Function roc_curve 

pub fn roc_curve(
    scores: &[f64],
    labels: &[bool],
) -> Result<RocCurveData, StatsError>

Compute an ROC curve from predicted scores and true binary labels (true = positive class).

Higher scores are treated as “more positive”. The sweep visits each distinct score as a threshold, emitting one point per distinct score (ties handled together), and always begins at (0, 0).

Errors

• StatsError::LengthMismatch if scores and labels differ in length.
• StatsError::EmptyInput if empty.
• StatsError::NoPositiveLabels / StatsError::NoNegativeLabels if a class is absent, since TPR or FPR would be 0/0.