Struct MetricStats 

pub struct MetricStats {
    pub mean: f64,
    pub std_dev: f64,
    pub ci_half_width: f64,
    pub ci_lower: f64,
    pub ci_upper: f64,
    pub min: f64,
    pub max: f64,
    pub n: usize,
}

Statistical metrics with variance and confidence intervals.

This is the canonical type for evaluation metrics that need to track statistical properties across multiple runs or samples.

Features

• Mean, standard deviation, and min/max range
• 95% confidence interval (both ± and lower/upper bounds)
• Sample count for determining statistical significance
• Formatting helpers for display

Examples

use anno_core::types::MetricStats;

let stats = MetricStats::from_samples(&[0.85, 0.87, 0.82, 0.88, 0.84]);
println!("F1: {}", stats.format_ci());  // "85.2% ± 2.1%"
println!("Range: {}", stats.format_range());  // "85.2% (82.0%-88.0%)"

Fields

mean: f64

Mean value of the metric

std_dev: f64

Standard deviation (sample, Bessel-corrected)

ci_half_width: f64

95% confidence interval half-width (±)

ci_lower: f64

95% CI lower bound

ci_upper: f64

95% CI upper bound

min: f64

Minimum observed value

max: f64

Maximum observed value

n: usize

Number of samples

Implementations

impl MetricStats

pub fn from_samples(samples: &[f64]) -> MetricStats

Create from a slice of sample values.

Uses sample standard deviation (Bessel’s correction) and t-distribution approximation for 95% CI.

Examples

use anno_core::types::MetricStats;

let stats = MetricStats::from_samples(&[0.80, 0.82, 0.85, 0.83, 0.80]);
assert!((stats.mean - 0.82).abs() < 0.01);
assert_eq!(stats.n, 5);

pub fn from_single(value: f64) -> MetricStats

Create from a single value (no variance).

pub fn format_ci(&self) -> String

Format as “mean ± ci” string (percentages).

Examples

use anno_core::types::MetricStats;

let stats = MetricStats::from_samples(&[0.85, 0.87, 0.83]);
println!("{}", stats.format_ci());  // e.g., "85.0% ± 2.3%"

pub fn format_range(&self) -> String

Format as “mean (min-max)” string (percentages).

pub fn format_ci_bounds(&self) -> String

Format as “mean [ci_lower, ci_upper]” string (percentages).

pub fn coefficient_of_variation(&self) -> f64

Get coefficient of variation (CV = std_dev / mean).

Lower CV indicates more stable/consistent results.

• CV < 0.05: Excellent stability
• CV < 0.10: Good stability
• CV < 0.20: Moderate stability
• CV >= 0.20: High variance, investigate

pub fn is_stable(&self, cv_threshold: f64) -> bool

Check if results are statistically stable.

Returns true if coefficient of variation is below threshold.

pub fn ci_95_tuple(&self) -> (f64, f64)

Get 95% CI as a tuple (lower, upper).

For compatibility with code expecting tuple representation.

pub fn merge(&self, other: &MetricStats) -> MetricStats

Merge with another MetricStats (combine samples).

Uses Welford’s online algorithm for numerically stable combination.

Trait Implementations

impl Clone for MetricStats

fn clone(&self) -> MetricStats

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for MetricStats

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

Formats the value using the given formatter.

impl Default for MetricStats

fn default() -> MetricStats

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for MetricStats

fn deserialize<__D>( __deserializer: __D, ) -> Result<MetricStats, <__D as Deserializer<'de>>::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Display for MetricStats

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

Formats the value using the given formatter.

impl PartialEq for MetricStats

fn eq(&self, other: &MetricStats) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for MetricStats

fn serialize<__S>( &self, __serializer: __S, ) -> Result<<__S as Serializer>::Ok, <__S as Serializer>::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl Copy for MetricStats

impl StructuralPartialEq for MetricStats