diff PsiCLASS-1.0.2/samtools-0.1.19/bam_aux.c @ 0:903fc43d6227 draft default tip

Uploaded
author lsong10
date Fri, 26 Mar 2021 16:52:45 +0000
parents
children
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/PsiCLASS-1.0.2/samtools-0.1.19/bam_aux.c	Fri Mar 26 16:52:45 2021 +0000
@@ -0,0 +1,217 @@
+#include <ctype.h>
+#include "bam.h"
+#include "khash.h"
+typedef char *str_p;
+KHASH_MAP_INIT_STR(s, int)
+KHASH_MAP_INIT_STR(r2l, str_p)
+
+void bam_aux_append(bam1_t *b, const char tag[2], char type, int len, uint8_t *data)
+{
+	int ori_len = b->data_len;
+	b->data_len += 3 + len;
+	b->l_aux += 3 + len;
+	if (b->m_data < b->data_len) {
+		b->m_data = b->data_len;
+		kroundup32(b->m_data);
+		b->data = (uint8_t*)realloc(b->data, b->m_data);
+	}
+	b->data[ori_len] = tag[0]; b->data[ori_len + 1] = tag[1];
+	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 == '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 x = (int)s[0]<<8 | s[1];
+		s += 2;
+		if (x == y) return 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;
+}
+
+int bam_aux_drop_other(bam1_t *b, uint8_t *s)
+{
+	if (s) {
+		uint8_t *p, *aux;
+		aux = bam1_aux(b);
+		p = s - 2;
+		__skip_tag(s);
+		memmove(aux, p, s - p);
+		b->data_len -= b->l_aux - (s - p);
+		b->l_aux = s - p;
+	} else {
+		b->data_len -= b->l_aux;
+		b->l_aux = 0;
+	}
+	return 0;
+}
+
+void bam_init_header_hash(bam_header_t *header)
+{
+	if (header->hash == 0) {
+		int ret, i;
+		khiter_t iter;
+		khash_t(s) *h;
+		header->hash = h = kh_init(s);
+		for (i = 0; i < header->n_targets; ++i) {
+			iter = kh_put(s, h, header->target_name[i], &ret);
+			kh_value(h, iter) = i;
+		}
+	}
+}
+
+void bam_destroy_header_hash(bam_header_t *header)
+{
+	if (header->hash)
+		kh_destroy(s, (khash_t(s)*)header->hash);
+}
+
+int32_t bam_get_tid(const bam_header_t *header, const char *seq_name)
+{
+	khint_t k;
+	khash_t(s) *h = (khash_t(s)*)header->hash;
+	k = kh_get(s, h, 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 *beg, int *end)
+{
+	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;
+
+	*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);
+        if (iter == kh_end(h)) {
+          free(s); 
+          return -1;
+        }
+	*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);
+	return *beg <= *end? 0 : -1;
+}
+
+int32_t bam_aux2i(const uint8_t *s)
+{
+	int type;
+	if (s == 0) return 0;
+	type = *s++;
+	if (type == 'c') return (int32_t)*(int8_t*)s;
+	else if (type == 'C') return (int32_t)*(uint8_t*)s;
+	else if (type == 's') return (int32_t)*(int16_t*)s;
+	else if (type == 'S') return (int32_t)*(uint16_t*)s;
+	else if (type == 'i' || type == 'I') return *(int32_t*)s;
+	else return 0;
+}
+
+float bam_aux2f(const uint8_t *s)
+{
+	int type;
+	type = *s++;
+	if (s == 0) return 0.0;
+	if (type == 'f') return *(float*)s;
+	else return 0.0;
+}
+
+double bam_aux2d(const uint8_t *s)
+{
+	int type;
+	type = *s++;
+	if (s == 0) return 0.0;
+	if (type == 'd') return *(double*)s;
+	else return 0.0;
+}
+
+char bam_aux2A(const uint8_t *s)
+{
+	int type;
+	type = *s++;
+	if (s == 0) return 0;
+	if (type == 'A') return *(char*)s;
+	else return 0;
+}
+
+char *bam_aux2Z(const uint8_t *s)
+{
+	int type;
+	type = *s++;
+	if (s == 0) return 0;
+	if (type == 'Z' || type == 'H') return (char*)s;
+	else return 0;
+}
+
+#ifdef _WIN32
+double drand48()
+{
+	return (double)rand() / RAND_MAX;
+}
+#endif