti_conf_vcf, ti_conf_sam; +extern ti_conf_t ti_conf_gff, ti_conf_bed, ti_conf_psltbl, ti_conf_vcf, ti_conf_sam; // preset #ifdef __cplusplus extern "C" { #endif - int ti_readline(BGZF *fp, kstring_t *str); - + /* Build the index for file . File .tbi will be generated + * and overwrite the file of the same name. Return -1 on failure. */ int ti_index_build(const char *fn, const ti_conf_t *conf); + + /* Load the index from file .tbi. If is a URL and the index + * file is not in the working directory, .tbi will be + * downloaded. Return NULL on failure. */ ti_index_t *ti_index_load(const char *fn); - int ti_list_chromosomes(const char *fn); + + /* Destroy the index */ void ti_index_destroy(ti_index_t *idx); + + /* Get the list of sequence names. Each "char*" pointer points to a + * internal member of the index, so DO NOT modify the returned + * pointer; otherwise the index will be corrupted. The returned + * pointer should be freed by a single free() call by the routine + * calling this function. The number of sequences is returned at *n. */ + const char **ti_seqname(const ti_index_t *idx, int *n); + + /* Parse a region like: chr2, chr2:100, chr2:100-200. Return -1 on failure. */ int ti_parse_region(ti_index_t *idx, const char *str, int *tid, int *begin, int *end); + + /* Get the iterator pointing to the first record at the current file + * position. If the file is just openned, the iterator points to the + * first record in the file. */ + ti_iter_t ti_first(BGZF *fp); + + /* Get the iterator pointing to the first record in region tid:beg-end */ + ti_iter_t ti_query(BGZF *fp, const ti_index_t *idx, int tid, int beg, int end); + + /* Get the data line pointed by the iterator and iterate to the next record. */ + const char *ti_iter_read(ti_iter_t iter, int *len); + + /* Destroy the iterator */ + void ti_iter_destroy(ti_iter_t iter); + + /* The callback version for random access */ int ti_fetch(BGZF *fp, const ti_index_t *idx, int tid, int beg, int end, void *data, ti_fetch_f func); + /* Read one line. */ + int ti_readline(BGZF *fp, kstring_t *str); + #ifdef __cplusplus } #endif -- 2.30.2