rxnfit.lsoda_p_vector module

Construct the numbalsoda parameter vector p for LSODA integration.

The ordering rules here are part of the public ABI and must mirror the Numba RHS compiled in rxnfit.numbalsoda_rhs.

rxnfit.lsoda_p_vector.build_lsoda_parameter_vector_p(rate_const_values: Mapping[str, float], symbolic_rate_const_keys: Sequence[str]) → ndarray

Assemble the numbalsoda parameter vector p from rate-constant values.

Parameter layout for numbalsoda RHS (must match numbalsoda_rhs cfunc):

  • Let n_fitted = len(symbolic_rate_const_keys).

  • p[0] … p[n_fitted-1]: values for symbolic_rate_const_keys in that exact order (same order as create_system_rhs / fitters).

  • p[n_fitted] … p[n_p-1]: every other key present in rate_const_values, excluding those in symbolic_rate_const_keys, ordered by lexicographic sort of the key strings (not natural sort).

  • n_p = len(p) is fixed for a given RHS / build.

Example: symbolic_rate_const_keys = ("k2", "k1") and the only remaining keys in rate_const_values are k3 and k4 → p = [v_k2, v_k1, v_k3, v_k4].

The concentration vector u (species order / function_names) does not need to align with the ordering of entries in p.

Parameters:

  • rate_const_values – All rate constants needed for the current integration (numeric).

  • symbolic_rate_const_keys – Fitted (or explicit) keys in the order used by create_system_rhs for extra ODE arguments.

Returns:

One-dimensional float64 ndarray of length n_p.