Function wilson_interval 

pub fn wilson_interval(
    successes: usize,
    trials: usize,
    confidence_level: f64,
) -> MetricValue<ConfidenceInterval>

Computes a two-sided Wilson score interval for a binomial proportion.

confidence_level must be strictly between zero and one. A zero-trial input is reported as insufficient data rather than represented as 0 / 0.