Parse CIGAR string in C/C++

[tedwong]

What'd be the best way to parse a CIGAR string fully according to the specification in C/C++? Would regular expression work?

[Brian Bushnell]

What'd be the best way to parse a CIGAR string fully according to the specification in C/C++? Would regular expression work?

No. Unless you simply want to detect the presence of some operation, the best way is with a custom loop.

Here's an example in Java that can easily be translated to C++:

http://seqanswers.com/forums/showthread.php?t=51162

[dpryan]

If you're using C/C++ already, then just use htslib. The functions for this have already been written (afterall, it's what samtools uses) and the API is generally convenient.

[lindenb]

cross posted: http://stackoverflow.com/questions/2...lar-expression

[student-t]

For reference (using the htslib library)



#include <htslib/sam.h>

auto f = sam_open(file.c_str(), "r");
auto h = sam_hdr_read(f);
auto t = bam_init1();

while (sam_read1(f, h, t) >= 0)
{
auto id = std::string(h->target_name[0]);
auto mapped = !(t->core.flag & BAM_FUNMAP);

const auto cigar = bam_get_cigar(t);

for (int k = 0; k < t->core.n_cigar; k++)
{
const int op = bam_cigar_op(cigar[k]);
const int ol = bam_cigar_oplen(cigar[k]);

if (op == BAM_CMATCH || op == BAM_CINS || op == BAM_CDEL)
{
// your code, you have the length in ol (eg: 101M -> ol == 101)
}
}
}

sam_close(f);