test xxhash distribution

test_xxhash_distribution.go

package main

import (
	"bufio"
	"fmt"
	"math"
	"os"

	"github.com/cespare/xxhash"
	"gonum.org/v1/gonum/floats"
	"gonum.org/v1/gonum/stat"
)

func main() {
	scanner := bufio.NewScanner(os.Stdin)

	distribution := make([]float64, 10)

	for scanner.Scan() {
		line := scanner.Text()
		f := hashToFloat(line)

		distribution[int(math.Floor(f*10))]++
	}

	fmt.Println("\nBuckets:\n")

	for i := range distribution {
		f := float64(i)
		fmt.Printf("[%.1f, %.1f]  = %.1f\n", f/10, (f+1)/10, distribution[i])
	}

	stddev := stat.StdDev(distribution, nil)
	mean := stat.Mean(distribution, nil)
	stderr := stat.StdErr(stddev, floats.Sum(distribution))

	fmt.Println("\nTotal:\n")

	fmt.Printf("mean: %.2f ± %.2f\n", mean, stderr)
}

func hashToFloat(in string) float64 {
	h := xxhash.New()
	h.Write([]byte(in))

	return float64(h.Sum64()) / math.MaxUint64
}