Vectorise edge_effect using tskit 1.0 mutation arrays

[jeromekelleher]

Replace the Python loop over causal sites and their mutations with an array-based computation using ts.mutations_edge,
ts.mutations_inherited_state and ts.mutations_derived_state. Each causal site owns a contiguous run of mutation IDs, so the per-site boundaries come from a single searchsorted, the causal allele state changes are array comparisons, and the effects are accumulated with np.bincount. This is around 20 times faster: 1.44s -> 0.07s for 168k trait rows over a tree sequence with 338k edges.

Extract the trait dataframe validation shared by genetic_value and edge_effect into _check_trait_df, and add a site_id bounds check there. Both functions now raise a ValueError for a site_id outside [0, ts.num_sites), where a negative value previously wrapped around to the end of the site table.

Expand the edge_effect tests with a state-based reference implementation that is independent of the mutation walking approach, run over the tests/data.py tree sequences, all_trees_ts(2..5) with recurrent and back mutations, and simulations with and without recombination. Cross check against genetic_value(level="node") using the single tree identity edge_effect[e] == value[child(e)] - value[parent(e)], and cover multiple traits, shared causal sites, sites without mutations, stacked mutations, multi-character alleles, isolated nodes and multi-root forests.

Closes #190

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (c1d5dba) to head (6c9d01f).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff            @@
##              main      #193   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            9         9           
  Lines          417       417           
  Branches        56        50    -6     
=========================================
  Hits           417       417           

Flag	Coverage Δ
python-tests	100.00% <100.00%> (ø)

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

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[jeromekelleher]

Vectorised numpy version is about 20X faster than Python loops, so good enough for now. Also use some tskit 1.0 infrastructure to simplify the code.