fix: CJK 转写分段拼接修复 - #94
Merged
Merged
Conversation
WhisperKit returns one text per VAD chunk; joining with a single space inserted spaces between Chinese/Japanese/Korean segments. Add TranscriptSegmentJoiner that selects the separator from the language (auto falls back to boundary characters) and use it for both file and streaming transcription. Also add a regression test locking hotword prompt tokens when the language is auto. Co-authored-by: multica-agent <github@multica.ai>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
背景
WhisperKit 在 VAD 分段下每个 chunk 返回一条文本，原实现统一用空格拼接，导致中文/日文/韩文段间插入空格，污染原始转写与后续 LLM 输入。
改动
Sources/Speech/TranscriptSegmentJoiner.swift：zh/yue/ja/ko段间无分隔符，其他语言用单空格；语言未知（auto）时按相邻边界字符判断（任一侧为 CJK 则无空格）。逐段 trim、丢弃空段。Sources/Speech/WhisperEngine.swift：文件转写改用该拼接器，并把language透传给流式会话。Sources/Speech/WhisperStreamingSession.swift：新增language属性（init 默认nil），partial 拼接同样走拼接器。Tests/OpenTypeTests/TranscriptSegmentJoinerTests.swift（新增）：zh/ja/ko 无空格、en 有空格、auto 内容回退、trim/空段。Tests/OpenTypeTests/SpeechRecognitionQualityTests.swift：新增 auto 语言下热词 prompt token 非空的回归测试（复核确认 Add Volcengine Doubao ASR speech engine support #1 为假阳性，仅锁定现有行为）。测试
swift build→Build complete! (164.77 secs)swift test --filter 'TranscriptSegmentJoinerTests|SpeechRecognitionQualityTests'→ 13 passed / 0 failuresStreamingSpeechSupportTests|TranscriptionSanitizerTests|RealtimeAudioConversionTests→ 合计 38 passed / 0 failures备注
#3/#4（录音能量闸门与弱音频启发式阈值）本轮不改：缺带实测
averageRMS/maxRMS的真实语料，无法标定，后续单独立项。