Base mod API improvements

Makefile

@@ -214,6 +214,7 @@ LIBHTS_OBJS = \
 	regidx.o \
 	region.o \
 	sam.o \
+	sam_mods.o \
 	synced_bcf_reader.o \
 	vcf_sweep.o \
 	tbx.o \
@@ -457,6 +458,7 @@ hts_expr.o hts_expr.pico: hts_expr.c config.h $(htslib_hts_expr_h) $(htslib_hts_
 hts_os.o hts_os.pico: hts_os.c config.h $(htslib_hts_defs_h) os/rand.c
 vcf.o vcf.pico: vcf.c config.h $(htslib_vcf_h) $(htslib_bgzf_h) $(htslib_tbx_h) $(htslib_hfile_h) $(hts_internal_h) $(htslib_khash_str2int_h) $(htslib_kstring_h) $(htslib_sam_h) $(htslib_khash_h) $(htslib_kseq_h) $(htslib_hts_endian_h)
 sam.o sam.pico: sam.c config.h $(htslib_hts_defs_h) $(htslib_sam_h) $(htslib_bgzf_h) $(cram_h) $(hts_internal_h) $(sam_internal_h) $(htslib_hfile_h) $(htslib_hts_endian_h) $(htslib_hts_expr_h) $(header_h) $(htslib_khash_h) $(htslib_kseq_h) $(htslib_kstring_h)
+sam_mods.o sam_mods.pico: sam_mods.c config.h $(htslib_sam_h) $(textutils_internal_h)
 tbx.o tbx.pico: tbx.c config.h $(htslib_tbx_h) $(htslib_bgzf_h) $(htslib_hts_endian_h) $(hts_internal_h) $(htslib_khash_h)
 faidx.o faidx.pico: faidx.c config.h $(htslib_bgzf_h) $(htslib_faidx_h) $(htslib_hfile_h) $(htslib_khash_h) $(htslib_kstring_h) $(hts_internal_h)
 bcf_sr_sort.o bcf_sr_sort.pico: bcf_sr_sort.c config.h $(bcf_sr_sort_h) $(htslib_khash_str2int_h) $(htslib_kbitset_h)

htslib/sam.h

@@ -2224,6 +2224,12 @@ typedef struct hts_base_mod {
     int qual;
 } hts_base_mod;
 
+#define HTS_MOD_UNKNOWN   -1  // In MM but no ML
+#define HTS_MOD_UNCHECKED -2  // Not in MM and in explicit mode
+
+// Flags for hts_parse_basemod2
+#define HTS_MOD_REPORT_UNCHECKED 1
+
 /// Allocates an hts_base_mode_state.
 /**
  * @return An hts_base_mode_state pointer on success,
@@ -2260,6 +2266,22 @@ void hts_base_mod_state_free(hts_base_mod_state *state);
 HTSLIB_EXPORT
 int bam_parse_basemod(const bam1_t *b, hts_base_mod_state *state);
 
+/// Parses the Mm and Ml tags out of a bam record.
+/**
+ * @param b        BAM alignment record
+ * @param state    The base modification state pointer.
+ * @param flags    A bit-field controlling base modification processing
+ *
+ * @return 0 on success,
+ *         -1 on failure.
+ *
+ * This fills out the contents of the modification state, resetting the
+ * iterator location to the first sequence base.
+ */
+HTSLIB_EXPORT
+int bam_parse_basemod2(const bam1_t *b, hts_base_mod_state *state,
+                       uint32_t flags);
+
 /// Returns modification status for the next base position in the query seq.
 /**
  * @param b        BAM alignment record
@@ -2349,6 +2371,26 @@ HTSLIB_EXPORT
 int bam_mods_query_type(hts_base_mod_state *state, int code,
                         int *strand, int *implicit, char *canonical);
 
+/// Returns data about the i^th modification type for the alignment record.
+/**
+ * @param b          BAM alignment record
+ * @param state      The base modification state pointer.
+ * @param i          Modification index, from 0 to ntype-1
+ * @param strand     Boolean for top (0) or bottom (1) strand
+ * @param implicit   Boolean for whether unlisted positions should be
+ *                   implicitly assumed to be unmodified, or require an
+ *                   explicit score and should be considered as unknown.
+ *                   Returned.
+ * @param canonical  Canonical base type associated with this modification
+ *                   Returned.
+ *
+ * @return 0 on success or -1 if not found.  The strand, implicit and canonical
+ * fields are filled out if passed in as non-NULL pointers.
+ */
+HTSLIB_EXPORT
+int bam_mods_queryi(hts_base_mod_state *state, int i,
+                    int *strand, int *implicit, char *canonical);
+
 /// Returns the list of base modification codes provided for this
 /// alignment record as an array of character codes (+ve) or ChEBI numbers
 /// (negative).