Merge upstream VariantKey changes

Author: nicolaasuni

plugins/variantkey-hex.c

@@ -33,7 +33,7 @@ DEALINGS IN THE SOFTWARE.  */
 
 const char *FILE_VKRS = "vkrs.unsorted.hex";
 const char *FILE_RSVK = "rsvk.unsorted.hex";
-const char *FILE_VKNR = "vknr.unsorted.hex";
+const char *FILE_VKNR = "vknr.unsorted.tsv";
 
 FILE *fp_vkrs; // VariantKey -> rsID
 FILE *fp_rsvk; // rsID -> VariantKey
@@ -99,15 +99,6 @@ int init(int argc, char **argv, bcf_hdr_t *in, bcf_hdr_t *out)
     return 1;
 }
 
-void str2hex(const char *str, char *hex)
-{
-    int i = 0;
-    while (*str)
-    {
-        sprintf(hex+(i++)*2, "%02x", (uint8_t) *str++);
-    }
-}
-
 // Called for each VCF record. Return rec to output the line or NULL to suppress output.
 bcf1_t *process(bcf1_t *rec)
 {
@@ -129,11 +120,7 @@ bcf1_t *process(bcf1_t *rec)
     if (vk & 1)
     {
         // map VariantKey to REF and ALT
-        char hex_ref[2*len_ref+1];
-        char hex_alt[2*len_alt+1];
-        str2hex(rec->d.allele[0], hex_ref);
-        str2hex(rec->d.allele[1], hex_alt);
-        fprintf(fp_vknr, "%016" PRIx64 "\t%016" PRIx64 "\t%02" PRIx8 "\t%02" PRIx8 "\t%s\t%s\n", vk, (uint64_t)(len_ref + len_alt + 2), len_ref, len_alt, hex_ref, hex_alt);
+        fprintf(fp_vknr, "%016" PRIx64 "\t%s\t%s\n", vk, rec->d.allele[0], rec->d.allele[1]);
         nrv++;
     }
     numvar++;

test/vknr.unsorted.hex

@@ -0,0 +1,3 @@
+080013f9a00e1d03	TAACCCCTAACCCTAACCCTAAACCCTA	T
+0800142b90367897	AACCCCTAACCCTAACCCTAACCCT	A
+080014bb8ad3d64f	CCGCCGTTGCAAAGGCGCGCCG	C

test/vknr.unsorted.tsv

@@ -329,9 +329,9 @@ inline int compare_variantkey_chrom_pos(uint64_t vka, uint64_t vkb)
     return compare_uint64_t((vka >> VKSHIFT_POS), (vkb >> VKSHIFT_POS));
 }
 
-inline size_t variantkey_hex(uint64_t code, char *str)
+inline size_t variantkey_hex(uint64_t vk, char *str)
 {
-    return sprintf(str, "%016" PRIx64, code);
+    return sprintf(str, "%016" PRIx64, vk);
 }
 
 inline uint64_t parse_variantkey_hex(const char *vs)

variantkey.c

@@ -65,7 +65,7 @@ extern "C" {
 typedef struct variantkey_t
 {
     uint8_t chrom;   //!< Chromosome encoded number (only the LSB 5 bit are used)
-    uint32_t pos;    //!< Reference position, with the 1st base having position 0 (only the LSB 28 bit are used)
+    uint32_t pos;    //!< Reference position, with the first base having position 0 (only the LSB 28 bit are used)
     uint32_t refalt; //!< Code for Reference and Alternate allele (only the LSB 31 bits are used)
 } variantkey_t;
 
@@ -130,7 +130,7 @@ size_t decode_refalt(uint32_t code, char *ref, size_t *sizeref, char *alt, size_
 /** @brief Returns a 64 bit variant key based on the pre-encoded CHROM, POS (0-based) and REF+ALT.
  *
  * @param chrom      Encoded Chromosome (see encode_chrom).
- * @param pos        Position. The reference position, with the 1st base having position 0.
+ * @param pos        Position. The reference position, with the first base having position 0.
  * @param refalt     Encoded Reference + Alternate (see encode_refalt).
  *
  * @return      VariantKey 64 bit code.
@@ -174,7 +174,7 @@ void decode_variantkey(uint64_t code, variantkey_t *vk);
  *
  * @param chrom      Chromosome. An identifier from the reference genome, no white-space or leading zeros permitted.
  * @param sizechrom  Length of the chrom string, excluding the terminating null byte.
- * @param pos        Position. The reference position, with the 1st base having position 0.
+ * @param pos        Position. The reference position, with the first base having position 0.
  * @param ref        Reference allele. String containing a sequence of nucleotide letters.
  *                   The value in the pos field refers to the position of the first nucleotide in the String.
  *                   Characters must be A-Z, a-z or *
@@ -190,8 +190,8 @@ uint64_t variantkey(const char *chrom, size_t sizechrom, uint32_t pos, const cha
 /** @brief Returns minimum and maximum VariantKeys for range searches.
  *
  * @param chrom     Chromosome encoded number.
- * @param pos_min   Start reference position, with the 1st base having position 0.
- * @param pos_max   End reference position, with the 1st base having position 0.
+ * @param pos_min   Start reference position, with the first base having position 0.
+ * @param pos_max   End reference position, with the first base having position 0.
  * @param range     VariantKey range values.
  *
  * @return      Min and Max variant keys for any given REF+ALT encoding.
@@ -223,15 +223,15 @@ int compare_variantkey_chrom_pos(uint64_t vka, uint64_t vkb);
  *   - 28 bit for POS
  *   - 31 bit for REF+ALT
  *
- * @param code  VariantKey code.
- * @param str   String buffer to be returned (it must be sized 17 bytes at leasr).
+ * @param vk    VariantKey code.
+ * @param str   String buffer to be returned (it must be sized 17 bytes at least).
  *
  * @return      Upon successful return, these function returns the number of characters processed
  *              (excluding the null byte used to end output to strings).
  *              If the buffer size is not sufficient, then the return value is the number of characters required for
  *              buffer string, including the terminating null byte.
  */
-size_t variantkey_hex(uint64_t code, char *str);
+size_t variantkey_hex(uint64_t vk, char *str);
 
 /** @brief Parses a VariantKey hexadecimal string and returns the code.
  *