Fix TS multiplicity from YAML input

[alongd]

Issue

A TS species defined in input.yml with only is_ts: true and xyz (no smiles/adjlist) could be assigned the wrong multiplicity, causing Gaussian to reject the input with GL301. For example, a closed-shell TS with 24 electrons was written as 0 2 (mult=2), which is electronically impossible.

Root cause

ARCSpecies.from_dict (the path taken when main.py loads species from YAML via ARCSpecies(species_dict=spc)) called mol_from_xyz to perceive a 2D graph from the TS geometry, then blindly assigned self.multiplicity = self.mol.multiplicity. RMG's perception of TS-like geometries (partial bonds, near-radical atoms) routinely returns spurious biradical structures — for the failing case it returned O=[C][H][OH] with mult=2. Direct construction (ARCSpecies(label=..., is_ts=True, xyz=...)) was unaffected because init routes through determine_multiplicity, which skips the descriptor path for TSs and uses electron parity from xyz.

Fix

In from_dict, when is_ts=True and xyz is available, fall back to determine_multiplicity_from_xyz (electron-parity) instead of trusting the perceived mol.multiplicity. Non-TS behavior is unchanged; TSs without xyz still fall through to mol.multiplicity since there's no parity signal.

Test

Added a dict-construction assertion to test_determine_multiplicity for the failing TS xyz — this is the path the bug actually went through. The pre-existing direct-construction assertion was insufficient because it bypassed from_dict entirely.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 60.40%. Comparing base (f30f9cc) to head (85b76f7).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #882      +/-   ##
==========================================
- Coverage   60.42%   60.40%   -0.02%     
==========================================
  Files         102      102              
  Lines       31096    31098       +2     
  Branches     8103     8104       +1     
==========================================
- Hits        18791    18786       -5     
- Misses       9961     9969       +8     
+ Partials     2344     2343       -1     

Flag	Coverage Δ
functionaltests	60.40% <ø> (-0.02%)	⬇️
unittests	60.40% <ø> (-0.02%)	⬇️

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

☔ View full report in Codecov by Sentry.
📢 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.