Let pKa/redox helpers consume conformer ensembles via EnsembleThermo

[galjos]

`pKa`, `calibrate_proton_reference`, `reduction_potential`, and `reaction_free_energy` all duck-type on a single argument's `total_EeGtot()` -- they never inspect the type, just call that one method. `ensemble.py` already has proper Boltzmann-weighting (`boltzmann_weights`, `ensemble_free_energy`, `lowest_gibbs`), but `ensemble_free_energy` returns a bare float, so there was no way to plug a conformer ensemble directly into those helpers -- today's usage docs pick a single conformer (`generate(..., max_conformers=1)`) for exactly that reason.

The fix

`EnsembleThermo(thermos, temperature=298.15)` wraps `ensemble_free_energy` behind a `.total_EeGtot()` method, so it duck-types as a `Thermo` and can be passed anywhere a single `Thermo` is expected today -- no changes needed to pka.py or reactions.py themselves.

Verification

• Unit tests confirm `EnsembleThermo(...).total_EeGtot()` matches `ensemble_free_energy(..., unit="H")`, and that it drops into `pKa`/`reduction_potential` producing results identical to manually pre-reducing the ensemble.
• A real end-to-end test runs actual xtb-cli (GFN2-xTB) on 4-hydroxybutanoic acid / its conjugate base. Picked deliberately: smaller/more rigid acids (e.g. glycolic acid) collapse to a single RDKit conformer after RMSD pruning, which wouldn't actually exercise ensemble averaging -- 4-hydroxybutanoic acid's flexible C-C-C-C backbone gives 4-5 genuinely distinct conformers per species, so this is real Boltzmann averaging over real semiempirical energies, not a degenerate case.
• Full suite: 401 passed, 10 skipped with a real `xtb` binary; 397 passed, 14 skipped without (clean skip). 100% coverage on `ensemble.py`. Docs build clean.

Closes #118

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 97.31%. Comparing base (923a4c4) to head (434989d).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #119   +/-   ##
=======================================
  Coverage   97.31%   97.31%           
=======================================
  Files          32       32           
  Lines        1935     1940    +5     
=======================================
+ Hits         1883     1888    +5     
  Misses         52       52           

Flag	Coverage Δ
unittests	97.31% <100.00%> (+<0.01%)	⬆️

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.