feat: 页面接入 tts 服务

.env.example

@@ -50,3 +50,27 @@ VITE_API_BASE_URL=http://localhost:8088/api
 
 # Optional: Analytics
 # VITE_ANALYTICS_ID=
+
+# -----------------------------------------------------------------------------
+# TTS Providers
+# -----------------------------------------------------------------------------
+# Default provider used when a request omits `provider`. Edge TTS works
+# out of the box (no key required). To enable the Mi (Xiaomi) provider,
+# set both MI_TTS_ENDPOINT and MI_TTS_API_KEY. These are read by the
+# Python tts-service and shared across all users.
+
+# TTS_DEFAULT_PROVIDER=edge
+TTS_SERVICE_URL=http://localhost:5000
+TTS_AUDIO_CACHE_DIR=./data/audio
+# TTS_PORT=5000
+
+# Edge TTS — public, no key needed
+EDGE_TTS_ENABLED=1
+
+# Mi / Xiaomi TTS — fill in to enable
+# Get the endpoint & API key from your Xiaomi Cloud / Mi AI integration.
+# Primary (recommended):
+TTS_MIMO_API_TOKEN=sk-xxxxxxxxxxxxxxxxxxxx
+TTS_MIMO_MODEL=mimo-v2.5-tts
+TTS_MIMO_ENDPOINT=https://api.xiaomimimo.com/v1
+

.github/workflows/android-release.yml

@@ -6,11 +6,12 @@ on:
       - "v*.*.*"
   workflow_dispatch:
 
+permissions:
+  contents: write
+
 jobs:
   build:
     runs-on: ubuntu-latest
-    permissions:
-      contents: write
 
     steps:
       - name: Checkout
@@ -41,11 +42,6 @@ jobs:
           NPM_CONFIG_REGISTRY: https://registry.npmjs.org/
         run: pnpm install --no-frozen-lockfile
 
-      - name: Prebuild Android
-        run: |
-          cd apps/mobile
-          npx expo prebuild --clean --platform android
-
       - name: Build Android APK
         run: |
           cd apps/mobile/android

.github/workflows/feat-test-release.yml

@@ -6,11 +6,12 @@ on:
       - "feat/test"
   workflow_dispatch:
 
+permissions:
+  contents: write
+
 jobs:
   build-desktop:
     name: Build Desktop (${{ matrix.platform }})
-    permissions:
-      contents: write
     strategy:
       matrix:
         include:
@@ -69,8 +70,6 @@ jobs:
   build-android:
     name: Build Android
     runs-on: ubuntu-latest
-    permissions:
-      contents: write
     steps:
       - name: Checkout
         uses: actions/checkout@v4
@@ -100,11 +99,6 @@ jobs:
           NPM_CONFIG_REGISTRY: https://registry.npmjs.org/
         run: pnpm install --no-frozen-lockfile
 
-      - name: Prebuild Android
-        run: |
-          cd apps/mobile
-          npx expo prebuild --platform android
-
       - name: Build Android APK
         run: |
           cd apps/mobile/android
@@ -129,5 +123,4 @@ jobs:
           draft: true
           files: |
             ${{ steps.prepare_apk.outputs.apk_path }}
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          token: ${{ secrets.GITHUB_TOKEN }}

.gitignore

@@ -32,8 +32,6 @@ credentials.json
 # Docker
 .docker/
 
-# Prisma
-apps/server/prisma/migrations/
 
 # Test
 coverage/
@@ -42,4 +40,21 @@ coverage/
 my-release-key.jks
 
 ebooks/
-covers/
+covers/
+
+.claude
+
+.tts-cache
+
+# Python bytecode / cache
+__pycache__/
+*.py[cod]
+*$py.class
+*.egg-info/
+.eggs/
+build/
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+venv/
+.venv/

Dockerfile

@@ -78,54 +78,32 @@ RUN SOURCE_PRISMA=$(find /app/node_modules/.pnpm -path '*/@prisma+client@*/node_
       cp -r "$SOURCE_PRISMA" "$TARGET_PRISMA/"; \
     fi
 
-# ── Stage 4: TTS Model Downloader ────────────────────────────────────────────
-FROM python:${PYTHON_VERSION}-slim AS tts-model-builder
-
-RUN apt-get update && apt-get install -y --no-install-recommends curl && rm -rf /var/lib/apt/lists/*
-
-# Download default voice model (HuggingFace - GitHub release 404)
-ENV PIPER_MODEL_URL="https://huggingface.co/rhasspy/piper-voices/resolve/main/en/en_US/lessac/medium/en_US-lessac-medium.onnx"
-ENV PIPER_MODEL_JSON_URL="https://huggingface.co/rhasspy/piper-voices/resolve/main/en/en_US/lessac/medium/en_US-lessac-medium.onnx.json"
-
-RUN mkdir -p /models && \
-    curl -L --fail -o /models/voice.onnx "$PIPER_MODEL_URL" && \
-    curl -L --fail -o /models/voice.onnx.json "$PIPER_MODEL_JSON_URL"
-
 
 # ── Stage 5: Production Runner ───────────────────────────────────────────────
 FROM node:${NODE_VERSION}-bookworm AS runner
 
 LABEL org.opencontainers.image.title="BookDock"
-LABEL org.opencontainers.image.description="BookDock all-in-one: API + Web + TTS"
+LABEL org.opencontainers.image.description="BookDock all-in-one: API + Web + TTS (Edge/Mi)"
 LABEL maintainer="BookDock Team"
 
 WORKDIR /app
 
-# Install Python 3.11, pip, runtime deps, and TTS Python packages directly
+# Install Python 3.11, pip, runtime deps, and TTS Python packages.
+# Edge TTS only needs stdlib + httpx; no heavy native deps.
 RUN apt-get update && apt-get install -y --no-install-recommends \
     python3.11 \
     python3-pip \
-    python3.11-venv \
-    libespeak1 \
-    libsndfile1 \
-    espeak-ng \
-    sox \
-    libsox-fmt-mp3 \
     supervisor \
     tini \
     && rm -rf /var/lib/apt/lists/* \
     && pip3 install --break-system-packages --no-cache-dir \
-    piper-tts \
-    onnxruntime \
+    edge-tts \
+    httpx \
     fastapi \
     uvicorn \
     pydantic \
-    aiofiles \
     && npm install -g prisma@5.22.0
 
-# Copy voice model
-COPY --from=tts-model-builder /models /models
-
 # Copy deployed server (self-contained with all deps)
 COPY --from=server-builder /app/server-deploy ./
 
@@ -149,9 +127,12 @@ ENV DATABASE_URL=file:/data/db/bookdock.db
 ENV NAS_EBOOK_PATH=/data/ebooks
 ENV NAS_AUDIO_PATH=/data/audio
 ENV SOURCE_LOCAL_PATH=/data/sources
-ENV PIPER_VOICE_PATH=/models/voice.onnx
-ENV PIPER_SAMPLE_RATE=22050
-ENV PIPER_PORT=5000
+ENV TTS_SERVICE_URL=http://localhost:5000
+ENV TTS_DEFAULT_PROVIDER=edge
+ENV TTS_AUDIO_CACHE_DIR=/data/audio
+ENV TTS_PORT=5000
+ENV EDGE_TTS_ENABLED=1
+ENV MI_TTS_ENABLED=1
 
 # Write supervisord config
 RUN cat > /etc/supervisor/conf.d/bookdock.conf << 'EOF'
@@ -162,14 +143,13 @@ logfile=/var/log/supervisor/supervisord.log
 pidfile=/var/run/supervisord.pid
 
 [program:tts]
-command=python3 -m uvicorn tts_service:app --host 0.0.0.0 --port 5000 --app-dir /app/tts-service
+command=/app/tts-service/start.sh
 autostart=true
 autorestart=true
 stdout_logfile=/dev/stdout
 stdout_logfile_maxbytes=0
 stderr_logfile=/dev/stderr
 stderr_logfile_maxbytes=0
-environment=PIPER_VOICE_PATH="/models/voice.onnx",PIPER_SAMPLE_RATE="22050",PIPER_PORT="5000"
 
 [program:api]
 command=node /app/dist/main.js

apps/desktop/src/App.tsx

@@ -413,7 +413,7 @@ function AppLayout({ children }: { children: React.ReactNode }) {
       <main
         className={
           isReaderPage || isLoginPage
-            ? ""
+            ? "h-screen flex flex-col"
             : "max-w-7xl mx-auto px-4 sm:px-6 lg:px-8 py-8 pb-24 md:pb-8"
         }
       >