Fix ParameterValues.to_json for create_from_bpx partials

[rtimms]

Description

pybamm.ParameterValues.create_from_bpx(...).to_json() raised TypeError: ... got multiple values for argument 'D_ref' (and similar for OCP interpolants / exchange-current density).

create_from_bpx stores several functional parameters as functools.partial objects with keyword arguments already bound, e.g.:

functools.partial(_diffusivity, D_ref=..., Ea=..., constant=True)  # def _diffusivity(sto, T, D_ref, Ea, constant=False)
functools.partial(_interpolant_func, name=..., x=..., y=...)       # OCP interpolants

In convert_function_to_symbolic_expression, the partial branch built its argument list from the full partial signature — which still lists the bound and defaulted parameters (D_ref, Ea, constant) — then called the partial positionally for every one of them. The already-bound kwargs then collided with the positional values.

Fix

Only create symbolic Parameters for the parameters that still lack a value (no default and not variadic) — the genuine symbolic inputs (sto, T, …). Bound kwargs and defaulted control flags (e.g. constant, which is used for Python-level if branching, not as a symbol) keep their values.

Note the subtlety: filtering only on func.keywords is insufficient — a defaulted-but-unbound parameter (constant in the callable-diffusivity case) would still be turned into a symbol and shift positionally into a bound parameter's slot, re-triggering the same collision.

Also removes the dead get_name/get_args branch, which required an object exposing get_name(), get_args(), and .func simultaneously — a type that has never existed in the codebase or its history.

Tests

• TestConvertFunctionToSymbolicExpression in test_serialisation.py — kwarg-bound partials returning scalars and interpolants, plus the defaulted-unbound-arg regression.
• Parametrized test_bpx_to_json_roundtrip (function / constant / table forms) in test_bpx.py — full create_from_bpx → to_json → from_json round-trip.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.11%. Comparing base (e929c3c) to head (5527ae7).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #5641      +/-   ##
==========================================
+ Coverage   98.10%   98.11%   +0.01%     
==========================================
  Files         339      339              
  Lines       32069    32061       -8     
==========================================
- Hits        31460    31457       -3     
+ Misses        609      604       -5     

☔ 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.