* This revision is SERIOUSLY BUGGY. Please NOT use it.

[samtools.git] / bam_aux.c

diff --git a/bam_aux.c b/bam_aux.c
index a63e2aeadccf35e95b786097cd33394cb31a7a06..89e99f281adb652e144eea0a77074a1c6cec023a 100644 (file)
--- a/bam_aux.c
+++ b/bam_aux.c
@@ -19,30 +19,47 @@ void bam_aux_append(bam1_t *b, const char tag[2], char type, int len, uint8_t *d
        b->data[ori_len + 2] = type;
        memcpy(b->data + ori_len + 3, data, len);
 }
-/*
+
 uint8_t *bam_aux_get_core(bam1_t *b, const char tag[2])
 {
        return bam_aux_get(b, tag);
 }
-*/
+
+#define __skip_tag(s) do { \
+               int type = toupper(*(s));                                                                               \
+               ++(s);                                                                                                                  \
+               if (type == 'C' || type == 'A') ++(s);                                                  \
+               else if (type == 'S') (s) += 2;                                                                 \
+               else if (type == 'I' || type == 'F') (s) += 4;                                  \
+               else if (type == 'D') (s) += 8;                                                                 \
+               else if (type == 'Z' || type == 'H') { while (*(s)) ++(s); ++(s); } \
+       } while (0)
+
 uint8_t *bam_aux_get(const bam1_t *b, const char tag[2])
 {
        uint8_t *s;
        int y = tag[0]<<8 | tag[1];
        s = bam1_aux(b);
        while (s < b->data + b->data_len) {
-               int type, x = (int)s[0]<<8 | s[1];
+               int x = (int)s[0]<<8 | s[1];
                s += 2;
                if (x == y) return s;
-               type = toupper(*s); ++s;
-               if (type == 'C') ++s;
-               else if (type == 'S') s += 2;
-               else if (type == 'I' || type == 'F') s += 4;
-               else if (type == 'D') s += 8;
-               else if (type == 'Z' || type == 'H') { while (*s) putchar(*s++); ++s; }
+               __skip_tag(s);
        }
        return 0;
 }
+// s MUST BE returned by bam_aux_get()
+int bam_aux_del(bam1_t *b, uint8_t *s)
+{
+       uint8_t *p, *aux;
+       aux = bam1_aux(b);
+       p = s - 2;
+       __skip_tag(s);
+       memmove(p, s, b->l_aux - (s - aux));
+       b->data_len -= s - p;
+       b->l_aux -= s - p;
+       return 0;
+}
 
 void bam_init_header_hash(bam_header_t *header)
 {
@@ -72,7 +89,7 @@ int32_t bam_get_tid(const bam_header_t *header, const char *seq_name)
        return k == kh_end(h)? -1 : kh_value(h, k);
 }
 
-void bam_parse_region(bam_header_t *header, const char *str, int *ref_id, int *begin, int *end)
+int bam_parse_region(bam_header_t *header, const char *str, int *ref_id, int *begin, int *end)
 {
        char *s, *p;
        int i, l, k;
@@ -93,12 +110,12 @@ void bam_parse_region(bam_header_t *header, const char *str, int *ref_id, int *b
        iter = kh_get(s, h, s); /* get the ref_id */
        if (iter == kh_end(h)) { // name not found
                *ref_id = -1; free(s);
-               return;
+               return -1;
        }
        *ref_id = kh_value(h, iter);
        if (i == k) { /* dump the whole sequence */
                *begin = 0; *end = 1<<29; free(s);
-               return;
+               return -1;
        }
        for (p = s + i + 1; i != k; ++i) if (s[i] == '-') break;
        *begin = atoi(p);
@@ -107,8 +124,12 @@ void bam_parse_region(bam_header_t *header, const char *str, int *ref_id, int *b
                *end = atoi(p);
        } else *end = 1<<29;
        if (*begin > 0) --*begin;
-       assert(*begin <= *end);
        free(s);
+       if (*begin > *end) {
+               fprintf(stderr, "[bam_parse_region] invalid region.\n");
+               return -1;
+       }
+       return 0;
 }
 
 int32_t bam_aux2i(const uint8_t *s)
@@ -159,70 +180,3 @@ char *bam_aux2Z(const uint8_t *s)
        if (type == 'Z' || type == 'H') return (char*)s;
        else return 0;
 }
-
-/******************
- * rg2lib related *
- ******************/
-
-int bam_strmap_put(void *rg2lib, const char *rg, const char *lib)
-{
-       int ret;
-       khint_t k;
-       khash_t(r2l) *h = (khash_t(r2l)*)rg2lib;
-       char *key;
-       if (h == 0) return 1;
-       key = strdup(rg);
-       k = kh_put(r2l, h, key, &ret);
-       if (ret) kh_val(h, k) = strdup(lib);
-       else {
-               fprintf(stderr, "[bam_rg2lib_put] duplicated @RG ID: %s\n", rg);
-               free(key);
-       }
-       return 0;
-}
-
-const char *bam_strmap_get(const void *rg2lib, const char *rg)
-{
-       const khash_t(r2l) *h = (const khash_t(r2l)*)rg2lib;
-       khint_t k;
-       if (h == 0) return 0;
-       k = kh_get(r2l, h, rg);
-       if (k != kh_end(h)) return (const char*)kh_val(h, k);
-       else return 0;
-}
-
-void *bam_strmap_dup(const void *rg2lib)
-{
-       const khash_t(r2l) *h = (const khash_t(r2l)*)rg2lib;
-       khash_t(r2l) *g;
-       khint_t k, l;
-       int ret;
-       if (h == 0) return 0;
-       g = kh_init(r2l);
-       for (k = kh_begin(h); k < kh_end(h); ++k) {
-               if (kh_exist(h, k)) {
-                       char *key = strdup(kh_key(h, k));
-                       l = kh_put(r2l, g, key, &ret);
-                       kh_val(g, l) = strdup(kh_val(h, k));
-               }
-       }
-       return g;
-}
-
-void *bam_strmap_init()
-{
-       return (void*)kh_init(r2l);
-}
-
-void bam_strmap_destroy(void *rg2lib)
-{
-       khash_t(r2l) *h = (khash_t(r2l)*)rg2lib;
-       khint_t k;
-       if (h == 0) return;
-       for (k = kh_begin(h); k < kh_end(h); ++k) {
-               if (kh_exist(h, k)) {
-                       free((char*)kh_key(h, k)); free(kh_val(h, k));
-               }
-       }
-       kh_destroy(r2l, h);
-}