cc-metric-store/pkg/resampler/resampler.go

package resampler

import (
	"errors"
	"fmt"
	"math"

	"github.com/ClusterCockpit/cc-metric-store/internal/util"
)

func SimpleResampler(data []util.Float, old_frequency int64, new_frequency int64) ([]util.Float, error) {
	if old_frequency == 0 || new_frequency == 0 {
		return nil, errors.New("either old or new frequency is set to 0")
	}

	if new_frequency%old_frequency != 0 {
		return nil, errors.New("new sampling frequency should be multiple of the old frequency")
	}

	var step int = int(new_frequency / old_frequency)
	var new_data_length = len(data) / step

	if new_data_length == 0 || len(data) < 100 || new_data_length >= len(data) {
		return data, nil
	}

	new_data := make([]util.Float, new_data_length)

	for i := 0; i < new_data_length; i++ {
		new_data[i] = data[i*step]
	}

	return new_data, nil
}

// Inspired by one of the algorithms from https://skemman.is/bitstream/1946/15343/3/SS_MSthesis.pdf
// Adapted from https://github.com/haoel/downsampling/blob/master/core/lttb.go
func LargestTriangleThreeBucket(data []util.Float, old_frequency int64, new_frequency int64) ([]util.Float, int64, error) {

	if old_frequency == 0 || new_frequency == 0 {
		return data, old_frequency, nil
	}

	if new_frequency%old_frequency != 0 {
		return nil, 0, fmt.Errorf("new sampling frequency : %d should be multiple of the old frequency : %d", new_frequency, old_frequency)
	}

	var step int = int(new_frequency / old_frequency)
	var new_data_length = len(data) / step

	if new_data_length == 0 || len(data) < 100 || new_data_length >= len(data) {
		return data, old_frequency, nil
	}

	new_data := make([]util.Float, 0, new_data_length)

	// Bucket size. Leave room for start and end data points
	bucketSize := float64(len(data)-2) / float64(new_data_length-2)

	new_data = append(new_data, data[0]) // Always add the first point

	// We have 3 pointers represent for
	// > bucketLow - the current bucket's beginning location
	// > bucketMiddle - the current bucket's ending location,
	//                  also the beginning location of next bucket
	// > bucketHight - the next bucket's ending location.
	bucketLow := 1
	bucketMiddle := int(math.Floor(bucketSize)) + 1

	var prevMaxAreaPoint int

	for i := 0; i < new_data_length-2; i++ {

		bucketHigh := int(math.Floor(float64(i+2)*bucketSize)) + 1
		if bucketHigh >= len(data)-1 {
			bucketHigh = len(data) - 2
		}

		// Calculate point average for next bucket (containing c)
		avgPointX, avgPointY := calculateAverageDataPoint(data[bucketMiddle:bucketHigh+1], int64(bucketMiddle))

		// Get the range for current bucket
		currBucketStart := bucketLow
		currBucketEnd := bucketMiddle

		// Point a
		pointX := prevMaxAreaPoint
		pointY := data[prevMaxAreaPoint]

		maxArea := -1.0

		var maxAreaPoint int
		flag_ := 0
		for ; currBucketStart < currBucketEnd; currBucketStart++ {

			area := calculateTriangleArea(util.Float(pointX), pointY, avgPointX, avgPointY, util.Float(currBucketStart), data[currBucketStart])
			if area > maxArea {
				maxArea = area
				maxAreaPoint = currBucketStart
			}
			if math.IsNaN(float64(avgPointY)) {
				flag_ = 1
			}
		}

		if flag_ == 1 {
			new_data = append(new_data, util.NaN) // Pick this point from the bucket

		} else {
			new_data = append(new_data, data[maxAreaPoint]) // Pick this point from the bucket
		}
		prevMaxAreaPoint = maxAreaPoint // This MaxArea point is the next's prevMAxAreaPoint

		//move to the next window
		bucketLow = bucketMiddle
		bucketMiddle = bucketHigh
	}

	new_data = append(new_data, data[len(data)-1]) // Always add last

	return new_data, new_frequency, nil
}