77 lines
1.6 KiB
Go
77 lines
1.6 KiB
Go
|
package outreach
|
|||
|
|
|
|||
|
|
import (
|
|||
|
|
"strings"
|
|||
|
|
)
|
|||
|
|
|
|||
|
|
func NormalizeDraftText(text string) string {
|
|||
|
|
text = strings.ReplaceAll(text, "\r\n", "\n")
|
|||
|
|
text = strings.ReplaceAll(text, "\r", "\n")
|
|||
|
|
text = strings.TrimSpace(text)
|
|||
|
|
text = humanizeDraftLineBreaks(text)
|
|||
|
|
return trimDraftText(text)
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
func humanizeDraftLineBreaks(text string) string {
|
|||
|
|
if text == "" || strings.Contains(text, "\n") {
|
|||
|
|
return text
|
|||
|
|
}
|
|||
|
|
if len([]rune(text)) < 40 {
|
|||
|
|
return text
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
sentences := splitSentences(text)
|
|||
|
|
if len(sentences) <= 1 {
|
|||
|
|
return text
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
paragraphs := make([]string, 0, (len(sentences)+1)/2)
|
|||
|
|
for i := 0; i < len(sentences); {
|
|||
|
|
end := i + 2
|
|||
|
|
if end > len(sentences) {
|
|||
|
|
end = len(sentences)
|
|||
|
|
}
|
|||
|
|
paragraphs = append(paragraphs, strings.Join(sentences[i:end], ""))
|
|||
|
|
i = end
|
|||
|
|
}
|
|||
|
|
return strings.Join(paragraphs, "\n\n")
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
func splitSentences(text string) []string {
|
|||
|
|
runes := []rune(text)
|
|||
|
|
if len(runes) == 0 {
|
|||
|
|
return nil
|
|||
|
|
}
|
|||
|
|
out := make([]string, 0, 4)
|
|||
|
|
var current strings.Builder
|
|||
|
|
for i := 0; i < len(runes); i++ {
|
|||
|
|
ch := runes[i]
|
|||
|
|
current.WriteRune(ch)
|
|||
|
|
if !isSentenceEnd(ch) {
|
|||
|
|
continue
|
|||
|
|
}
|
|||
|
|
// Keep ellipsis / repeated punctuation attached to the same sentence.
|
|||
|
|
for i+1 < len(runes) && isSentenceEnd(runes[i+1]) {
|
|||
|
|
i++
|
|||
|
|
current.WriteRune(runes[i])
|
|||
|
|
}
|
|||
|
|
if sentence := strings.TrimSpace(current.String()); sentence != "" {
|
|||
|
|
out = append(out, sentence)
|
|||
|
|
}
|
|||
|
|
current.Reset()
|
|||
|
|
}
|
|||
|
|
if tail := strings.TrimSpace(current.String()); tail != "" {
|
|||
|
|
out = append(out, tail)
|
|||
|
|
}
|
|||
|
|
return out
|
|||
|
|
}
|
|||
|
|
|
|||
|
|
func isSentenceEnd(ch rune) bool {
|
|||
|
|
switch ch {
|
|||
|
|
case '。', '！', '？', '!', '?', '…':
|
|||
|
|
return true
|
|||
|
|
default:
|
|||
|
|
return false
|
|||
|
|
}
|
|||
|
|
}
|