mirror of
https://github.com/zhouxiaoka/autoclip.git
synced 2026-09-03 06:24:14 +08:00
6e59f40f01
Videos without embedded subtitles (e.g. B站 without AI字幕) need Whisper to generate subtitles, but bundling it would bloat every install. Instead let users install it on demand from Settings → 语音识别, and pick which model. Backend uses faster-whisper (CTranslate2, no PyTorch, ~214MB installed, several times faster than openai-whisper, cross-platform) — chosen over mlx-whisper, which hard-depends on torch (~2-3GB). - whisper_runtime.py (new): pip-install faster-whisper into a user-writable dir (<data>/whisper-runtime) using the bundled Python; add to sys.path; status + coarse progress; uninstall. Never writes into the signed .app bundle. - whisper_model_manager.py: tiny→large-v3 from Systran/faster-whisper-*, background download via huggingface_hub, real status; cache under <data>/whisper-models. - speech_recognizer.py: subtitle generation rewritten from the `whisper` CLI to faster-whisper's WhisperModel API → SRT; availability = runtime installed. - speech_recognition.py API: /whisper/install, /whisper/uninstall, /whisper/runtime-status (+ existing /whisper-models*). - SpeechRecognitionConfig.tsx: was a stub; now a full UI (install button + progress + log, model list with download/delete/status) wired to a new speechApi in services/api.ts. - build_macos_arm.sh: allowlist faster_whisper/ctranslate2/huggingface_hub in the dependency guard (they're installed at runtime, imported lazily). Verified end-to-end: install runtime → download tiny model → transcribe a real video into a valid SRT, all through the API/runtime. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
scripts/
构建与运维脚本。桌面客户端只有一条打包路线（python-build-standalone，下称 PBS）， 历史上的 PyInstaller / prepare_resources 路线及其脚本已全部移除。
脚本清单
| 脚本 | 用途 |
|---|---|
build_macos_arm.sh |
唯一的桌面打包脚本（macOS Apple Silicon）。端到端产出 .app + .dmg。 |
verify_desktop.sh |
后端冒烟测试：cargo check + 起后端，校验 /health 与 /api/v1/video-categories。被 nightly-desktop-smoke.yml 调用。 |
monitor_whisper.py |
运行期 Whisper 任务监控，被根目录 start_autoclip.sh / check_whisper_status.sh 调用。 |
打包桌面客户端（macOS arm64）
./scripts/build_macos_arm.sh
产物：
src-tauri/target/release/bundle/macos/
├── AutoClip Desktop.app
└── AutoClip Desktop_1.0.0_aarch64.dmg
这个脚本做了什么
- 下载便携 Python 运行时（python-build-standalone，缓存在
build/pbs-cache/） - 用便携 Python 安装
requirements.txt的全部依赖 - 依赖完整性检查：AST 扫描后端所有第三方 import，缺任何一个就让构建失败 （防止"开发机能跑、打包就 500"）
- rsync 拷贝后端源码到
src-tauri/resources/backend/ - 下载静态 ffmpeg + ffprobe（arm64，自包含，零 homebrew 依赖，缓存在
build/ffmpeg-cache/） - 构建前端（
npm ci && npm run build） cargo tauri build --bundles app- 把
python/、backend/、ffmpeg/注入.app/Contents/Resources/resources/ - ad-hoc 签名 +
hdiutil手动打 DMG
前置依赖
- Node.js 18+、Rust（带
aarch64-apple-darwintarget）、cargo-tauri (cargo install tauri-cli) - 系统 不需要 预装 Python / ffmpeg —— 脚本会自带便携版
环境变量
PIP_INDEX_URL：pip 源，默认清华镜像；CI 里设为https://pypi.org/simple
CI
.github/workflows/desktop-build.yml 在 workflow_dispatch 和 v* tag 上跑同一个
build_macos_arm.sh，打 tag 时把 DMG 挂到 GitHub Release。
当前仅 macOS arm64。PBS 与静态 ffmpeg 对 Intel/Windows/Linux 都有对应版本， 后续把 runner、下载 URL、tauri target 参数化即可泛化。
开发模式（不打包）
直接用 Tauri 开发模式，热重载：
cd src-tauri && cargo tauri dev
（前端 :3000 + 后端动态端口，由 backend_manager.rs 拉起）