Struct RankReport 

pub struct RankReport {
    pub rows: Vec<RankRow>,
    pub n_cols: usize,
    pub singular_values: Vec<Number>,
    pub tol: Number,
    pub rank: usize,
    pub cond: Number,
    pub culprits: Vec<RankCulprit>,
}

Numerical rank diagnosis of the active-constraint Jacobian at one iterate. See the module docs for the underlying SVD analysis.

Fields

rows: Vec<RankRow>

The active rows analyzed, in block order (currently the equality block J_c). rows[i] is the identity of dense row i.

n_cols: usize

Number of variables (columns of the analyzed block).

singular_values: Vec<Number>

Singular values, nonincreasing.

tol: Number

Numerical-rank threshold τ = σ_max · max(m, n) · ε.

rank: usize

Numerical rank (count of σ > τ).

cond: Number

Condition number σ_max / σ_min (∞ if σ_min == 0).

culprits: Vec<RankCulprit>

Rows participating in the near-null space, sorted by weight descending. Empty when the block has full row rank.

Implementations

impl RankReport

pub fn n_rows(&self) -> usize

Number of active rows analyzed (m).

pub fn deficiency(&self) -> usize

Row-rank deficiency m − rank — the number of dependent rows.

pub fn is_rank_deficient(&self) -> bool

Whether the analyzed block is (numerically) row-rank-deficient.

pub fn sigma_min(&self) -> Number

Smallest singular value (0 if there are none).

pub fn sigma_max(&self) -> Number

Largest singular value (0 if there are none).

Trait Implementations

impl Clone for RankReport

fn clone(&self) -> RankReport

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for RankReport

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.