func checkPackedAlpha(alpha alphabet.Nucleic) error {
if alpha.Len() != 4 {
return bio.NewError("Cannot create packed sequence with alphabet length != 4", 0, alpha)
}
for _, v := range alphabet.BytesToLetters([]byte(alpha.String())) {
if c, ok := alpha.Complement(v); ok && alpha.IndexOf(v) != alpha.IndexOf(c)^0x3 {
// TODO: Resolution to the following problem:
// Normal nucleotide alphabets (ACGT/ACGU) are safe with this in either case sensitive or
// insensitive. Other alphabets may not be, in this case specify case sensitive.
return bio.NewError("alphabet order not consistent with bit operations for packed.", 0, alpha)
}
}
return nil
}
// Pack bytes that conform to a into a slice of alphabet.Pack. Panics if a byte in s does not conform.
func PackLetters(a alphabet.Nucleic, s ...alphabet.Letter) (p *Packing) {
p = &Packing{
Letters: make([]alphabet.Pack, (len(s)+3)/4),
RightPad: int8(4-len(s)&3) & 3,
}
for i, c := range s {
if !a.IsValid(c) {
panic("packed: invalid letter")
}
p.Letters[i/4] <<= 2
p.Letters[i/4] |= alphabet.Pack(a.IndexOf(c))
}
if sc := uint(len(s)) & 3; sc != 0 {
p.Letters[len(p.Letters)-1] <<= (4 - sc) << 1
}
return
}
// Create a new Seq with the given id, letter sequence and alphabet.
func NewSeq(id string, b []alphabet.Letter, alpha alphabet.Nucleic) (p *Seq, err error) {
defer func() {
if r := recover(); r != nil {
_, pos := alpha.AllValid(b)
err = bio.NewError(fmt.Sprintf("Encoding error: %s %q at position %d.", r, b[pos], pos), 1, b)
}
}()
err = checkPackedAlpha(alpha)
if err != nil {
return
}
p = &Seq{
ID: id,
S: PackLetters(alpha, b...),
alphabet: alpha,
Strand: 1,
Stringify: Stringify,
}
return
}
