fix DecodeCIGAR argument hint

add test to ensure CIGAR correctness in the case of different traceback results,
add DecodeCIGAR function to exports

Makefile

@@ -3,7 +3,7 @@
 build: clean
 	@echo "======================== Building Binary ======================="
 	minify wfa.js > dist/wfa.js
-	GOOS=js GOARCH=wasm CGO_ENABLED=0 tinygo build -no-debug -opt=2 -target=wasm -o dist/wfa.wasm .
+	GOOS=js GOARCH=wasm CGO_ENABLED=0 tinygo build -panic=trap -no-debug -opt=2 -target=wasm -o dist/wfa.wasm .
 
 clean:
 	@echo "======================== Cleaning Project ======================"

go.mod

@@ -2,10 +2,14 @@ module wfa
 
 go 1.23.2
 
+require (
+	github.com/schollz/progressbar/v3 v3.17.0
+	golang.org/x/exp v0.0.0-20241009180824-f66d83c29e7c
+)
+
 require (
 	github.com/mitchellh/colorstring v0.0.0-20190213212951-d06e56a500db // indirect
 	github.com/rivo/uniseg v0.4.7 // indirect
-	github.com/schollz/progressbar/v3 v3.16.1 // indirect
 	golang.org/x/sys v0.26.0 // indirect
 	golang.org/x/term v0.25.0 // indirect
 )

main.go

@@ -9,6 +9,7 @@ import (
 func main() {
 	c := make(chan bool)
 	js.Global().Set("wfAlign", js.FuncOf(wfAlign))
+	js.Global().Set("DecodeCIGAR", js.FuncOf(DecodeCIGAR))
 	<-c
 }
 
@@ -70,3 +71,21 @@ func wfAlign(this js.Value, args []js.Value) interface{} {
 
 	return js.ValueOf(resultMap)
 }
+
+func DecodeCIGAR(this js.Value, args []js.Value) interface{} {
+	if len(args) != 1 {
+		fmt.Println("invalid number of args, requires 1: CIGAR")
+		return nil
+	}
+
+	if args[0].Type() != js.TypeString {
+		fmt.Println("CIGAR should be a string")
+		return nil
+	}
+
+	CIGAR := args[0].String()
+
+	decoded := wfa.RunLengthDecode(CIGAR)
+
+	return js.ValueOf(decoded)
+}

pkg/custom_slice.go

@@ -1,65 +1,5 @@
 package wfa
 
-type IntegerSlice[T any] struct {
-	data         []T
-	valid        []bool
-	defaultValue T
-}
-
-func (a *IntegerSlice[T]) TranslateIndex(idx int) int {
-	if idx >= 0 { // 0 -> 0, 1 -> 2, 2 -> 4, 3 -> 6, ...
-		return 2 * idx
-	} else { // -1 -> 1, -2 -> 3, -3 -> 5, ...
-		return (-2 * idx) - 1
-	}
-}
-
-func (a *IntegerSlice[T]) Valid(idx int) bool {
-	actualIdx := a.TranslateIndex(idx)
-	return 0 <= actualIdx && actualIdx < len(a.valid) && a.valid[actualIdx]
-}
-
-func (a *IntegerSlice[T]) Get(idx int) T {
-	actualIdx := a.TranslateIndex(idx)
-	if 0 <= actualIdx && actualIdx < len(a.valid) && a.valid[actualIdx] { // idx is in the slice
-		return a.data[actualIdx]
-	} else { // idx is out of the slice
-		return a.defaultValue
-	}
-}
-
-func (a *IntegerSlice[T]) Set(idx int, value T) {
-	actualIdx := a.TranslateIndex(idx)
-	if actualIdx >= len(a.valid) { // idx is outside the slice
-		// expand data array to actualIdx
-		newData := make([]T, 2*actualIdx+1)
-		copy(newData, a.data)
-		a.data = newData
-
-		// expand valid array to actualIdx
-		newValid := make([]bool, 2*actualIdx+1)
-		copy(newValid, a.valid)
-		a.valid = newValid
-	}
-
-	a.data[actualIdx] = value
-	a.valid[actualIdx] = true
-}
-
-func (a *IntegerSlice[T]) Preallocate(lo int, hi int) {
-	actualLo := a.TranslateIndex(lo)
-	actualHi := a.TranslateIndex(hi)
-	size := max(actualHi, actualLo)
-
-	// expand data array to actualIdx
-	newData := make([]T, size+1)
-	a.data = newData
-
-	// expand valid array to actualIdx
-	newValid := make([]bool, size+1)
-	a.valid = newValid
-}
-
 type PositiveSlice[T any] struct {
 	data         []T
 	valid        []bool

pkg/debug.go

@@ -0,0 +1,79 @@
+package wfa
+
+import (
+	"fmt"
+	"math"
+)
+
+func (w *WavefrontComponent) String(score int) string {
+	traceback_str := []string{"OI", "EI", "OD", "ED", "SB", "IN", "DL", "EN"}
+	s := "<"
+	min_lo := math.MaxInt
+	max_hi := math.MinInt
+
+	for i := 0; i <= score; i++ {
+		if w.lo.Valid(i) && w.lo.Get(i) < min_lo {
+			min_lo = w.lo.Get(i)
+		}
+		if w.hi.Valid(i) && w.hi.Get(i) > max_hi {
+			max_hi = w.hi.Get(i)
+		}
+	}
+
+	for k := min_lo; k <= max_hi; k++ {
+		s = s + fmt.Sprintf("%02d", k)
+		if k < max_hi {
+			s = s + "|"
+		}
+	}
+
+	s = s + ">\t<"
+
+	for k := min_lo; k <= max_hi; k++ {
+		s = s + fmt.Sprintf("%02d", k)
+		if k < max_hi {
+			s = s + "|"
+		}
+	}
+
+	s = s + ">\n"
+
+	for i := 0; i <= score; i++ {
+		s = s + "["
+		lo := w.lo.Get(i)
+		hi := w.hi.Get(i)
+		// print out wavefront matrix
+		for k := min_lo; k <= max_hi; k++ {
+			valid, val, _ := UnpackWavefrontValue(w.W.Get(i).Get(k))
+			if valid {
+				s = s + fmt.Sprintf("%02d", val)
+			} else if k < lo || k > hi {
+				s = s + "--"
+			} else {
+				s = s + "  "
+			}
+
+			if k < max_hi {
+				s = s + "|"
+			}
+		}
+		s = s + "]\t["
+		// print out traceback matrix
+		for k := min_lo; k <= max_hi; k++ {
+			valid, _, tb := UnpackWavefrontValue(w.W.Get(i).Get(k))
+			if valid {
+				s = s + traceback_str[tb]
+			} else if k < lo || k > hi {
+				s = s + "--"
+			} else {
+				s = s + "  "
+			}
+
+			if k < max_hi {
+				s = s + "|"
+			}
+		}
+		s = s + "]\n"
+	}
+	return s
+}

pkg/types.go

@@ -1,10 +1,5 @@
 package wfa
 
-import (
-	"fmt"
-	"math"
-)
-
 type Result struct {
 	Score int
 	CIGAR string
@@ -17,10 +12,10 @@ type Penalty struct {
 	E int
 }
 
-type traceback byte
+type Traceback byte
 
 const (
-	OpenIns traceback = iota
+	OpenIns Traceback = iota
 	ExtdIns
 	OpenDel
 	ExtdDel
@@ -30,19 +25,87 @@ const (
 	End
 )
 
-type WavefrontComponent struct {
-	lo *PositiveSlice[int]                      // lo for each wavefront
-	hi *PositiveSlice[int]                      // hi for each wavefront
-	W  *PositiveSlice[*IntegerSlice[int]]       // wavefront diag distance for each wavefront
-	A  *PositiveSlice[*IntegerSlice[traceback]] // compact CIGAR for backtrace for each wavefront
+// bitpacked wavefront values with 1 valid bit, 3 traceback bits, and 28 bits for the diag distance
+// technically this restricts to solutions within 268 million score but that should be sufficient for most cases
+type WavefrontValue uint32
+
+// TODO: add 64 bit packed value in case more than 268 million score is needed
+
+// PackWavefrontValue: packs a diag value and traceback into a WavefrontValue
+func PackWavefrontValue(value uint32, traceback Traceback) WavefrontValue {
+	valueBM := value & 0x0FFF_FFFF
+	tracebackBM := uint32(traceback&0x0000_0007) << 28
+	return WavefrontValue(0x8000_0000 | valueBM | tracebackBM)
 }
 
+// UnpackWavefrontValue: opens a WavefrontValue into a valid bool, diag value and traceback
+func UnpackWavefrontValue(wfv WavefrontValue) (bool, uint32, Traceback) {
+	valueBM := uint32(wfv & 0x0FFF_FFFF)
+	tracebackBM := uint8(wfv & 0x7000_0000 >> 28)
+	validBM := wfv&0x8000_0000 != 0
+	return validBM, valueBM, Traceback(tracebackBM)
+}
+
+// Wavefront: stores a single wavefront, stores wavefront's lo value and hi is naturally lo + len(data)
+type Wavefront struct { // since wavefronts store diag distance, they should never be negative, and traceback data can be stored as uint8
+	data []WavefrontValue
+	lo   int
+}
+
+// NewWavefront: returns a new wavefront with size accomodating lo and hi (inclusive)
+func NewWavefront(lo int, hi int) *Wavefront {
+	a := &Wavefront{}
+
+	a.lo = lo
+	size := a.TranslateIndex(hi)
+
+	newData := make([]WavefrontValue, size+1)
+	a.data = newData
+
+	return a
+}
+
+// TranslateIndex: utility function for getting the data index given a diagonal
+func (a *Wavefront) TranslateIndex(diagonal int) int {
+	return diagonal - a.lo
+}
+
+// Get: returns WavefrontValue for given diagonal
+func (a *Wavefront) Get(diagonal int) WavefrontValue {
+	actualIdx := a.TranslateIndex(diagonal)
+	if 0 <= actualIdx && actualIdx < len(a.data) { // idx is in the slice
+		return a.data[actualIdx]
+	} else { // idx is out of the slice
+		return 0
+	}
+}
+
+// Set: the diagonal to a WavefrontValue
+func (a *Wavefront) Set(diagonal int, value WavefrontValue) {
+	actualIdx := a.TranslateIndex(diagonal)
+
+	/* in theory idx is always in bounds because the wavefront is preallocated
+	if actualIdx < 0 || actualIdx >= len(a.data) {
+		return
+	}
+	*/
+
+	a.data[actualIdx] = value
+}
+
+// WavefrontComponent: each M/I/D wavefront matrix including the wavefront data, lo and hi
+type WavefrontComponent struct {
+	lo *PositiveSlice[int]        // lo for each wavefront
+	hi *PositiveSlice[int]        // hi for each wavefront
+	W  *PositiveSlice[*Wavefront] // wavefront diag distance and traceback for each wavefront
+}
+
+// NewWavefrontComponent: returns initialized WavefrontComponent
 func NewWavefrontComponent(preallocateSize int) WavefrontComponent {
 	// new wavefront component = {
 	// lo = [0]
 	// hi = [0]
 	// W = []
-	// A = []
 	// }
 	w := WavefrontComponent{
 		lo: &PositiveSlice[int]{
@@ -53,16 +116,9 @@ func NewWavefrontComponent(preallocateSize int) WavefrontComponent {
 			data:  []int{0},
 			valid: []bool{true},
 		},
-		W: &PositiveSlice[*IntegerSlice[int]]{
-			defaultValue: &IntegerSlice[int]{
-				data:  []int{},
-				valid: []bool{},
-			},
-		},
-		A: &PositiveSlice[*IntegerSlice[traceback]]{
-			defaultValue: &IntegerSlice[traceback]{
-				data:  []traceback{},
-				valid: []bool{},
+		W: &PositiveSlice[*Wavefront]{
+			defaultValue: &Wavefront{
+				data: []WavefrontValue{0},
 			},
 		},
 	}
@@ -70,32 +126,21 @@ func NewWavefrontComponent(preallocateSize int) WavefrontComponent {
 	w.lo.Preallocate(preallocateSize)
 	w.hi.Preallocate(preallocateSize)
 	w.W.Preallocate(preallocateSize)
-	w.A.Preallocate(preallocateSize)
 
 	return w
 }
 
-// get value for wavefront=score, diag=k => returns ok, value
-func (w *WavefrontComponent) GetVal(score int, k int) (bool, int) {
-	return w.W.Valid(score) && w.W.Get(score).Valid(k), w.W.Get(score).Get(k)
+// GetVal: get value for wavefront=score, diag=k => returns ok, value, traceback
+func (w *WavefrontComponent) GetVal(score int, k int) (bool, uint32, Traceback) {
+	return UnpackWavefrontValue(w.W.Get(score).Get(k))
 }
 
-// set value for wavefront=score, diag=k
-func (w *WavefrontComponent) SetVal(score int, k int, val int) {
-	w.W.Get(score).Set(k, val)
+// SetVal: set value, traceback for wavefront=score, diag=k
+func (w *WavefrontComponent) SetVal(score int, k int, val uint32, tb Traceback) {
+	w.W.Get(score).Set(k, PackWavefrontValue(val, tb))
 }
 
-// get alignment traceback for wavefront=score, diag=k => returns ok, value
-func (w *WavefrontComponent) GetTraceback(score int, k int) (bool, traceback) {
-	return w.A.Valid(score) && w.A.Get(score).Valid(k), w.A.Get(score).Get(k)
-}
-
-// set alignment traceback for wavefront=score, diag=k
-func (w *WavefrontComponent) SetTraceback(score int, k int, val traceback) {
-	w.A.Get(score).Set(k, val)
-}
-
-// get hi for wavefront=score
+// GetLoHi: get lo and hi for wavefront=score
 func (w *WavefrontComponent) GetLoHi(score int) (bool, int, int) {
 	// if lo[score] and hi[score] are valid
 	if w.lo.Valid(score) && w.hi.Valid(score) {
@@ -106,89 +151,14 @@ func (w *WavefrontComponent) GetLoHi(score int) (bool, int, int) {
 	}
 }
 
-// set hi for wavefront=score
+// SetLoHi: set lo and hi for wavefront=score
 func (w *WavefrontComponent) SetLoHi(score int, lo int, hi int) {
 	// lo[score] = lo
 	w.lo.Set(score, lo)
 	// hi[score] = hi
 	w.hi.Set(score, hi)
 
-	// preemptively setup w.A
-	w.A.Set(score, &IntegerSlice[traceback]{})
-	w.A.Get(score).Preallocate(lo, hi)
-
 	// preemptively setup w.W
-	w.W.Set(score, &IntegerSlice[int]{})
-	w.W.Get(score).Preallocate(lo, hi)
-}
-
-func (w *WavefrontComponent) String(score int) string {
-	traceback_str := []string{"OI", "EI", "OD", "ED", "SB", "IN", "DL", "EN"}
-	s := "<"
-	min_lo := math.MaxInt
-	max_hi := math.MinInt
-
-	for i := 0; i <= score; i++ {
-		if w.lo.Valid(i) && w.lo.Get(i) < min_lo {
-			min_lo = w.lo.Get(i)
-		}
-		if w.hi.Valid(i) && w.hi.Get(i) > max_hi {
-			max_hi = w.hi.Get(i)
-		}
-	}
-
-	for k := min_lo; k <= max_hi; k++ {
-		s = s + fmt.Sprintf("%02d", k)
-		if k < max_hi {
-			s = s + "|"
-		}
-	}
-
-	s = s + ">\t<"
-
-	for k := min_lo; k <= max_hi; k++ {
-		s = s + fmt.Sprintf("%02d", k)
-		if k < max_hi {
-			s = s + "|"
-		}
-	}
-
-	s = s + ">\n"
-
-	for i := 0; i <= score; i++ {
-		s = s + "["
-		lo := w.lo.Get(i)
-		hi := w.hi.Get(i)
-		// print out wavefront matrix
-		for k := min_lo; k <= max_hi; k++ {
-			if w.W.Valid(i) && w.W.Get(i).Valid(k) {
-				s = s + fmt.Sprintf("%02d", w.W.Get(i).Get(k))
-			} else if k < lo || k > hi {
-				s = s + "--"
-			} else {
-				s = s + "  "
-			}
-
-			if k < max_hi {
-				s = s + "|"
-			}
-		}
-		s = s + "]\t["
-		// print out traceback matrix
-		for k := min_lo; k <= max_hi; k++ {
-			if w.A.Valid(i) && w.A.Get(i).Valid(k) {
-				s = s + traceback_str[w.A.Get(i).Get(k)]
-			} else if k < lo || k > hi {
-				s = s + "--"
-			} else {
-				s = s + "  "
-			}
-
-			if k < max_hi {
-				s = s + "|"
-			}
-		}
-		s = s + "]\n"
-	}
-	return s
+	b := NewWavefront(lo, hi)
+	w.W.Set(score, b)
 }

pkg/utils.go

@@ -2,26 +2,72 @@ package wfa
 
 import (
 	"math"
-	"unicode/utf8"
+	"strings"
+
+	"golang.org/x/exp/constraints"
 )
 
-func SafeMin(values []int, idx int) int {
+func UIntToString(num uint) string { // num assumed to be positive
+	var builder strings.Builder
+
+	for num > 0 {
+		digit := num % 10
+		builder.WriteRune(rune('0' + digit))
+		num /= 10
+	}
+
+	// Reverse the string as we built it in reverse order
+	str := []rune(builder.String())
+	for i, j := 0, len(str)-1; i < j; i, j = i+1, j-1 {
+		str[i], str[j] = str[j], str[i]
+	}
+
+	return string(str)
+}
+
+func RunLengthDecode(encoded string) string {
+	decoded := strings.Builder{}
+	length := len(encoded)
+	i := 0
+
+	for i < length {
+		// If the current character is a digit, we need to extract the run length
+		runLength := 0
+		for i < length && encoded[i] >= '0' && encoded[i] <= '9' {
+			runLength = runLength*10 + int(encoded[i]-'0')
+			i++
+		}
+
+		// The next character will be the character to repeat
+		if i < length {
+			char := encoded[i]
+			for j := 0; j < runLength; j++ {
+				decoded.WriteByte(char)
+			}
+			i++ // Move past the character
+		}
+	}
+
+	return decoded.String()
+}
+
+func SafeMin[T constraints.Integer](values []T, idx int) T {
 	return values[idx]
 }
 
-func SafeMax(values []int, idx int) int {
+func SafeMax[T constraints.Integer](values []T, idx int) T {
 	return values[idx]
 }
 
-func SafeArgMax(valids []bool, values []int) (bool, int) {
+func SafeArgMax[T constraints.Integer](valids []bool, values []T) (bool, int) {
 	hasValid := false
 	maxIndex := 0
 	maxValue := math.MinInt
 	for i := 0; i < len(valids); i++ {
-		if valids[i] && values[i] > maxValue {
+		if valids[i] && int(values[i]) > maxValue {
 			hasValid = true
 			maxIndex = i
-			maxValue = values[i]
+			maxValue = int(values[i])
 		}
 	}
 	if hasValid {
@@ -31,15 +77,15 @@ func SafeArgMax(valids []bool, values []int) (bool, int) {
 	}
 }
 
-func SafeArgMin(valids []bool, values []int) (bool, int) {
+func SafeArgMin[T constraints.Integer](valids []bool, values []T) (bool, int) {
 	hasValid := false
 	minIndex := 0
 	minValue := math.MaxInt
 	for i := 0; i < len(valids); i++ {
-		if valids[i] && values[i] < minValue {
+		if valids[i] && int(values[i]) < minValue {
 			hasValid = true
 			minIndex = i
-			minValue = values[i]
+			minValue = int(values[i])
 		}
 	}
 	if hasValid {
@@ -49,21 +95,6 @@ func SafeArgMin(valids []bool, values []int) (bool, int) {
 	}
 }
 
-func Reverse(s string) string {
-	size := len(s)
-	buf := make([]byte, size)
-	for start := 0; start < size; {
-		r, n := utf8.DecodeRuneInString(s[start:])
-		start += n
-		utf8.EncodeRune(buf[size-start:], r)
-	}
-	return string(buf)
-}
-
-func Splice(s string, c rune, idx int) string {
-	return s[:idx] + string(c) + s[idx:]
-}
-
 func NextLoHi(M WavefrontComponent, I WavefrontComponent, D WavefrontComponent, score int, penalties Penalty) (int, int) {
 	x := penalties.X
 	o := penalties.O
@@ -98,14 +129,13 @@ func NextI(M WavefrontComponent, I WavefrontComponent, score int, k int, penalti
 	o := penalties.O
 	e := penalties.E
 
-	a_ok, a := M.GetVal(score-o-e, k-1)
-	b_ok, b := I.GetVal(score-e, k-1)
+	a_ok, a, _ := M.GetVal(score-o-e, k-1)
+	b_ok, b, _ := I.GetVal(score-e, k-1)
 
-	ok, nextITraceback := SafeArgMax([]bool{a_ok, b_ok}, []int{a, b})
-	nextIVal := SafeMax([]int{a, b}, nextITraceback) + 1 // important that the +1 is here
+	ok, nextITraceback := SafeArgMax([]bool{a_ok, b_ok}, []uint32{a, b})
+	nextIVal := SafeMax([]uint32{a, b}, nextITraceback) + 1 // important that the +1 is here
 	if ok {
-		I.SetVal(score, k, nextIVal)
-		I.SetTraceback(score, k, []traceback{OpenIns, ExtdIns}[nextITraceback])
+		I.SetVal(score, k, nextIVal, []Traceback{OpenIns, ExtdIns}[nextITraceback])
 	}
 }
 
@@ -113,33 +143,27 @@ func NextD(M WavefrontComponent, D WavefrontComponent, score int, k int, penalti
 	o := penalties.O
 	e := penalties.E
 
-	a_ok, a := M.GetVal(score-o-e, k+1)
-	b_ok, b := D.GetVal(score-e, k+1)
+	a_ok, a, _ := M.GetVal(score-o-e, k+1)
+	b_ok, b, _ := D.GetVal(score-e, k+1)
 
-	ok, nextDTraceback := SafeArgMax(
-		[]bool{a_ok, b_ok},
-		[]int{a, b},
-	)
-	nextDVal := SafeMax([]int{a, b}, nextDTraceback) // nothing special
+	ok, nextDTraceback := SafeArgMax([]bool{a_ok, b_ok}, []uint32{a, b})
+	nextDVal := SafeMax([]uint32{a, b}, nextDTraceback)
 	if ok {
-		D.SetVal(score, k, nextDVal)
-		D.SetTraceback(score, k, []traceback{OpenDel, ExtdDel}[nextDTraceback])
+		D.SetVal(score, k, nextDVal, []Traceback{OpenDel, ExtdDel}[nextDTraceback])
 	}
 }
 
 func NextM(M WavefrontComponent, I WavefrontComponent, D WavefrontComponent, score int, k int, penalties Penalty) {
 	x := penalties.X
 
-	a_ok, a := M.GetVal(score-x, k)
+	a_ok, a, _ := M.GetVal(score-x, k)
 	a++ // important to have +1 here
-	b_ok, b := I.GetVal(score, k)
-	c_ok, c := D.GetVal(score, k)
-
-	ok, nextMTraceback := SafeArgMax([]bool{a_ok, b_ok, c_ok}, []int{a, b, c})
-	nextMVal := SafeMax([]int{a, b, c}, nextMTraceback)
+	b_ok, b, _ := I.GetVal(score, k)
+	c_ok, c, _ := D.GetVal(score, k)
 
+	ok, nextMTraceback := SafeArgMax([]bool{a_ok, b_ok, c_ok}, []uint32{a, b, c})
+	nextMVal := SafeMax([]uint32{a, b, c}, nextMTraceback)
 	if ok {
-		M.SetVal(score, k, nextMVal)
-		M.SetTraceback(score, k, []traceback{Sub, Ins, Del}[nextMTraceback])
+		M.SetVal(score, k, nextMVal, []Traceback{Sub, Ins, Del}[nextMTraceback])
 	}
 }

pkg/wfa.go

@@ -1,22 +1,25 @@
 package wfa
 
+import (
+	"strings"
+)
+
 func WFAlign(s1 string, s2 string, penalties Penalty, doCIGAR bool) Result {
 	n := len(s1)
 	m := len(s2)
 	A_k := m - n
-	A_offset := m
+	A_offset := uint32(m)
 	score := 0
 	estimatedScore := (max(n, m) * max(penalties.M, penalties.X, penalties.O, penalties.E)) / 4
 	M := NewWavefrontComponent(estimatedScore)
 	M.SetLoHi(0, 0, 0)
-	M.SetVal(0, 0, 0)
-	M.SetTraceback(0, 0, End)
+	M.SetVal(0, 0, 0, End)
 	I := NewWavefrontComponent(estimatedScore)
 	D := NewWavefrontComponent(estimatedScore)
 
 	for {
 		WFExtend(M, s1, n, s2, m, score)
-		ok, val := M.GetVal(score, A_k)
+		ok, val, _ := M.GetVal(score, A_k)
 		if ok && val >= A_offset {
 			break
 		}
@@ -26,7 +29,7 @@ func WFAlign(s1 string, s2 string, penalties Penalty, doCIGAR bool) Result {
 
 	CIGAR := ""
 	if doCIGAR {
-		CIGAR = WFBacktrace(M, I, D, score, penalties, A_k, s1, s2)
+		CIGAR = WFBacktrace(M, I, D, score, penalties, A_k, A_offset, s1, s2)
 	}
 
 	return Result{
@@ -40,7 +43,8 @@ func WFExtend(M WavefrontComponent, s1 string, n int, s2 string, m int, score in
 	for k := lo; k <= hi; k++ {
 		// v = M[score][k] - k
 		// h = M[score][k]
-		ok, h := M.GetVal(score, k)
+		ok, hu, _ := M.GetVal(score, k)
+		h := int(hu)
 		v := h - k
 
 		// exit early if v or h are invalid
@@ -48,8 +52,8 @@ func WFExtend(M WavefrontComponent, s1 string, n int, s2 string, m int, score in
 			continue
 		}
 		for v < n && h < m && s1[v] == s2[h] {
-			_, val := M.GetVal(score, k)
-			M.SetVal(score, k, val+1)
+			_, val, tb := M.GetVal(score, k)
+			M.SetVal(score, k, val+1, tb)
 			v++
 			h++
 		}
@@ -67,76 +71,128 @@ func WFNext(M WavefrontComponent, I WavefrontComponent, D WavefrontComponent, sc
 	}
 }
 
-func WFBacktrace(M WavefrontComponent, I WavefrontComponent, D WavefrontComponent, score int, penalties Penalty, A_k int, s1 string, s2 string) string {
-	traceback_CIGAR := []string{"I", "I", "D", "D", "X", "", "", ""}
+func WFBacktrace(M WavefrontComponent, I WavefrontComponent, D WavefrontComponent, score int, penalties Penalty, A_k int, A_offset uint32, s1 string, s2 string) string {
 	x := penalties.X
 	o := penalties.O
 	e := penalties.E
-	CIGAR_rev := ""
+
 	tb_s := score
 	tb_k := A_k
-	_, current_traceback := M.GetTraceback(tb_s, tb_k)
 	done := false
 
+	_, current_dist, current_traceback := M.GetVal(tb_s, tb_k)
+
+	Ops := []rune{'~'}
+	Counts := []uint{0}
+	idx := 0
+
 	for !done {
-		CIGAR_rev = CIGAR_rev + traceback_CIGAR[current_traceback]
 		switch current_traceback {
 		case OpenIns:
+			if Ops[idx] == 'I' {
+				Counts[idx]++
+			} else {
+				Ops = append(Ops, 'I')
+				Counts = append(Counts, 1)
+				idx++
+			}
+
 			tb_s = tb_s - o - e
 			tb_k = tb_k - 1
-			_, current_traceback = M.GetTraceback(tb_s, tb_k)
+			_, current_dist, current_traceback = M.GetVal(tb_s, tb_k)
 		case ExtdIns:
+			if Ops[idx] == 'I' {
+				Counts[idx]++
+			} else {
+				Ops = append(Ops, 'I')
+				Counts = append(Counts, 1)
+				idx++
+			}
+
 			tb_s = tb_s - e
 			tb_k = tb_k - 1
-			_, current_traceback = I.GetTraceback(tb_s, tb_k)
+			_, current_dist, current_traceback = I.GetVal(tb_s, tb_k)
 		case OpenDel:
+			if Ops[idx] == 'D' {
+				Counts[idx]++
+			} else {
+				Ops = append(Ops, 'D')
+				Counts = append(Counts, 1)
+				idx++
+			}
+
 			tb_s = tb_s - o - e
 			tb_k = tb_k + 1
-			_, current_traceback = M.GetTraceback(tb_s, tb_k)
+			_, current_dist, current_traceback = M.GetVal(tb_s, tb_k)
 		case ExtdDel:
+			if Ops[idx] == 'D' {
+				Counts[idx]++
+			} else {
+				Ops = append(Ops, 'D')
+				Counts = append(Counts, 1)
+				idx++
+			}
+
 			tb_s = tb_s - e
 			tb_k = tb_k + 1
-			_, current_traceback = D.GetTraceback(tb_s, tb_k)
+			_, current_dist, current_traceback = D.GetVal(tb_s, tb_k)
 		case Sub:
 			tb_s = tb_s - x
 			// tb_k = tb_k;
-			_, current_traceback = M.GetTraceback(tb_s, tb_k)
+			_, next_dist, next_traceback := M.GetVal(tb_s, tb_k)
+
+			if int(current_dist-next_dist)-1 > 0 {
+				Ops = append(Ops, 'M')
+				Counts = append(Counts, uint(current_dist-next_dist)-1)
+				idx++
+			}
+
+			if Ops[idx] == 'X' {
+				Counts[idx]++
+			} else {
+				Ops = append(Ops, 'X')
+				Counts = append(Counts, 1)
+				idx++
+			}
+
+			current_dist = next_dist
+			current_traceback = next_traceback
 		case Ins:
 			// tb_s = tb_s;
 			// tb_k = tb_k;
-			_, current_traceback = I.GetTraceback(tb_s, tb_k)
+			_, next_dist, next_traceback := I.GetVal(tb_s, tb_k)
+
+			Ops = append(Ops, 'M')
+			Counts = append(Counts, uint(current_dist-next_dist))
+			idx++
+
+			current_dist = next_dist
+			current_traceback = next_traceback
 		case Del:
 			// tb_s = tb_s;
 			// tb_k = tb_k;
-			_, current_traceback = D.GetTraceback(tb_s, tb_k)
+			_, next_dist, next_traceback := D.GetVal(tb_s, tb_k)
+
+			Ops = append(Ops, 'M')
+			Counts = append(Counts, uint(current_dist-next_dist))
+			idx++
+
+			current_dist = next_dist
+			current_traceback = next_traceback
 		case End:
+			Ops = append(Ops, 'M')
+			Counts = append(Counts, uint(current_dist))
+			idx++
+
 			done = true
 		}
 	}
 
-	CIGAR_part := Reverse(CIGAR_rev)
-	c := 0
-	i := 0
-	j := 0
-	for i < len(s1) && j < len(s2) {
-		if s1[i] == s2[j] {
-			//CIGAR_part.splice(c, 0, "M")
-			CIGAR_part = Splice(CIGAR_part, 'M', c)
-			c++
-			i++
-			j++
-		} else if CIGAR_part[c] == 'X' {
-			c++
-			i++
-			j++
-		} else if CIGAR_part[c] == 'I' {
-			c++
-			j++
-		} else if CIGAR_part[c] == 'D' {
-			c++
-			i++
-		}
+	CIGAR := strings.Builder{}
+	for i := len(Ops) - 1; i > 0; i-- {
+		CIGAR.WriteString(UIntToString(Counts[i]))
+		CIGAR.WriteRune(Ops[i])
 	}
 
-	return CIGAR_part
+	return CIGAR.String()
 }

test/wfa_test.go

@@ -3,6 +3,8 @@ package tests
 import (
 	"bufio"
 	"encoding/json"
+	"log"
+	"math/rand/v2"
 	"os"
 	"strconv"
 	"strings"
@@ -27,6 +29,89 @@ type TestCase struct {
 	Solutions string      `json:"solutions"`
 }
 
+func randRange(min, max int) uint32 {
+	return uint32(rand.IntN(max-min) + min)
+}
+
+func TestWavefrontPacking(t *testing.T) {
+	for range 1000 {
+		val := randRange(0, 1000)
+		tb := wfa.Traceback(randRange(0, 7))
+		v := wfa.PackWavefrontValue(val, tb)
+
+		valid, gotVal, gotTB := wfa.UnpackWavefrontValue(v)
+
+		if !valid || gotVal != val || gotTB != tb {
+			t.Errorf(`test WavefrontPack/Unpack, val: %d, tb: %d, packedval: %x, gotok: %t, gotval: %d, gottb: %d\n`, val, tb, v, valid, gotVal, gotTB)
+		}
+	}
+}
+
+func GetScoreFromCIGAR(CIGAR string, penalties wfa.Penalty) int {
+	unpackedCIGAR := wfa.RunLengthDecode(CIGAR)
+	previousOp := '~'
+	score := 0
+	for _, Op := range unpackedCIGAR {
+		if Op == 'M' {
+			score = score + penalties.M
+		} else if Op == 'X' {
+			score = score + penalties.X
+		} else if (Op == 'I' && previousOp != 'I') || (Op == 'D' && previousOp != 'D') {
+			score = score + penalties.O + penalties.E
+		} else if (Op == 'I' && previousOp == 'I') || (Op == 'D' && previousOp == 'D') {
+			score = score + penalties.E
+		}
+		previousOp = Op
+	}
+	return score
+}
+
+func CheckCIGARCorrectness(s1 string, s2 string, CIGAR string) bool {
+	unpackedCIGAR := wfa.RunLengthDecode(CIGAR)
+	i := 0
+	j := 0
+
+	s1Aligned := strings.Builder{}
+	alignment := strings.Builder{}
+	s2Aligned := strings.Builder{}
+
+	for c := 0; c < len(unpackedCIGAR); c++ {
+		Op := unpackedCIGAR[c]
+		if Op == 'M' {
+			s1Aligned.WriteByte(s1[i])
+			alignment.WriteRune('|')
+			s2Aligned.WriteByte(s2[j])
+			i++
+			j++
+		} else if Op == 'X' {
+			s1Aligned.WriteByte(s1[i])
+			alignment.WriteRune(' ')
+			s2Aligned.WriteByte(s2[j])
+			i++
+			j++
+		} else if Op == 'I' {
+
+			s1Aligned.WriteRune('-')
+			alignment.WriteRune(' ')
+			s2Aligned.WriteByte(s2[j])
+
+			j++
+		} else if Op == 'D' {
+			s1Aligned.WriteByte(s1[i])
+			alignment.WriteRune('|')
+			s2Aligned.WriteRune('-')
+			i++
+		}
+	}
+
+	if i == len(s1) && j == len(s2) {
+		return true
+	} else {
+		log.Printf("\n%s\n%s\n%s\n i=%d, j=%d, |s1|=%d, |s2|=%d\n", s1Aligned.String(), alignment.String(), s2Aligned.String(), i, j, len(s1), len(s2))
+		return false
+	}
+}
+
 func TestWFA(t *testing.T) {
 	content, _ := os.ReadFile(testJsonPath)
 
@@ -54,7 +139,9 @@ func TestWFA(t *testing.T) {
 
 		for solutions.Scan() {
 			solution := solutions.Text()
+
 			expectedScore, _ := strconv.Atoi(strings.Split(solution, "\t")[0])
+			expectedCIGAR := strings.Split(solution, "\t")[1]
 
 			sequences.Scan()
 			s1 := sequences.Text()
@@ -64,11 +151,27 @@ func TestWFA(t *testing.T) {
 			s2 := sequences.Text()
 			s2 = s2[1:]
 
-			x := wfa.WFAlign(s1, s2, testPenalties, false)
+			x := wfa.WFAlign(s1, s2, testPenalties, true)
 			gotScore := x.Score
+			gotCIGAR := x.CIGAR
 
 			if gotScore != -1*expectedScore {
-				t.Errorf(`test: %s#%d, s1: %s, s2: %s, got: %d, expected: %d\n`, testName, idx, s1, s2, gotScore, expectedScore)
+				t.Errorf(`test: %s#%d, s1: %s, s2: %s, got: %d, expected: %d`, testName, idx, s1, s2, gotScore, expectedScore)
+				os.Exit(1)
+			}
+
+			if gotCIGAR != expectedCIGAR {
+				checkScore := GetScoreFromCIGAR(gotCIGAR, testPenalties)
+				CIGARCorrectness := CheckCIGARCorrectness(s1, s2, gotCIGAR)
+				if checkScore != gotScore && checkScore != -1*expectedScore { // nonequivalent alignment
+					t.Errorf(`test: %s#%d, s1: %s, s2: %s, got: [%s], expected: [%s]`, testName, idx, s1, s2, gotCIGAR, expectedCIGAR)
+					t.Errorf(`test: %s#%d, recalculated score: %d`, testName, idx, checkScore)
+					os.Exit(1)
+				}
+				if !CIGARCorrectness {
+					t.Errorf(`test: %s#%d, s1: %s, s2: %s, got: [%s], expected: [%s]`, testName, idx, s1, s2, gotCIGAR, expectedCIGAR)
+					os.Exit(1)
+				}
 			}
 
 			idx++

wfa.js

@@ -300,10 +300,14 @@
 
 					// func finalizeRef(v ref)
 					"syscall/js.finalizeRef": (v_ref) => {
-						// Note: TinyGo does not support finalizers so this should never be
-						// called.
-						//console.error('syscall/js.finalizeRef not implemented');
-						// for whatever reason this is called by wfajs but doesnt impact the results at all??
+						const id = mem().getUint32(unboxValue(v_ref), true);
+						this._goRefCounts[id]--;
+						if (this._goRefCounts[id] === 0) {
+							const v = this._values[id];
+							this._values[id] = null;
+							this._ids.delete(v);
+							this._idPool.push(id);
+						}
 					},
 
 					// func stringVal(value string) ref