feat(F9-2h): render 旁白套該課 glossary 讀音表（job↔課 → TTS）#75
Merged
Conversation
把 F9-2g 落地的 JobRecord.project_id 接到 TTS：render 前以
ProjectStore.get_glossary(project_id).to_pronunciation_map() 現讀該課讀音表，
render 期間掛上 → 旁白照課程術語讀音念。
- tts_backend.py：新增 course_pronunciation_override context manager（比照
既有 render-scoped override 慣例 video_dimensions_override/talking_head_override，
module-level 覆寫、sequential job 設計下非 thread-safe 可接受），normalize_text
在呼叫端未顯式給 extra_pronunciation 時自動沿用該覆寫（顯式 arg 含 {} 永遠優先）。
深埋的 pipeline.main→gen_tts→synthesize→normalize_text 一條鏈因此不必逐層穿參。
- server/runner.py：新增 _resolve_course_pronunciation(rec)（fail-soft：無 project_id
/ 課不存在 / 無 glossary / 讀音表空 → None＝沿用全域，glossary 解析絕不讓 render
失敗）並把 _run_render 的 inner render 包進第三層 override。
- 守紀律：完全不碰 R-2 render 入口 assert / 狀態機 / reviewed（只影響旁白怎麼念）。
補 tests/test_glossary_tts_render.py 15 測（normalize_text 套用/顯式優先/{} 停用/
還原·例外·巢狀·None no-op；_resolve_course_pronunciation fail-soft 各路；_run_render
wiring inner 期間掛上·出去還原，全 offline 不打 API、不真跑 TTS）。
全套 2665 passed（3 個 QR/journal 字型像素為容器缺 Noto CJK 假象，CI 權威）。
https://claude.ai/code/session_01Ae7ZGWzWw4kfVLcsyAXQv9
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
做什麼
F9-2「課程術語/讀音表 glossary」的下一塊 offline slice（接 F9-2g #74）。把已落地的
JobRecord.project_id接到 TTS：render 旁白前現讀該課 glossary 讀音表，render 期間掛上 → 旁白照課程術語讀音念。對應docs/JOB_COURSE_ASSOCIATION_RFC.md§4.2。為什麼這樣做（架構選擇）
render 旁白 → TTS 深埋在
pipeline.main → gen_tts → synthesize → normalize_text一條鏈。把extra_pronunciation逐層穿過去要動synthesize抽象介面 + 三個 backend（Edge/F5/Google）+ FallbackTTS + gen_tts + render_video 簽章（>5 檔、碰抽象介面）。改採 codebase 既有的 render-scoped override 慣例（
core.config.video_dimensions_override/talking_head_override）：tts_backend.py：新增course_pronunciation_overridecontext manager（module-level 覆寫；sequential job 設計下非 thread-safe 可接受，與既有 override 同取捨）。normalize_text在呼叫端未顯式給extra_pronunciation時自動沿用該覆寫；顯式 arg（含{}）永遠優先。server/runner.py：新增_resolve_course_pronunciation(rec)從rec.project_id取ProjectStore.get_glossary(...).to_pronunciation_map()，並把_run_render的 inner render 包進第三層 override。不可妥協紀律
reviewed（只影響「旁白怎麼念」，硬規則 Add pronunciation mapping for mathematical and Greek symbols in TTS #1）。project_id（無主 job）/ 課不存在 / 無 glossary / 讀音表空 → 一律 None ＝沿用全域pronunciation.json行為，glossary 解析絕不讓 render 失敗。怎麼測
新增
tests/test_glossary_tts_render.py15 測，全 offline（不打 API、不真跑 TTS/ffmpeg、tmp 隔離 ProjectStore）：normalize_text× override：套用 / 顯式優先 /{}停用 / 還原·例外還原·巢狀還原 / None·空 no-op_resolve_course_pronunciation：無 pid / 有 glossary 回 map / 課不存在 fail-soft / 無 glossary / 空 map 收斂 None_run_renderwiring：override 在 inner render 期間掛上、出去還原、無 glossary 沿用全域本機相關子集 105 passed、全套 2665 passed（3 個 QR/journal 字型像素失敗為容器缺 Noto CJK 假象，CI 裝了字型為權威）。
Reviewer 決策點
無新架構抉擇——Option A 已於 #73 RFC 拍板，本 PR 是 RFC §4.2 列好的 offline 接線。唯一風格選擇是「render-scoped override vs 逐層穿參」，已按既有慣例擇前者（理由見上）。
https://claude.ai/code/session_01Ae7ZGWzWw4kfVLcsyAXQv9
Generated by Claude Code