Struct stream_histogram::Histogram

pub struct Histogram { /* fields omitted */ }

A Histogram struct include a double linklist and some attributes for manage data. using linkedList for fast insert and merge items in a sorted data structure

Implementations

impl Histogram

pub fn new(max: usize) -> Histogram

Histogram is a data structure for calculation some statistic value from stream data, for example the mean、max、min、 q-quantiles and cumulative distribution etc. You must set the bucket number first, greater bucket number will bring more precision， but also bring more load for process. Default is 10, any number less than 10 will reset to default.

Examples

 use crate::histogram::Histogram;
 let mut histogram = Histogram::new(20);
 for i in 1..=100000 {
      histogram.add(i as f64);
 }

pub fn add(&mut self, number: f64)

add method update histogram data structure from stream each time update a new value current value must be f64 after update, the inner process will check the bucket number and merge two smallest gap bucket into one bucket.

Examples

 use crate::histogram::Histogram;
 let mut histogram = Histogram::new(20);
 for i in 1..=100000 {
      histogram.add(i as f64);
 }

pub fn quantile(&self, q: f64) -> Option<f64>

q-quantiles are values that partition a finite set of values into q subsets of (nearly) equal sizes. for more detail please check wikipedia

pub fn cdf(&mut self, x: f64) -> Option<f64>

Cumulative distribution function(aka: cdf) returns the value of the cumulative distribution at value x. for more detail, please check wikipedia

pub fn mean(&self) -> Option<f64>

pub fn variance(&self) -> Option<f64>

pub fn to_string(&self) -> String

to_string method print the result in a simple visulazation style, each line print the bucket and the size in a dot mode.

Examples

 use rand::distributions::{Normal, Distribution};
 use crate::histogram::Histogram;
 let normal = Normal::new(10.0, 10.0);
 let mut histogram = Histogram::new(20);
 for _i in 1..=100000 {
      let v = normal.sample(&mut rand::thread_rng());
      histogram.add(v as f64);
 }
 println!("{}", histogram.to_string());

pub fn report(&self) -> Option<HistogramReport>

return a report from current histogram

Example

use crate::histogram::Histogram;
let mut histogram = Histogram::new(10);
for i in 1..=100 {
    histogram.add(i as f64);
}
if let Some(report) = histogram.report(){
   println!("{:?}", report);
}

Trait Implementations

impl Default for Histogram

pub fn default() -> Self

Returns the "default value" for a type.