C++ module for on-the-fly Statistics

Overview

PyPi module	N/A
git repository	https://bitbucket.org/arrizza-public/cpp-on-the-fly-stats
git command	git clone git@bitbucket.org:arrizza-public/cpp-on-the-fly-stats.git
Verification Report	https://arrizza.com/web-ver/cpp-on-the-fly-stats-report.html
Version Info	macOS 14.5, Python 3.10 Ubuntu 20.04 focal, Python 3.10 Ubuntu 22.04 jammy, Python 3.10 Ubuntu 24.04 noble, Python 3.10 win32 Windows 10, Python 3.10

installation: see https://arrizza.com/setup-common

Summary

This C++ library contains various statistic functions (e.g. standard deviation) that can be updated on the fly i.e. the entire dataset is not needed up front.

For a Python version see on-the-fly-stats

./doit

Typical code:

auto otfs1 = OTFStats("ofts1");   # <== created a new set of stats
for (int i = 1; i <= 10; ++i) {
    otfs1.update((double)i);      # <== adds a new value to on-the-fly stream of values 
    otfs1.report(writeln);        # <== write a formatted line of values
}
printf("%s: avg=%5.2f stddev=%5.2f\n", otfs1.tag().c_str(), otfs1.average(), otfs1.stddev());

Typical output:

 --  2] ofts1                         1        1.000000        1.000000        1.000000             nan
 --  3] ofts1                         2        1.000000        2.000000        1.500000        0.707107
 --  4] ofts1                         3        1.000000        3.000000        2.000000        1.000000
 --  5] ofts1                         4        1.000000        4.000000        2.500000        1.290994
 --  6] ofts1                         5        1.000000        5.000000        3.000000        1.581139
 --  7] ofts1                         6        1.000000        6.000000        3.500000        1.870829
 --  8] ofts1                         7        1.000000        7.000000        4.000000        2.160247
 --  9] ofts1                         8        1.000000        8.000000        4.500000        2.449490
 -- 10] ofts1                         9        1.000000        9.000000        5.000000        2.738613
 -- 11] ofts1                        10        1.000000       10.000000        5.500000        3.027650
 -- 12] ofts1: avg= 5.50 stddev= 3.03

You can maintain multiple counters using OTFList. Typical code:

#include "otf_list.h"
// snip

OTFList stats;
stats.create("stats1");
stats.create("stats2");
stats.create("stats3");
stats.get("stats1")->update(1.0);
stats.get("stats2")->update(8.0);
stats.get("stats3")->update(22.0);
stats.report(writeln);


// defines a callback that prints a line from the report
// this can be used to write to a file or stdout as needed
// see ut/ut_tp002_report for an example of writing to an internal vector of lines 
static void writeln(const std::string& s)
{
    std::cout << s << std::endl;
}

Typical output:

 -- 17] Name                      Count             Min             Max         Average         Std Dev
 -- 18] --------------- --------------- --------------- --------------- --------------- ---------------
 -- 19] stats1                        1        1.000000        1.000000        1.000000             nan
 -- 20] stats2                        1        8.000000        8.000000        8.000000             nan
 -- 21] stats3                        1       22.000000       22.000000       22.000000             nan