Struct StreamingStats 

pub struct StreamingStats { /* private fields */ }

Online statistics accumulator based on Welford’s one-pass algorithm.

All operations are O(1) per data point (or O(n) for a batch update). After processing all data the accumulator exposes:

• count() – number of samples seen
• mean() – arithmetic mean
• variance() – Bessel-corrected sample variance
• std_dev() – sample standard deviation
• min() / max() – extremes

The merge method combines two accumulators in O(1) using Chan’s parallel formula, enabling map-reduce over partitioned data.

Implementations

impl StreamingStats

pub fn new() -> Self

Create a new, empty accumulator.

pub fn update(&mut self, value: f64)

Incorporate one more observation using Welford’s update rule.

NaN values are silently ignored.

pub fn update_batch(&mut self, values: &[f64])

Process a batch of values, equivalent to calling update for each element but potentially more cache-friendly.

pub fn count(&self) -> u64

Number of non-NaN observations accumulated so far.

pub fn mean(&self) -> f64

Arithmetic mean of all accumulated observations.

Returns 0.0 when count == 0.

pub fn variance(&self) -> f64

Bessel-corrected sample variance (divides by count - 1).

Returns 0.0 when count <= 1.

pub fn population_variance(&self) -> f64

Population variance (divides by count).

Returns 0.0 when count == 0.

pub fn std_dev(&self) -> f64

Sample standard deviation (square root of the sample variance).

Returns 0.0 when count <= 1.

pub fn min(&self) -> f64

Minimum value seen so far.

Returns f64::INFINITY when count == 0.

pub fn max(&self) -> f64

Maximum value seen so far.

Returns f64::NEG_INFINITY when count == 0.

pub fn range(&self) -> f64

Range (max − min).

Returns 0.0 when fewer than two distinct values have been seen.

pub fn merge(self, other: Self) -> Self

Merge two accumulators into one using Chan’s parallel update formula.

The result is mathematically equivalent to having accumulated all observations from both sources into a single accumulator.

This enables map-reduce patterns:

use scirs2_core::streaming_stats::StreamingStats;

let mut a = StreamingStats::new();
a.update_batch(&[1.0, 2.0, 3.0]);

let mut b = StreamingStats::new();
b.update_batch(&[4.0, 5.0, 6.0]);

let combined = a.merge(b);
assert!((combined.mean() - 3.5).abs() < 1e-12);

pub fn merge_all(stats: Vec<Self>) -> Self

Convenience: merge a slice of accumulators into one.

Returns StreamingStats::new() when the slice is empty.

Trait Implementations

impl Clone for StreamingStats

fn clone(&self) -> StreamingStats

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for StreamingStats

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

Formats the value using the given formatter.

impl Default for StreamingStats

fn default() -> Self

Returns the “default value” for a type.