Struct ComplexityEstimator 

pub struct ComplexityEstimator<const W: usize> { /* private fields */ }

Windowed complexity estimator using normalized entropy.

Maintains a rolling histogram of residual norms quantized into BINS bins. Computes Shannon entropy H and normalizes by log₂(BINS) to produce a score in [0, 1]:

• 0.0: all observations fall in one bin (maximally compressible)
• 1.0: uniform distribution across bins (maximally incompressible)

Implementations

impl<const W: usize> ComplexityEstimator<W>

pub fn new(max_val: f32) -> Self

Create a new estimator.

max_val = maximum residual norm for binning. Values above this are clamped to the last bin. Typically set to 2ρ.

pub fn push(&mut self, norm: f32) -> ComplexityResult

Push a residual norm and return the current complexity estimate.

pub fn reset(&mut self)

Reset the estimator.