X-Git-Url: http://woldlab.caltech.edu/gitweb/?p=samtools.git;a=blobdiff_plain;f=bam_import.c;h=29516c92ff42593313caf2833c38213d62ebdc52;hp=1dc906eb49f36428460be547b1582ab52432f651;hb=2347ad62090e036841008e054c234868e21f7a80;hpb=4a17fa7e1f91b2fe04ad334a63fc2b0d5e859d8a

diff --git a/bam_import.c b/bam_import.c
index 1dc906e..29516c9 100644
--- a/bam_import.c
+++ b/bam_import.c
@@ -10,6 +10,7 @@
 #endif
 #include "kstring.h"
 #include "bam.h"
+#include "sam_header.h"
 #include "kseq.h"
 #include "khash.h"
 
@@ -115,7 +116,7 @@ static bam_header_t *hash2header(const kh_ref_t *hash)
 bam_header_t *sam_header_read2(const char *fn)
 {
 	bam_header_t *header;
-	int c, dret, ret;
+	int c, dret, ret, error = 0;
 	gzFile fp;
 	kstream_t *ks;
 	kstring_t *str;
@@ -134,6 +135,10 @@ bam_header_t *sam_header_read2(const char *fn)
 		ks_getuntil(ks, 0, str, &dret);
 		len = atoi(str->s);
 		k = kh_put(ref, hash, s, &ret);
+		if (ret == 0) {
+			fprintf(stderr, "[sam_header_read2] duplicated sequence name: %s\n", s);
+			error = 1;
+		}
 		kh_value(hash, k) = (uint64_t)len<<32 | i;
 		if (dret != '\n')
 			while ((c = ks_getc(ks)) != '\n' && c != -1);
@@ -142,6 +147,7 @@ bam_header_t *sam_header_read2(const char *fn)
 	gzclose(fp);
 	free(str->s); free(str);
 	fprintf(stderr, "[sam_header_read2] %d sequences loaded.\n", kh_size(hash));
+	if (error) return 0;
 	header = hash2header(hash);
 	kh_destroy(ref, hash);
 	return header;
@@ -162,115 +168,49 @@ static inline void parse_error(int64_t n_lines, const char * __restrict msg)
 }
 static inline void append_text(bam_header_t *header, kstring_t *str)
 {
-	int x = header->l_text, y = header->l_text + str->l + 2; // 2 = 1 byte dret + 1 byte null
+	size_t x = header->l_text, y = header->l_text + str->l + 2; // 2 = 1 byte dret + 1 byte null
 	kroundup32(x); kroundup32(y);
-	if (x < y) header->text = (char*)realloc(header->text, y);
+	if (x < y) 
+    {
+        header->n_text = y;
+        header->text = (char*)realloc(header->text, y);
+        if ( !header->text ) 
+        {
+            fprintf(stderr,"realloc failed to alloc %ld bytes\n", y);
+            abort();
+        }
+    }
+    // Sanity check
+    if ( header->l_text+str->l+1 >= header->n_text )
+    {
+        fprintf(stderr,"append_text FIXME: %ld>=%ld, x=%ld,y=%ld\n",  header->l_text+str->l+1,header->n_text,x,y);
+        abort();
+    }
 	strncpy(header->text + header->l_text, str->s, str->l+1); // we cannot use strcpy() here.
 	header->l_text += str->l + 1;
 	header->text[header->l_text] = 0;
 }
 
-int sam_header_parse_rg(bam_header_t *h)
-{
-	kstring_t *rgid, *rglib;
-	char *p, *q, *s, *r;
-	int n = 0;
-
-	// free
-	if (h == 0) return 0;
-	bam_strmap_destroy(h->rg2lib); h->rg2lib = 0;
-	if (h->l_text < 3) return 0;
-	// parse @RG lines
-	h->rg2lib = bam_strmap_init();
-	rgid = calloc(1, sizeof(kstring_t));
-	rglib = calloc(1, sizeof(kstring_t));
-	s = h->text;
-	while ((s = strstr(s, "@RG")) != 0) {
-		if (rgid->l && rglib->l) {
-			bam_strmap_put(h->rg2lib, rgid->s, rglib->s);
-			++n;
-		}
-		rgid->l = rglib->l = 0;
-		s += 3;
-		r = s;
-		if ((p = strstr(s, "ID:")) != 0) {
-			q = p + 3;
-			for (p = q; *p && *p != '\t' && *p != '\r' && *p != '\n'; ++p);
-			kputsn(q, p - q, rgid);
-		} else {
-			fprintf(stderr, "[bam_header_parse] missing ID tag in @RG lines.\n");
-			break;
-		}
-		if (r < p) r = p;
-		if ((p = strstr(s, "LB:")) != 0) {
-			q = p + 3;
-			for (p = q; *p && *p != '\t' && *p != '\r' && *p != '\n'; ++p);
-			kputsn(q, p - q, rglib);
-		} else {
-			fprintf(stderr, "[bam_header_parse] missing LB tag in @RG lines.\n");
-			break;
-		}
-		if (r < p) r = p;
-		s = r + 3;
-	}
-	if (rgid->l && rglib->l) {
-		bam_strmap_put(h->rg2lib, rgid->s, rglib->s);
-		++n;
-	}
-	free(rgid->s); free(rgid);
-	free(rglib->s); free(rglib);
-	if (n == 0) {
-		bam_strmap_destroy(h->rg2lib);
-		h->rg2lib = 0;
-	}
-	return n;
-}
-
 int sam_header_parse(bam_header_t *h)
 {
+	char **tmp;
 	int i;
-	char *s, *p, *q, *r;
-
-	// free
 	free(h->target_len); free(h->target_name);
 	h->n_targets = 0; h->target_len = 0; h->target_name = 0;
 	if (h->l_text < 3) return 0;
-	// count number of @SQ
-	s = h->text;
-	while ((s = strstr(s, "@SQ")) != 0) {
-		++h->n_targets;
-		s += 3;
-	}
+	if (h->dict == 0) h->dict = sam_header_parse2(h->text);
+	tmp = sam_header2list(h->dict, "SQ", "SN", &h->n_targets);
 	if (h->n_targets == 0) return 0;
-	h->target_len = (uint32_t*)calloc(h->n_targets, 4);
-	h->target_name = (char**)calloc(h->n_targets, sizeof(void*));
-	// parse @SQ lines
-	i = 0;
-	s = h->text;
-	while ((s = strstr(s, "@SQ")) != 0) {
-		s += 3;
-		r = s;
-		if ((p = strstr(s, "SN:")) != 0) {
-			q = p + 3;
-			for (p = q; *p && *p != '\t' && *p != '\r' && *p != '\n'; ++p);
-			h->target_name[i] = (char*)calloc(p - q + 1, 1);
-			strncpy(h->target_name[i], q, p - q);
-		} else goto header_err_ret;
-		if (r < p) r = p;
-		if ((p = strstr(s, "LN:")) != 0) h->target_len[i] = strtol(p + 3, 0, 10);
-		else goto header_err_ret;
-		if (r < p) r = p;
-		s = r + 3;
-		++i;
-	}
-	sam_header_parse_rg(h);
+	h->target_name = calloc(h->n_targets, sizeof(void*));
+	for (i = 0; i < h->n_targets; ++i)
+		h->target_name[i] = strdup(tmp[i]);
+	free(tmp);
+	tmp = sam_header2list(h->dict, "SQ", "LN", &h->n_targets);
+	h->target_len = calloc(h->n_targets, 4);
+	for (i = 0; i < h->n_targets; ++i)
+		h->target_len[i] = atoi(tmp[i]);
+	free(tmp);
 	return h->n_targets;
-
-header_err_ret:
-	fprintf(stderr, "[bam_header_parse] missing SN or LN tag in @SQ lines.\n");
-	free(h->target_len); free(h->target_name);
-	h->n_targets = 0; h->target_len = 0; h->target_name = 0;
-	return 0;
 }
 
 bam_header_t *sam_header_read(tamFile fp)
@@ -487,6 +427,27 @@ int sam_read1(tamFile fp, bam_header_t *header, bam1_t *b)
 				memcpy(s, str->s + 5, str->l - 5);
 				s[str->l - 5] = 0;
 				doff += size;
+			} else if (type == 'B') {
+				int32_t n = 0, Bsize, k = 0, size;
+				char *p;
+				if (str->l < 8) parse_error(fp->n_lines, "too few values in aux type B");
+				Bsize = bam_aux_type2size(str->s[5]); // the size of each element
+				for (p = (char*)str->s + 6; *p; ++p) // count the number of elements in the array
+					if (*p == ',') ++n;
+				p = str->s + 7; // now p points to the first number in the array
+				size = 6 + Bsize * n; // total number of bytes allocated to this tag
+				s = alloc_data(b, doff + 6 * Bsize * n) + doff; // allocate memory
+				*s++ = 'B'; *s++ = str->s[5];
+				memcpy(s, &n, 4); s += 4; // write the number of elements
+				if (str->s[5] == 'c')      while (p < str->s + str->l) ((int8_t*)s)[k++]   = (int8_t)strtol(p, &p, 0),   ++p;
+				else if (str->s[5] == 'C') while (p < str->s + str->l) ((uint8_t*)s)[k++]  = (uint8_t)strtol(p, &p, 0),  ++p;
+				else if (str->s[5] == 's') while (p < str->s + str->l) ((int16_t*)s)[k++]  = (int16_t)strtol(p, &p, 0),  ++p; // FIXME: avoid unaligned memory
+				else if (str->s[5] == 'S') while (p < str->s + str->l) ((uint16_t*)s)[k++] = (uint16_t)strtol(p, &p, 0), ++p;
+				else if (str->s[5] == 'i') while (p < str->s + str->l) ((int32_t*)s)[k++]  = (int32_t)strtol(p, &p, 0),  ++p;
+				else if (str->s[5] == 'I') while (p < str->s + str->l) ((uint32_t*)s)[k++] = (uint32_t)strtol(p, &p, 0), ++p;
+				else if (str->s[5] == 'f') while (p < str->s + str->l) ((float*)s)[k++]    = (float)strtod(p, &p),       ++p;
+				else parse_error(fp->n_lines, "unrecognized array type");
+				s += Bsize * n; doff += size;
 			} else parse_error(fp->n_lines, "unrecognized type");
 			if (dret == '\n' || dret == '\r') break;
 		}