X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=bam_sort.c;h=7d00cd1a83321729c1384cf56f066d2c83f65548;hb=0ccd9d36ebf35ce620a8248ecf4336c84065e6c0;hp=6c3f52d17ce4dc0dfd8c9d1277cf8050c7e267ca;hpb=d07731e7c769ac5f2709ac60eb60ddd3fa0120fd;p=samtools.git diff --git a/bam_sort.c b/bam_sort.c index 6c3f52d..7d00cd1 100644 --- a/bam_sort.c +++ b/bam_sort.c @@ -288,6 +288,7 @@ int bam_merge(int argc, char *argv[]) fprintf(stderr, " -u uncompressed BAM output\n"); fprintf(stderr, " -f overwrite the output BAM if exist\n"); fprintf(stderr, " -1 compress level 1\n"); + fprintf(stderr, " -l INT compression level, from 0 to 9 [-1]\n"); fprintf(stderr, " -@ INT number of BAM compression threads [0]\n"); fprintf(stderr, " -R STR merge file in the specified region STR [all]\n"); fprintf(stderr, " -h FILE copy the header in FILE to [in1.bam]\n\n"); @@ -318,6 +319,40 @@ int bam_merge(int argc, char *argv[]) typedef bam1_t *bam1_p; +static int change_SO(bam_header_t *h, const char *so) +{ + char *p, *q, *beg = 0, *end = 0, *newtext; + if (h->l_text > 3) { + if (strncmp(h->text, "@HD", 3) == 0) { + if ((p = strchr(h->text, '\n')) == 0) return -1; + *p = '\0'; + if ((q = strstr(h->text, "\tSO:")) != 0) { + *p = '\n'; // change back + if (strncmp(q + 4, so, p - q - 4) != 0) { + beg = q; + for (q += 4; *q != '\n' && *q != '\t'; ++q); + end = q; + } else return 0; // no need to change + } else beg = end = p, *p = '\n'; + } + } + if (beg == 0) { // no @HD + h->l_text += strlen(so) + 15; + newtext = malloc(h->l_text + 1); + sprintf(newtext, "@HD\tVN:1.3\tSO:%s\n", so); + strcat(newtext, h->text); + } else { // has @HD but different or no SO + h->l_text = (beg - h->text) + (4 + strlen(so)) + (h->text + h->l_text - end); + newtext = malloc(h->l_text + 1); + strncpy(newtext, h->text, beg - h->text); + sprintf(newtext + (beg - h->text), "\tSO:%s", so); + strcat(newtext, end); + } + free(h->text); + h->text = newtext; + return 0; +} + static inline int bam1_lt(const bam1_p a, const bam1_p b) { if (g_is_by_qname) { @@ -406,7 +441,7 @@ static int sort_blocks(int n_files, size_t k, bam1_p *buf, const char *prefix, c */ void bam_sort_core_ext(int is_by_qname, const char *fn, const char *prefix, size_t _max_mem, int is_stdout, int n_threads, int level) { - int ret, i, extra_mem, n_files = 0; + int ret, i, n_files = 0; size_t mem, max_k, k, max_mem; bam_header_t *header; bamFile fp; @@ -414,7 +449,6 @@ void bam_sort_core_ext(int is_by_qname, const char *fn, const char *prefix, size char *fnout = 0; if (n_threads < 2) n_threads = 1; - extra_mem = sizeof(void*) + sizeof(void*) + (sizeof(bam1_t) - sizeof(bam1_core_t)); g_is_by_qname = is_by_qname; max_k = k = 0; mem = 0; max_mem = _max_mem * n_threads; @@ -425,6 +459,8 @@ void bam_sort_core_ext(int is_by_qname, const char *fn, const char *prefix, size return; } header = bam_header_read(fp); + if (is_by_qname) change_SO(header, "queryname"); + else change_SO(header, "coordinate"); // write sub files for (;;) { if (k == max_k) { @@ -436,7 +472,12 @@ void bam_sort_core_ext(int is_by_qname, const char *fn, const char *prefix, size if (buf[k] == 0) buf[k] = (bam1_t*)calloc(1, sizeof(bam1_t)); b = buf[k]; if ((ret = bam_read1(fp, b)) < 0) break; - mem += ret + extra_mem; + if (b->data_len < b->m_data>>2) { // shrink + b->m_data = b->data_len; + kroundup32(b->m_data); + b->data = realloc(b->data, b->m_data); + } + mem += sizeof(bam1_t) + b->m_data + sizeof(void*) + sizeof(void*); // two sizeof(void*) for the data allocated to pointer arrays ++k; if (mem >= max_mem) { n_files = sort_blocks(n_files, k, buf, prefix, header, n_threads); @@ -510,7 +551,14 @@ int bam_sort(int argc, char *argv[]) } } if (optind + 2 > argc) { - fprintf(stderr, "Usage: samtools sort [-on] [-m maxMem=1G] \n"); + fprintf(stderr, "\n"); + fprintf(stderr, "Usage: samtools sort [options] \n\n"); + fprintf(stderr, "Options: -n sort by read name\n"); + fprintf(stderr, " -o final output to stdout\n"); + fprintf(stderr, " -l INT compression level, from 0 to 9 [-1]\n"); + fprintf(stderr, " -@ INT number of sorting and compression threads [1]\n"); + fprintf(stderr, " -m INT max memory per thread; suffix K/M/G recognized [768M]\n"); + fprintf(stderr, "\n"); return 1; } bam_sort_core_ext(is_by_qname, argv[optind], argv[optind+1], max_mem, is_stdout, n_threads, level);