X-Git-Url: http://woldlab.caltech.edu/gitweb/?p=pysam.git;a=blobdiff_plain;f=samtools%2Fsam.c.pysam.c;fp=samtools%2Fsam.c.pysam.c;h=7f71f7a9818b2a36f5e6f79a21a5daf133bd50c4;hp=0000000000000000000000000000000000000000;hb=d02fe5283ed7a93a2f76a5d6dc6e37b40c11b9b1;hpb=d828f9c9aa78e3d1687265b52de841f3f3852089 diff --git a/samtools/sam.c.pysam.c b/samtools/sam.c.pysam.c new file mode 100644 index 0000000..7f71f7a --- /dev/null +++ b/samtools/sam.c.pysam.c @@ -0,0 +1,181 @@ +#include "pysam.h" + +#include +#include +#include "faidx.h" +#include "sam.h" + +#define TYPE_BAM 1 +#define TYPE_READ 2 + +bam_header_t *bam_header_dup(const bam_header_t *h0) +{ + bam_header_t *h; + int i; + h = bam_header_init(); + *h = *h0; + h->hash = h->dict = h->rg2lib = 0; + h->text = (char*)calloc(h->l_text + 1, 1); + memcpy(h->text, h0->text, h->l_text); + h->target_len = (uint32_t*)calloc(h->n_targets, 4); + h->target_name = (char**)calloc(h->n_targets, sizeof(void*)); + for (i = 0; i < h->n_targets; ++i) { + h->target_len[i] = h0->target_len[i]; + h->target_name[i] = strdup(h0->target_name[i]); + } + return h; +} +static void append_header_text(bam_header_t *header, char* text, int len) +{ + int x = header->l_text + 1; + int y = header->l_text + len + 1; // 1 byte null + if (text == 0) return; + kroundup32(x); + kroundup32(y); + if (x < y) header->text = (char*)realloc(header->text, y); + strncpy(header->text + header->l_text, text, len); // we cannot use strcpy() here. + header->l_text += len; + header->text[header->l_text] = 0; +} + +samfile_t *samopen(const char *fn, const char *mode, const void *aux) +{ + samfile_t *fp; + fp = (samfile_t*)calloc(1, sizeof(samfile_t)); + if (strchr(mode, 'r')) { // read + fp->type |= TYPE_READ; + if (strchr(mode, 'b')) { // binary + fp->type |= TYPE_BAM; + fp->x.bam = strcmp(fn, "-")? bam_open(fn, "r") : bam_dopen(fileno(stdin), "r"); + if (fp->x.bam == 0) goto open_err_ret; + fp->header = bam_header_read(fp->x.bam); + } else { // text + fp->x.tamr = sam_open(fn); + if (fp->x.tamr == 0) goto open_err_ret; + fp->header = sam_header_read(fp->x.tamr); + if (fp->header->n_targets == 0) { // no @SQ fields + if (aux) { // check if aux is present + bam_header_t *textheader = fp->header; + fp->header = sam_header_read2((const char*)aux); + if (fp->header == 0) goto open_err_ret; + append_header_text(fp->header, textheader->text, textheader->l_text); + bam_header_destroy(textheader); + } + if (fp->header->n_targets == 0 && bam_verbose >= 1) + fprintf(pysamerr, "[samopen] no @SQ lines in the header.\n"); + } else if (bam_verbose >= 2) fprintf(pysamerr, "[samopen] SAM header is present: %d sequences.\n", fp->header->n_targets); + } + } else if (strchr(mode, 'w')) { // write + fp->header = bam_header_dup((const bam_header_t*)aux); + if (strchr(mode, 'b')) { // binary + char bmode[3]; + int i, compress_level = -1; + for (i = 0; mode[i]; ++i) if (mode[i] >= '0' && mode[i] <= '9') break; + if (mode[i]) compress_level = mode[i] - '0'; + if (strchr(mode, 'u')) compress_level = 0; + bmode[0] = 'w'; bmode[1] = compress_level < 0? 0 : compress_level + '0'; bmode[2] = 0; + fp->type |= TYPE_BAM; + fp->x.bam = strcmp(fn, "-")? bam_open(fn, bmode) : bam_dopen(fileno(stdout), bmode); + if (fp->x.bam == 0) goto open_err_ret; + bam_header_write(fp->x.bam, fp->header); + } else { // text + // open file + fp->x.tamw = strcmp(fn, "-")? fopen(fn, "w") : stdout; + if (fp->x.tamr == 0) goto open_err_ret; + if (strchr(mode, 'X')) fp->type |= BAM_OFSTR<<2; + else if (strchr(mode, 'x')) fp->type |= BAM_OFHEX<<2; + else fp->type |= BAM_OFDEC<<2; + // write header + if (strchr(mode, 'h')) { + int i; + bam_header_t *alt; + // parse the header text + alt = bam_header_init(); + alt->l_text = fp->header->l_text; alt->text = fp->header->text; + sam_header_parse(alt); + alt->l_text = 0; alt->text = 0; + // check if there are @SQ lines in the header + fwrite(fp->header->text, 1, fp->header->l_text, fp->x.tamw); // FIXME: better to skip the trailing NULL + if (alt->n_targets) { // then write the header text without dumping ->target_{name,len} + if (alt->n_targets != fp->header->n_targets && bam_verbose >= 1) + fprintf(pysamerr, "[samopen] inconsistent number of target sequences. Output the text header.\n"); + } else { // then dump ->target_{name,len} + for (i = 0; i < fp->header->n_targets; ++i) + fprintf(fp->x.tamw, "@SQ\tSN:%s\tLN:%d\n", fp->header->target_name[i], fp->header->target_len[i]); + } + bam_header_destroy(alt); + } + } + } + return fp; + +open_err_ret: + free(fp); + return 0; +} + +void samclose(samfile_t *fp) +{ + if (fp == 0) return; + if (fp->header) bam_header_destroy(fp->header); + if (fp->type & TYPE_BAM) bam_close(fp->x.bam); + else if (fp->type & TYPE_READ) sam_close(fp->x.tamr); + else fclose(fp->x.tamw); + free(fp); +} + +int samread(samfile_t *fp, bam1_t *b) +{ + if (fp == 0 || !(fp->type & TYPE_READ)) return -1; // not open for reading + if (fp->type & TYPE_BAM) return bam_read1(fp->x.bam, b); + else return sam_read1(fp->x.tamr, fp->header, b); +} + +int samwrite(samfile_t *fp, const bam1_t *b) +{ + if (fp == 0 || (fp->type & TYPE_READ)) return -1; // not open for writing + if (fp->type & TYPE_BAM) return bam_write1(fp->x.bam, b); + else { + char *s = bam_format1_core(fp->header, b, fp->type>>2&3); + int l = strlen(s); + fputs(s, fp->x.tamw); fputc('\n', fp->x.tamw); + free(s); + return l + 1; + } +} + +int sampileup(samfile_t *fp, int mask, bam_pileup_f func, void *func_data) +{ + bam_plbuf_t *buf; + int ret; + bam1_t *b; + b = bam_init1(); + buf = bam_plbuf_init(func, func_data); + bam_plbuf_set_mask(buf, mask); + while ((ret = samread(fp, b)) >= 0) + bam_plbuf_push(b, buf); + bam_plbuf_push(0, buf); + bam_plbuf_destroy(buf); + bam_destroy1(b); + return 0; +} + +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(pysamerr, "[samfaipath] fail to read file %s.\n", fn_ref); + } else { + if (bam_verbose >= 3) fprintf(pysamerr, "[samfaipath] build FASTA index...\n"); + if (fai_build(fn_ref) == -1) { + fprintf(pysamerr, "[samfaipath] fail to build FASTA index.\n"); + free(fn_list); fn_list = 0; + } + } + } + return fn_list; +}