feat: split text support multiple languages #593
This commit is contained in:
@@ -107,6 +107,13 @@ export function getUUID(): string {
|
||||
|
||||
export function splitText(text: string): string[] {
|
||||
const data: string[] = []
|
||||
if (Intl.Segmenter) {
|
||||
const segmenter = new Intl.Segmenter()
|
||||
const segments = segmenter.segment(text)
|
||||
for (const { segment } of segments) {
|
||||
data.push(segment)
|
||||
}
|
||||
} else {
|
||||
const symbolMap = new Map<number, string>()
|
||||
for (const match of text.matchAll(UNICODE_SYMBOL_REG)) {
|
||||
symbolMap.set(match.index!, match[0])
|
||||
@@ -122,6 +129,7 @@ export function splitText(text: string): string[] {
|
||||
t++
|
||||
}
|
||||
}
|
||||
}
|
||||
return data
|
||||
}
|
||||
|
||||
|
||||
Reference in New Issue
Block a user