Merge pull request #50 from wtsi-npg/devel

release 0.9.5 prep

.travis.yml

@@ -9,11 +9,11 @@ matrix:
   include:
   - os: linux
     compiler: clang
+    env: CFLAGS="-Wno-pointer-sign""
   - os: linux
     compiler: gcc
   - os: linux
     compiler: gcc
-    env: CFLAGS="-std=gnu99 -O0"
 
 env:
   global:

Makefile.am

@@ -7,7 +7,8 @@ TESTS = test/t_array \
         test/t_decode \
         test/t_filterfile \
         test/t_posfile \
-        test/t_i2b
+        test/t_i2b \
+        test/t_read2tags
 SUBDIRS = src test
 dist_doc_DATA = README.md LICENSE
 AM_COLOR_TESTS=always

src/Makefile.am

@@ -1,10 +1,10 @@
-AM_CPPFLAGS = $(HTSLIB_CPPFLAGS) -std=gnu99
+AM_CPPFLAGS = $(HTSLIB_CPPFLAGS) -std=gnu99 -Werror
 AM_LDFLAGS = -rdynamic $(HTSLIB_LDFLAGS)
 
 bin_PROGRAMS = bambi 
-bambi_SOURCES = bambi.c bclfile.c  decode.c  filterfile.c  hts_addendum.c  i2b.c  posfile.c array.c select.c \
-                bambi.h bclfile.h            filterfile.h  hts_addendum.h         posfile.h array.h \
-                parse.c bamit.c    chrsplit.c\
+bambi_SOURCES = bambi.c bclfile.c  decode.c   filterfile.c  hts_addendum.c  i2b.c  posfile.c array.c select.c \
+                bambi.h bclfile.h             filterfile.h  hts_addendum.h         posfile.h array.h \
+                parse.c bamit.c    chrsplit.c read2tags.c   \
                 parse.h bamit.h
 nobase_include_HEADERS = cram/cram_samtools.h cram/pooled_alloc.h cram/sam_header.h cram/string_alloc.h
 bambi_CFLAGS = -I/usr/include/libxml2

src/array.c

@@ -123,6 +123,10 @@ void va_free(va_t *va)
     free(va);
 }
 
+/*
+ * join entries into a 'delim' separated list
+ * NB assumes that entries are strings
+ */
 char *va_join(va_t *va, char *delim)
 {
     int m = 64;
@@ -140,4 +144,15 @@ char *va_join(va_t *va, char *delim)
     return s;
 }
 
+/*
+ * return array number containing 's', or -1 if not found
+ * NB assumes all entries are strings
+ */
+int va_contains(va_t *va, char *s)
+{
+    for (int n=0; n < va->end; n++) {
+        if (strcmp(va->entries[n],s) == 0) return n;
+    }
+    return -1;
+}
 

src/array.h

@@ -59,7 +59,10 @@ va_t *va_init(int max, void(*free_entry)(void*));
 void va_push(va_t *va, void *ent);
 void va_free(va_t *va);
 static inline bool va_isEmpty(va_t *va) { return va->end == 0; }
+
+// NB these functions assume an array of strings
 char *va_join(va_t *va, char *delim);
+int va_contains(va_t *va, char *s);
 
 #endif
 

src/bambi.c

@@ -33,6 +33,7 @@ int main_decode(int argc, char *argv[]);
 int main_i2b(int argc, char *argv[]);
 int main_select(int argc, char *argv[]);
 int main_chrsplit(int argc, char *argv[]);
+int main_read2tags(int argc, char *argv[]);
 
 const char *bambi_version()
 {
@@ -83,6 +84,7 @@ static void usage(FILE *fp)
 "     i2b            converts illumina files to SAM/BAM/CRAM files\n"
 "     select         select reads by alignment\n"
 "     chrsplit       split reads by chromosome\n"
+"     read2tags      convert reads into tags\n"
 "\n"
 "bambi <command> for help on a particular command\n"
 "\n");
@@ -107,6 +109,7 @@ int main(int argc, char *argv[])
     else if (strcmp(argv[1], "i2b") == 0)       ret = main_i2b(argc-1, argv+1);
     else if (strcmp(argv[1], "select") == 0)    ret = main_select(argc-1, argv+1);
     else if (strcmp(argv[1], "chrsplit") == 0)  ret = main_chrsplit(argc-1, argv+1);
+    else if (strcmp(argv[1], "read2tags") == 0) ret = main_read2tags(argc-1, argv+1);
     else if (strcmp(argv[1], "--version") == 0) {
         printf( "bambi %s\n"
                 "Using htslib %s\n"

src/bamit.c

@@ -16,6 +16,8 @@ You should have received a copy of the GNU Affero General Public License
 along with this program.  If not, see <http://www.gnu.org/licenses/>.
 */
 
+#include <stdio.h>
+
 #include "bamit.h"
 
 BAMit_t *BAMit_init(samFile *f, bam_hdr_t *h)
@@ -30,6 +32,42 @@ BAMit_t *BAMit_init(samFile *f, bam_hdr_t *h)
     return bit;
 }
 
+/*
+ * Open a BAM file
+ * arguments are: char *fname               filename to open
+ *                char mode                 'r' or 'w'
+ *                char *fmt                 format [bam,sam,cram]
+ *                char compression level    [0..9]
+ */
+BAMit_t *BAMit_open(char *fname, char mode, char *fmt, char compression_level)
+{
+    samFile *f = NULL;
+    bam_hdr_t *h = NULL;
+    htsFormat *format = NULL;
+    char m[] = "xbC";
+
+    if (fmt) {
+        format = calloc(1,sizeof(htsFormat));
+        if (hts_parse_format(format, fmt) < 0) {
+            fprintf(stderr,"Unknown input format: %s\n", fmt);
+            exit(1);
+        }
+    }
+    m[0] = mode;
+    m[2] = compression_level;
+    f = hts_open_format(fname, m, format);
+    free(format);
+    if (!f) {
+        fprintf(stderr,"Could not open file (%s)\n", fname);
+        exit(1);
+    }
+
+    if (mode == 'r') h = sam_hdr_read(f);
+    else             h = bam_hdr_init();
+
+    return BAMit_init(f,h);
+}
+
 void BAMit_free(void *ptr)
 {
     BAMit_t *bit = (BAMit_t *)ptr;

src/bamit.h

@@ -34,11 +34,41 @@ typedef struct {
     bam1_t *nextRec;
 } BAMit_t;
 
+/*
+ * Open a BAM file
+ * arguments are: char *fname               filename to open
+ *                char mode                 'r' or 'w'
+ *                char *fmt                 format [bam,sam,cram] or NULL
+ *                char compression level    [0..9] or NULL
+ */
+BAMit_t *BAMit_open(char *fname, char mode, char *fmt, char compression_level);
+
+/*
+ * initialise with open file pointer and header
+ */
 BAMit_t *BAMit_init(samFile *f, bam_hdr_t *h);
-void BAMit_free(void *bit);
+
+/*
+ * read next record and advance to next record
+ */
 bam1_t *BAMit_next(BAMit_t *bit);
+
+/*
+ * read next record *without* advancing to next record
+ */
+
 bam1_t *BAMit_peek(BAMit_t *bit);
+
+/*
+ * return true if there is a next record
+ */
+
 bool BAMit_hasnext(BAMit_t *bit);
 
+/*
+ * free data and close file
+ */
+void BAMit_free(void *bit);
+
 #endif