fix: sample_from_fva_output crashes with NameError on every call

[xovishnukosuri]

Summary

PolytopeSampler.sample_from_fva_output is a @staticmethod but contained two bugs that made it crash with NameError on every call:

• It referenced self._parameters["tol"] (lines 294-295). There is no self in a static method.
• It called get_matrices_of_low_dim_polytope(S, min_fluxes, max_fluxes, opt_percentage, tol) where tol was also undefined, and the argument order did not match the function's signature (S, lb, ub, min_fluxes, max_fluxes).

Fix

• Define tol = 1e-06 locally, matching the default in PolytopeSampler.__init__.
• Pass min_fluxes and max_fluxes as the lb/ub arguments to get_matrices_of_low_dim_polytope. When FVA output is the input, the FVA bounds serve as both the flux bounds and the tight bounds for equality detection — the same pattern used in get_polytope().

Reproducer

from dingo import MetabolicNetwork, PolytopeSampler

model = MetabolicNetwork.from_json("ext_data/e_coli_core.json")
min_fluxes, max_fluxes, flux_vector, objective = model.fva()

# Before this fix: NameError: name 'tol' is not defined
steady_states = PolytopeSampler.sample_from_fva_output(
    min_fluxes, max_fluxes,
    model.objective_function, objective,
    model.S,
)

Test plan

• Run python -m pytest tests/fba.py tests/sampling.py to confirm no regressions
• Call PolytopeSampler.sample_from_fva_output with e_coli_core FVA output and confirm it no longer raises NameError

🤖 Generated with Claude Code