phodal/coca

package evaluator

import (
    "unicode"
    "unicode/utf8"
)

func SplitCamelcase(src string) (entries []string) {
    // don't split invalid utf8
    if !utf8.ValidString(src) {
        return []string{src}
    }
    entries = []string{}
    var runes [][]rune
    lastClass := 0
    class := 0
    // split into fields based on class of unicode character
    for _, r := range src {
        switch true {
        case unicode.IsLower(r):
            class = 1
        case unicode.IsUpper(r):
            class = 2
        case unicode.IsDigit(r):
            class = 3
        default:
            class = 4
        }
        if class == lastClass {
            runes[len(runes)-1] = append(runes[len(runes)-1], r)
        } else {
            runes = append(runes, []rune{r})
        }
        lastClass = class
    }
    // handle upper case -> lower case sequences, e.g.
    // "PDFL", "oader" -> "PDF", "Loader"
    for i := 0; i < len(runes)-1; i++ {
        if unicode.IsUpper(runes[i][0]) && unicode.IsLower(runes[i+1][0]) {
            runes[i+1] = append([]rune{runes[i][len(runes[i])-1]}, runes[i+1]...)
            runes[i] = runes[i][:len(runes[i])-1]
        }
    }
    // construct []string from results
    for _, s := range runes {
        if len(s) > 0 {
            entries = append(entries, string(s))
        }
    }
    return
}