LS model refactoring

[jeromekelleher]

Some refactoring of the LS hmm code. WIP

based on #2786 (so has a bunch of extra commits).

[codecov]

Codecov Report

Merging #2799 (64aedbf) into main (11dd2f5) will decrease coverage by 0.19%.
The diff coverage is 81.42%.

@@            Coverage Diff             @@
##             main    #2799      +/-   ##
==========================================
- Coverage   89.97%   89.78%   -0.19%     
==========================================
  Files          30       30              
  Lines       29322    29447     +125     
  Branches     5707     5731      +24     
==========================================
+ Hits        26383    26440      +57     
- Misses       1669     1723      +54     
- Partials     1270     1284      +14     

Flag	Coverage Δ
c-tests	86.08% <74.50%> (-0.42%)	⬇️
lwt-tests	80.78% <ø> (ø)
python-c-tests	67.89% <100.00%> (+0.07%)	⬆️
python-tests	99.02% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Files Changed	Coverage Δ
c/tskit/haplotype_matching.c	84.47% <74.50%> (-6.20%)	⬇️
python/_tskitmodule.c	88.68% <100.00%> (+0.06%)	⬆️

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 11dd2f5...64aedbf. Read the comment docs.

[jeromekelleher]

@astheeggeggs I've made some progress here on refactoring the Python tree LS HMM code, mainly to deduplicate things a bit and also to remove the need for mirroring the coords of the ts. I think it's pretty close to the point where I can implement it in C.

It would be helpful if you could take a look at the code at this point to see if you spot any howlers (just the test_haplotype_matching.py code, everthing else is incidental). I've made a start on refactoring the tests a bit as well, to move away from the current class-based approach where we internally iterate over a bunch of parameters, towards something that's much more modular, and where you can run a single example with a single set of parameters easily. It's a WIP though

[jeromekelleher]

I think this is breaking on some lshmm bugs that have been fixed, so the simplest thing is to wait for version 0.0.5 (astheeggeggs/lshmm#25)

I've added a commit to bump the requirements.