AlignedSegment.set_tag(): validate Hex values for tag type 'H'

[bentyeh]

Added validation for values in Hex format tags, using regular expression defined in the SAM specification.

---

Currently, set_tag() does not perform any validation of the tag value for Hex format tags (where value_type (typecode) is 'H'). This means that pysam allows writing reads to a SAM/BAM file that may result in errors when trying to read that same read via pysam.AlignmentFile.fetch().

Specifically, pysam.AlignmentFile.fetch() (via ... via pysam.AlignmentFile.cnext() via sam.c:sam_read1() via ... via sam.c:aux_parse() in htslib) checks that Hex format tag values contain an even number of characters.

pysam/htslib/sam.c

Lines 2835 to 2839 in 2f9d50d

	} else if (type == 'Z' || type == 'H') {
	char *end = strchr(q, '\t');
	if (!end) end = q + strlen(q);
	_parse_err(type == 'H' && ((end-q)&1) != 0,
	"hex field does not have an even number of digits");

Therefore, the proposed patch (which follows the SAM specification) for the write side of the pysam interface will be a bit stricter than what the read side checks for. Nonetheless, this helps ensure round-trip compatibility of reads written by pysam.