pympcc.jac_norms

pympcc.jac_norms(result, problem, *, tol=1e-06, zero_tol=1e-10)[source]

Row- and column-norm summary of the active-constraint Jacobian.

Builds the same active-gradient stack used by classify_cq() (rows: equality, comp G/H on the active sides, active inequality, active variable bounds) and returns row/column norm extrema plus near-zero counts.

Near-zero rows or columns are usually degeneracy signals (a row at norm ≈ 0 contributes no constraint information at the iterate; a column at norm ≈ 0 means a variable has no influence on any active constraint).

Returns:

Schema:

{
    "row": {"max": ..., "min": ..., "n_zero": ..., "n_rows": ...},
    "col": {"max": ..., "min": ..., "n_zero": ..., "n_cols": ...},
}

Each inner dict carries empty values (None / 0) when the active matrix is empty or the result did not converge.

Return type:

dict

Parameters: