Iterate alleles and genotypes at the same time

[benjeffery]

In preparation for tskit.
Stacked on #362 clean diff at 33d1d31

[coveralls]

coverage: 98.015% (+0.04%) from 97.976%
when pulling f19420c on benjeffery:iterate_alleles_genotypes
into d30940e on sgkit-dev:main.

[jeromekelleher]

Looks good overall, but I don't like the include_genotypes parameter - the only reason we wouldn't include genotypes is if they don't exist, right? So we can just check this, rather than passing responsibility to client code.

[benjeffery]

Looks good overall, but I don't like the include_genotypes parameter - the only reason we wouldn't include genotypes is if they don't exist, right? So we can just check this, rather than passing responsibility to client code.

The edge case here is if genotypes exist, but they aren't in the schema. If we're ok with wasting work fetching them in this odd case then we can remove it.

[jeromekelleher]

The edge case here is if genotypes exist, but they aren't in the schema. If we're ok with wasting work fetching them in this odd case then we can remove it.

I'm fine with that

[benjeffery]

Ok, should be good to go now.

[jeromekelleher]

I think we're over complicating this - can't you define the VCF case using the existing functions?

[jeromekelleher]

What's the continue for?

[jeromekelleher]

This is too much in one zip - let's break it up into simpler bits

What's wrong with

if self.gt_field is None:
     for alleles in self.iter_alleles(start, stop):
            yield alleles, None, None
else:
        yield from zip(self.iter_alleles(start, stop), self.iter_genotypes(start, stop)):

and just keep the old definitions?

[benjeffery]

Comments addressed!