X-Git-Url: http://woldlab.caltech.edu/gitweb/?p=samtools.git;a=blobdiff_plain;f=bam_aux.c;h=2247bdfe9ba15d39d5a1dcf193da9d68cdfab8bf;hp=7482500ef08a53f3471858793112d9c72263e33a;hb=5075acdfc387c56ca8411cde3e419e867da9804d;hpb=b27e00385f41769d03a8cca4dbd71275fc9fa906

diff --git a/bam_aux.c b/bam_aux.c
index 7482500..2247bdf 100644
--- a/bam_aux.c
+++ b/bam_aux.c
@@ -25,24 +25,39 @@ 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 == 'Z' || type == 'H') { while (*(s)) ++(s); ++(s); } \
+		else if (type == 'B') (s) += 5 + bam_aux_type2size(*(s)) * (*(int32_t*)((s)+1)); \
+		else (s) += bam_aux_type2size(type); \
+	} 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) ++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)
 {
@@ -72,47 +87,56 @@ int32_t bam_get_tid(const bam_header_t *header, const char *seq_name)
 	return k == kh_end(h)? -1 : kh_value(h, k);
 }
 
-int 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 *beg, int *end)
 {
-	char *s, *p;
-	int i, l, k;
+	char *s;
+	int i, l, k, name_end;
 	khiter_t iter;
 	khash_t(s) *h;
 
 	bam_init_header_hash(header);
 	h = (khash_t(s)*)header->hash;
 
-	l = strlen(str);
-	p = s = (char*)malloc(l+1);
-	/* squeeze out "," */
-	for (i = k = 0; i != l; ++i)
-		if (str[i] != ',' && !isspace(str[i])) s[k++] = str[i];
-	s[k] = 0;
-	for (i = 0; i != k; ++i) if (s[i] == ':') break;
-	s[i] = 0;
-	iter = kh_get(s, h, s); /* get the ref_id */
-	if (iter == kh_end(h)) { // name not found
-		*ref_id = -1; free(s);
-		return -1;
-	}
-	*ref_id = kh_value(h, iter);
-	if (i == k) { /* dump the whole sequence */
-		*begin = 0; *end = 1<<29; free(s);
-		return -1;
-	}
-	for (p = s + i + 1; i != k; ++i) if (s[i] == '-') break;
-	*begin = atoi(p);
-	if (i < k) {
-		p = s + i + 1;
-		*end = atoi(p);
-	} else *end = 1<<29;
-	if (*begin > 0) --*begin;
+	*ref_id = *beg = *end = -1;
+	name_end = l = strlen(str);
+	s = (char*)malloc(l+1);
+	// remove space
+	for (i = k = 0; i < l; ++i)
+		if (!isspace(str[i])) s[k++] = str[i];
+	s[k] = 0; l = k;
+	// determine the sequence name
+	for (i = l - 1; i >= 0; --i) if (s[i] == ':') break; // look for colon from the end
+	if (i >= 0) name_end = i;
+	if (name_end < l) { // check if this is really the end
+		int n_hyphen = 0;
+		for (i = name_end + 1; i < l; ++i) {
+			if (s[i] == '-') ++n_hyphen;
+			else if (!isdigit(s[i]) && s[i] != ',') break;
+		}
+		if (i < l || n_hyphen > 1) name_end = l; // malformated region string; then take str as the name
+		s[name_end] = 0;
+		iter = kh_get(s, h, s);
+		if (iter == kh_end(h)) { // cannot find the sequence name
+			iter = kh_get(s, h, str); // try str as the name
+			if (iter == kh_end(h)) {
+				if (bam_verbose >= 2) fprintf(stderr, "[%s] fail to determine the sequence name.\n", __func__);
+				free(s); return -1;
+			} else s[name_end] = ':', name_end = l;
+		}
+	} else iter = kh_get(s, h, str);
+	*ref_id = kh_val(h, iter);
+	// parse the interval
+	if (name_end < l) {
+		for (i = k = name_end + 1; i < l; ++i)
+			if (s[i] != ',') s[k++] = s[i];
+		s[k] = 0;
+		*beg = atoi(s + name_end + 1);
+		for (i = name_end + 1; i != k; ++i) if (s[i] == '-') break;
+		*end = i < k? atoi(s + i + 1) : 1<<29;
+		if (*beg > 0) --*beg;
+	} else *beg = 0, *end = 1<<29;
 	free(s);
-	if (*begin > *end) {
-		fprintf(stderr, "[bam_parse_region] invalid region.\n");
-		return -1;
-	}
-	return 0;
+	return *beg <= *end? 0 : -1;
 }
 
 int32_t bam_aux2i(const uint8_t *s)
@@ -164,69 +188,9 @@ char *bam_aux2Z(const uint8_t *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)
+#ifdef _WIN32
+double drand48()
 {
-	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);
+	return (double)rand() / RAND_MAX;
 }
+#endif