Release 1.1: various minor bug fixes

Makefile

@@ -81,7 +81,7 @@ lib-shared: libhts.so
 endif
 
 
-PACKAGE_VERSION  = 1.0
+PACKAGE_VERSION  = 1.1
 LIBHTS_SOVERSION = 1
 
 
@@ -234,11 +234,13 @@ bgzip.o: bgzip.c $(htslib_bgzf_h) $(htslib_hts_h)
 tabix.o: tabix.c $(htslib_tbx_h) $(htslib_sam_h) $(htslib_vcf_h) htslib/kseq.h $(htslib_bgzf_h) $(htslib_hts_h)
 
 
+# For tests that might use it, set $REF_PATH explicitly to use only reference
+# areas within the test suite (or set it to ':' to use no reference areas).
 check test: $(BUILT_TEST_PROGRAMS)
 	test/fieldarith test/fieldarith.sam
 	test/hfile
 	test/sam
-	cd test && ./test_view.pl
+	cd test && REF_PATH=: ./test_view.pl
 	cd test && ./test.pl
 
 test/fieldarith: test/fieldarith.o libhts.a
@@ -267,12 +269,12 @@ test/test-vcf-api.o: test/test-vcf-api.c $(htslib_hts_h) $(htslib_vcf_h) htslib/
 test/test-vcf-sweep.o: test/test-vcf-sweep.c $(htslib_vcf_sweep_h)
 
 
-install: installdirs install-$(SHLIB_FLAVOUR) install-pkgconfig
+install: libhts.a $(BUILT_PROGRAMS) installdirs install-$(SHLIB_FLAVOUR) install-pkgconfig
 	$(INSTALL_PROGRAM) $(BUILT_PROGRAMS) $(DESTDIR)$(bindir)
 	$(INSTALL_DATA) htslib/*.h $(DESTDIR)$(includedir)/htslib
 	$(INSTALL_DATA) libhts.a $(DESTDIR)$(libdir)/libhts.a
-	$(INSTALL_DATA) *.1 $(DESTDIR)$(man1dir)
-	$(INSTALL_DATA) *.5 $(DESTDIR)$(man5dir)
+	$(INSTALL_DATA) tabix.1 $(DESTDIR)$(man1dir)
+	$(INSTALL_DATA) faidx.5 sam.5 vcf.5 $(DESTDIR)$(man5dir)
 
 installdirs:
 	$(INSTALL_DIR) $(DESTDIR)$(bindir) $(DESTDIR)$(includedir) $(DESTDIR)$(includedir)/htslib $(DESTDIR)$(libdir) $(DESTDIR)$(man1dir) $(DESTDIR)$(man5dir) $(DESTDIR)$(pkgconfigdir)

README.md

@@ -12,6 +12,6 @@ This project also includes the popular tabix indexer, which indexes both `.tbi`
 and `.csi` formats, and the bgzip compression utility.
 
 [1]: http://samtools.github.io/hts-specs/
-[2]: http://samtools.github.io/bcftools/
-[3]: http://github.com/samtools/samtools
+[2]: http://github.com/samtools/samtools
+[3]: http://samtools.github.io/bcftools/
 [4]: http://zlib.net/

bgzf.c

@@ -240,11 +240,39 @@ static int bgzf_compress(void *_dst, int *dlen, void *src, int slen, int level)
     return 0;
 }
 
+static int bgzf_gzip_compress(BGZF *fp, void *_dst, int *dlen, void *src, int slen, int level)
+{
+    uint8_t *dst = (uint8_t*)_dst;
+    z_stream *zs = fp->gz_stream;
+    if ( !zs )
+    {
+        zs = fp->gz_stream = (z_stream*)calloc(1,sizeof(z_stream));
+        zs->zalloc = NULL;
+        zs->zfree  = NULL;
+        if ( deflateInit2(zs, level, Z_DEFLATED, 15|16, 8, Z_DEFAULT_STRATEGY)!=Z_OK ) return -1;   // gzip output
+    }
+    int flush = slen ? Z_NO_FLUSH : Z_FINISH;
+    zs->next_in   = (Bytef*)src;
+    zs->avail_in  = slen;
+    zs->next_out  = dst;
+    zs->avail_out = *dlen;
+    if ( deflate(zs, flush) == Z_STREAM_ERROR ) return -1;
+    *dlen = *dlen - zs->avail_out;
+    return 0;
+}
+
 // Deflate the block in fp->uncompressed_block into fp->compressed_block. Also adds an extra field that stores the compressed block length.
 static int deflate_block(BGZF *fp, int block_length)
 {
     int comp_size = BGZF_MAX_BLOCK_SIZE;
-    if (bgzf_compress(fp->compressed_block, &comp_size, fp->uncompressed_block, block_length, fp->compress_level) != 0) {
+    int ret;
+    if ( !fp->is_gzip )
+        ret = bgzf_compress(fp->compressed_block, &comp_size, fp->uncompressed_block, block_length, fp->compress_level);
+    else
+        ret = bgzf_gzip_compress(fp, fp->compressed_block, &comp_size, fp->uncompressed_block, block_length, fp->compress_level);
+
+    if ( ret != 0 )
+    {
         fp->errcode |= BGZF_ERR_ZLIB;
         return -1;
     }
@@ -778,7 +806,8 @@ int bgzf_close(BGZF* fp)
     }
     if ( fp->is_gzip )
     {
-        (void)inflateEnd(fp->gz_stream);
+        if (!fp->is_write) (void)inflateEnd(fp->gz_stream);
+        else (void)deflateEnd(fp->gz_stream);
         free(fp->gz_stream);
     }
     ret = hclose(fp->fp);

cram/cram_index.c

@@ -155,6 +155,8 @@ int cram_index_load(cram_fd *fd, const char *fn) {
 	}
 	kstr.s = s;
 	kstr.l = l;
+	kstr.m = l; // conservative estimate of the size allocated
+	kputsn("", 0, &kstr); // ensure kstr.s is NUL-terminated
     }
 
 

cram/cram_io.c

@@ -1434,7 +1434,7 @@ static int cram_populate_ref(cram_fd *fd, int id, ref_entry *r) {
 	fprintf(stderr, "cram_populate_ref on fd %p, id %d\n", fd, id);
 
     if (!ref_path || *ref_path == 0)
-	ref_path = "|http://www.ebi.ac.uk:80/ena/cram/md5/%s";
+	ref_path = "http://www.ebi.ac.uk:80/ena/cram/md5/%s";
 
     if (!r->name)
 	return -1;

cram/open_trace_file.c

@@ -302,43 +302,37 @@ mFILE *open_path_mfile(char *file, char *path, char *relative_to) {
      * special, otherwise we treat the element as a directory.
      */
     for (ele = newsearch; *ele; ele += strlen(ele)+1) {
-	int i;
-	char *suffix[6] = {"", ".gz", ".bz2", ".sz", ".Z", ".bz2"};
-	for (i = 0; i < 6; i++) {
-	    char file2[1024];
-	    char *ele2;
-	    int valid = 1;
-
-	    /*
-	     * '|' prefixing a path component indicates that we do not
-	     * wish to perform the compression extension searching in that
-	     * location.
-	     */
-	    if (*ele == '|') {
-		ele2 = ele+1;
-		valid = (i == 0);
-	    } else {
-		ele2 = ele;
-	    }
+	char *ele2;
+
+	/*
+	 * '|' prefixing a path component indicates that we do not
+	 * wish to perform the compression extension searching in that
+	 * location.
+	 *
+	 * NB: this has been removed from the htslib implementation.
+	 */
+	if (*ele == '|') {
+	    ele2 = ele+1;
+	} else {
+	    ele2 = ele;
+	}
 
-	    sprintf(file2, "%s%s", file, suffix[i]);
-
-	    if (0 == strncmp(ele2, "URL=", 4)) {
-		if (valid && (fp = find_file_url(file2, ele2+4))) {
-		    free(newsearch);
-		    return fp;
-		}
-	    } else if (!strncmp(ele2, "http:", 5) || !strncmp(ele2, "ftp:", 4)) {
-		if (valid && (fp = find_file_url(file2, ele2))) {
-		    free(newsearch);
-		    return fp;
-		}
-	    } else if (valid && (fp = find_file_dir(file2, ele2))) {
+	if (0 == strncmp(ele2, "URL=", 4)) {
+	    if ((fp = find_file_url(file, ele2+4))) {
 		free(newsearch);
 		return fp;
 	    }
-	}
-    }
+	} else if (!strncmp(ele2, "http:", 5) ||
+		   !strncmp(ele2, "ftp:", 4)) {
+	    if ((fp = find_file_url(file, ele2))) {
+		free(newsearch);
+		return fp;
+	    }
+	} else if ((fp = find_file_dir(file, ele2))) {
+	    free(newsearch);
+	    return fp;
+	} 
+   }
 
     free(newsearch);
 

faidx.c

@@ -409,10 +409,20 @@ char *fai_fetch(const faidx_t *fai, const char *str, int *len)
     return s;
 }
 
-int faidx_fetch_nseq(const faidx_t *fai)
+int faidx_nseq(const faidx_t *fai)
 {
     return fai->n;
 }
+const char *faidx_iseq(const faidx_t *fai, int i)
+{
+    return fai->name[i];
+}
+int faidx_seq_len(const faidx_t *fai, const char *seq)
+{
+    khint_t k = kh_get(s, fai->hash, seq);
+    if ( k == kh_end(fai->hash) ) return -1;
+    return kh_val(fai->hash, k).len;
+}
 
 char *faidx_fetch_seq(const faidx_t *fai, const char *c_name, int p_beg_i, int p_end_i, int *len)
 {

hts.c

@@ -708,10 +708,6 @@ int hts_idx_push(hts_idx_t *idx, int tid, int beg, int end, uint64_t offset, int
         idx->z.save_off = idx->z.last_off;
         idx->z.save_bin = idx->z.last_bin = bin;
         idx->z.save_tid = tid;
-        if (tid < 0) { // come to the end of the records having coordinates
-            hts_idx_finish(idx, offset);
-            return 0;
-        }
     }
     if (is_mapped) ++idx->z.n_mapped;
     else ++idx->z.n_unmapped;

htslib/faidx.h

@@ -113,6 +113,21 @@ extern "C" {
      */
     int faidx_has_seq(const faidx_t *fai, const char *seq);
 
+    /*!
+      @abstract    Return number of sequences in fai index
+     */
+    int faidx_nseq(const faidx_t *fai);
+
+    /*!
+      @abstract    Return name of i-th sequence
+     */
+    const char *faidx_iseq(const faidx_t *fai, int i);
+
+    /*!
+      @abstract    Return sequence length, -1 if not present
+     */
+    int faidx_seq_len(const faidx_t *fai, const char *seq);
+
 #ifdef __cplusplus
 }
 #endif

htslib/khash_str2int.h

@@ -90,7 +90,8 @@ static inline int khash_str2int_get(void *_hash, const char *str, int *value)
 /*
  *  Add a new string to the dictionary, auto-incrementing the value.
  *  On success returns the newly inserted integer id, on error -1
- *  is returned.
+ *  is returned. Note that the key must continue to exist throughout
+ *  the whole life of _hash.
  */
 static inline int khash_str2int_inc(void *_hash, const char *str)
 {
@@ -106,7 +107,8 @@ static inline int khash_str2int_inc(void *_hash, const char *str)
 
 /*
  *  Set a new key,value pair. On success returns the bin index, on
- *  error -1 is returned.
+ *  error -1 is returned. Note that the key must contnue to exist
+ *  throughout the whole life of _hash.
  */
 static inline int khash_str2int_set(void *_hash, const char *str, int value)
 {

htslib/vcf.h

@@ -590,6 +590,17 @@ extern "C" {
     bcf_fmt_t *bcf_get_fmt(const bcf_hdr_t *hdr, bcf1_t *line, const char *key);
     bcf_info_t *bcf_get_info(const bcf_hdr_t *hdr, bcf1_t *line, const char *key);
 
+    /**
+     * bcf_get_*_id() - returns pointer to FORMAT/INFO field data given the header index instead of the string ID
+     * @line: VCF line obtained from vcf_parse1
+     * @id:  The header index for the tag, obtained from bcf_hdr_id2int()
+     * 
+     * Returns bcf_fmt_t* / bcf_info_t*. These functions do not check if the index is valid 
+     * as their goal is to avoid the header lookup.
+     */
+    bcf_fmt_t *bcf_get_fmt_id(bcf1_t *line, const int id);
+    bcf_info_t *bcf_get_info_id(bcf1_t *line, const int id);
+
     /**
      *  bcf_get_info_*() - get INFO values, integers or floats
      *  @hdr:       BCF header

tabix.1

@@ -1,4 +1,4 @@
-.TH tabix 1 "15 August 2014" "tabix-1.0" "Bioinformatics tools"
+.TH tabix 1 "23 September 2014" "htslib-1.1" "Bioinformatics tools"
 .SH NAME
 .PP
 bgzip - Block compression/decompression utility