// Entropic returns the entropic complexity of a segment of s defined by
// start and end.
func Entropic(s seq.Sequence, start, end int) (ce float64, err error) {
if start < s.Start() || end > s.End() {
err = fmt.Errorf("complex: index out of range")
return
}
if start == end {
return 0, nil
}
var N float64
k := s.Alphabet().Len()
logk := math.Log(float64(k))
n := make([]float64, k)
// tally classes
it := s.Alphabet().LetterIndex()
for i := start; i < end; i++ {
if ind := it[s.At(i).L]; ind >= 0 {
N++
n[ind]++
}
}
// -∑i=1..k((n_i/N)*log_k(n_i/N))
for i := 0; i < k; i++ {
if n[i] != 0 { // ignore zero counts
ce += n[i] * logBaseK(logk, n[i]/N)
}
}
ce = -ce / N
return
}
// WF returns the Wootton and Federhen complexity of a segment of s defined by
// start and end.
func WF(s seq.Sequence, start, end int) (cwf float64, err error) {
if start < s.Start() || end > s.End() {
err = fmt.Errorf("complex: index out of range")
return
}
if start == end {
return 0, nil
}
var N int
k := s.Alphabet().Len()
logk := math.Log(float64(k))
n := make([]int, k)
// tally classes
it := s.Alphabet().LetterIndex()
for i := start; i < end; i++ {
if ind := it[s.At(i).L]; ind >= 0 {
N++
n[ind]++
}
}
// 1/N*log_k(N!/∏i=1..k(n_i!))
cwf = lnFac(N)
for i := 0; i < k; i++ {
cwf -= lnFac(n[i])
}
cwf /= float64(N) * logk
return
}
// Z returns the zlib compression estimate of complexity of a segment of s defined by
// start and end.
func Z(s seq.Sequence, start, end int) (cz float64, err error) {
if start < s.Start() || end > s.End() {
err = fmt.Errorf("complex: index out of range")
return
}
if start == end {
return 0, nil
}
bc := new(byteCounter)
z := zlib.NewWriter(bc)
defer z.Close()
it := s.Alphabet().LetterIndex()
var N float64
for i := start; i < end; i++ {
if b := byte(s.At(i).L); it[b] >= 0 {
N++
z.Write([]byte{b})
}
}
z.Close()
cz = (float64(*bc - overhead)) / N
return
}
