Struct DefectPredictor 

pub struct DefectPredictor { /* private fields */ }

Defect predictor using linear model (placeholder for GradientBoosting)

When ml feature is enabled, uses aprender’s GradientBoostingClassifier. Otherwise uses a simple linear model as fallback.

Implementations

impl DefectPredictor

pub fn new() -> Self

Create new defect predictor

pub fn with_weights(weights: CategoryWeights) -> Self

Create with custom category weights

pub fn train(&mut self, samples: &[DefectSample]) -> Result<()>

Train on samples (linear model fallback)

Uses gradient descent to optimize feature weights.

Errors

Returns an error if samples is empty.

pub fn predict(&self, features: &CommitFeatures, code: &str) -> DefectPrediction

Predict defect probability for code

pub fn predict_features(&self, features: &CommitFeatures) -> DefectPrediction

Predict for features only (without code for category)

pub fn stats(&self) -> &DefectPredictorStats

Get training statistics

pub fn is_trained(&self) -> bool

Is model trained?

pub fn category_weights(&self) -> &CategoryWeights

Get category weights

pub fn prioritize(&self, samples: &[(CommitFeatures, String)]) -> Vec<usize>

Prioritize samples by defect likelihood

Returns indices sorted by priority (highest defect probability first)

Trait Implementations

impl Debug for DefectPredictor

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

Formats the value using the given formatter.

impl Default for DefectPredictor

fn default() -> Self

Returns the “default value” for a type.