Updates to variant cooccurrence page

[phildarnowsky-broad]

This incorporates a number of updates to the variant cooccurrence page:

• We change the thresholds at which a variant pair are considered to be in cis and in trans respectively
• Singletons predicted to be in cis no longer get predictions, nor does the haplotype count table display for them
• Variants that are in cis but that are distant from one another have an accuracy warning

There are also a good many refactorings added along the way, and some basic tests laid down.

[rileyhgrant]

LGTM! I like the pattern in the testing to define a base test object then change certain properties with the spreader syntax -- very cool!

Approved with minor changes of the usual variety.

• One tiny comment on importing a constant value to reuse it
• Linter should be appeased (which I see you're currently in the process of doing)
• The DONTMERGE commit should be dropped, which I know you know, but I figure another friendly reminder (given its sufficiently friendly) almost never hurts

Nicely done as usual Phil :). I'm loving the tests.

[rileyhgrant]

Very minor comment, should we import these thresholds from VariantCoocurrencePage.tsx if they're being exported from that file anyways? One less thing to update if they ever change again.

[phildarnowsky-broad]

So it's maybe a little counterintuitive, but when I have key constants like this which drive the logic, I like to define them separately for tests. The reason is that if you change the value in the future, a bunch of your tests will break, which will reveal some implications of changing the value that you need to think about.