mirror of
https://github.com/stablyai/orca.git
synced 2026-11-17 03:15:01 +00:00
fix(dictation): write non-ASCII filler punctuation as escapes so the orcad bundle stays ASCII
This commit is contained in:
@@ -18,17 +18,17 @@ FILLERS_BY_LANGUAGE.yue = FILLERS_BY_LANGUAGE.zh
|
||||
// For an unknown language: English fillers that are no word elsewhere ("um" is German and Portuguese, "er" German).
|
||||
const LANGUAGE_NEUTRAL_FILLERS = 'u+h+m*|u+m{2,}|e+r+m+|h+m+|m{3,}'
|
||||
|
||||
const WORD_CHAR = String.raw`[\p{L}\p{N}\p{M}_/'’\-]`
|
||||
const WORD_CHAR = String.raw`[\p{L}\p{N}\p{M}_/'\u2019\-]`
|
||||
// A filler stands alone: nothing joins it to a word (a dot only does inside a name like "um.txt"),
|
||||
// and "hm?" or "eh!" is a question or a call, so it stays.
|
||||
const standalone = (fillers: string): string =>
|
||||
String.raw`(?<!${WORD_CHAR}|${WORD_CHAR}\.)(${fillers})(?!${WORD_CHAR}|\.${WORD_CHAR})(?!\s*[??!!])`
|
||||
String.raw`(?<!${WORD_CHAR}|${WORD_CHAR}\.)(${fillers})(?!${WORD_CHAR}|\.${WORD_CHAR})(?!\s*[?\uFF1F!\uFF01])`
|
||||
// The comma or ellipsis after a filler goes with it.
|
||||
const TRAILING_PAUSE = String.raw`(?:\s*(?:[,,、…]|\.{3}))?\s*`
|
||||
const TRAILING_PAUSE = String.raw`(?:\s*(?:[,\uFF0C\u3001\u2026]|\.{3}))?\s*`
|
||||
const SENTENCE_END_CHARS = '.!?。!?'
|
||||
const SENTENCE_END_RE = new RegExp(`^[${SENTENCE_END_CHARS}]+\\s*`, 'u')
|
||||
// A comma left in front of a removed filler that ended the sentence or the text.
|
||||
const TRAILING_SOFT_PUNCTUATION_RE = /[\s,,、]+$/u
|
||||
const TRAILING_SOFT_PUNCTUATION_RE = /[\s,\uFF0C\u3001]+$/u
|
||||
const NON_LATIN_LETTER_RE = /(?=\p{L})\P{Script=Latin}/u
|
||||
|
||||
function fillerPattern(languages: readonly string[] | undefined, text: string): RegExp | null {
|
||||
|
||||
Reference in New Issue
Block a user