0f81d5e70b
將 Tool_OCR 從 macOS conda 環境轉換為 Docker 單容器部署方案。 前後端整合於同一容器，通過 Nginx 反向代理，僅對外暴露單一端口。 ## 新增功能 - Docker 單容器架構（Frontend + Backend + Nginx） - 多階段構建優化鏡像大小 - Supervisor 進程管理 - 健康檢查機制 - 完整部署文檔 ## 技術細節 - 對外端口：12015（原 12010 已被佔用） - 內部架構：Nginx(12015) → FastAPI(8000) - 前端靜態文件由 Nginx 直接服務 - API 請求通過 Nginx 反向代理 ## 系統依賴完善 - libmagic1：文件類型檢測 - LibreOffice：Office 文檔轉換 - paddlex[ocr]：PP-StructureV3 版面分析 - 中日韓字體支援 ## 配置調整 - 環境變數路徑：macOS 路徑 → 容器絕對路徑 - 前端 API URL：修正為統一端口 12015 - Pip 安裝：延長超時至 600 秒，重試 5 次 - CRLF 轉換：自動處理 Windows 換行符 ## 清理 - 移除臨時文檔（API_FIX_SUMMARY.md 等 7 個文檔） 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
132 lines
3.8 KiB
Docker
132 lines
3.8 KiB
Docker
# ============================================
|
|
# Tool_OCR - Unified Docker Image
|
|
# Frontend (React + Vite) + Backend (FastAPI)
|
|
# Served by Nginx with reverse proxy
|
|
# ============================================
|
|
|
|
# ============================================
|
|
# Stage 1: Build Frontend
|
|
# ============================================
|
|
FROM node:20-alpine AS frontend-builder
|
|
|
|
WORKDIR /app/frontend
|
|
|
|
# Copy package files
|
|
COPY frontend/package*.json ./
|
|
|
|
# Install all dependencies (including devDependencies for build)
|
|
RUN npm ci
|
|
|
|
# Copy frontend source
|
|
COPY frontend/ ./
|
|
|
|
# Create production environment file
|
|
RUN echo "VITE_API_BASE_URL=" > .env.production
|
|
|
|
# Build frontend for production
|
|
RUN npm run build
|
|
|
|
|
|
# ============================================
|
|
# Stage 2: Build Backend + Final Image
|
|
# ============================================
|
|
FROM python:3.10-slim-bookworm
|
|
|
|
# Set working directory
|
|
WORKDIR /app
|
|
|
|
# Set environment variables
|
|
ENV PYTHONUNBUFFERED=1 \
|
|
PYTHONDONTWRITEBYTECODE=1 \
|
|
PIP_NO_CACHE_DIR=1 \
|
|
PIP_DISABLE_PIP_VERSION_CHECK=1 \
|
|
DEBIAN_FRONTEND=noninteractive
|
|
|
|
# Install system dependencies
|
|
# - nginx: web server and reverse proxy
|
|
# - supervisor: process manager for nginx + uvicorn
|
|
# - curl: for health checks
|
|
# - pandoc: for markdown to PDF conversion
|
|
# - poppler-utils: for pdf2image (PDF processing)
|
|
# - libpango-1.0-0, libpangocairo-1.0-0: for WeasyPrint
|
|
# - libgdk-pixbuf2.0-0: for WeasyPrint image handling
|
|
# - libffi-dev: for cryptography
|
|
# - fonts-noto-cjk: Chinese/Japanese/Korean font support
|
|
# - libgomp1, libgl1-mesa-glx, libglib2.0-0: for OpenCV and PaddleOCR
|
|
# - libmagic1: for python-magic file type detection
|
|
# - libreoffice-writer, libreoffice-impress: for Office document conversion (doc/docx/ppt/pptx)
|
|
RUN apt-get update && apt-get install -y --no-install-recommends \
|
|
nginx \
|
|
supervisor \
|
|
curl \
|
|
pandoc \
|
|
poppler-utils \
|
|
libpango-1.0-0 \
|
|
libpangocairo-1.0-0 \
|
|
libgdk-pixbuf2.0-0 \
|
|
libffi-dev \
|
|
fonts-noto-cjk \
|
|
fonts-noto-cjk-extra \
|
|
libgomp1 \
|
|
libgl1-mesa-glx \
|
|
libglib2.0-0 \
|
|
libmagic1 \
|
|
libreoffice-writer \
|
|
libreoffice-impress \
|
|
&& rm -rf /var/lib/apt/lists/*
|
|
|
|
# Copy Python requirements
|
|
COPY requirements.txt .
|
|
|
|
# Install Python dependencies with extended timeout
|
|
# PaddlePaddle is 189MB and may take time to download
|
|
# Timeout: 600 seconds (10 minutes), Retries: 5
|
|
RUN pip install --timeout 600 --retries 5 -r requirements.txt
|
|
|
|
# Copy backend application
|
|
COPY backend/ ./backend/
|
|
|
|
# Copy frontend build from frontend-builder stage
|
|
COPY --from=frontend-builder /app/frontend/dist /app/frontend/dist
|
|
|
|
# Copy Nginx configuration
|
|
COPY docker/nginx.conf /etc/nginx/nginx.conf
|
|
COPY docker/default.conf /etc/nginx/conf.d/default.conf
|
|
|
|
# Copy supervisor configuration
|
|
COPY docker/supervisord.conf /etc/supervisor/conf.d/supervisord.conf
|
|
|
|
# Copy startup script and fix line endings (Windows CRLF -> Linux LF)
|
|
COPY docker/entrypoint.sh /entrypoint.sh
|
|
RUN sed -i 's/\r$//' /entrypoint.sh && chmod +x /entrypoint.sh
|
|
|
|
# Create necessary directories with proper permissions
|
|
RUN mkdir -p \
|
|
/app/backend/uploads/temp \
|
|
/app/backend/uploads/processed \
|
|
/app/backend/uploads/images \
|
|
/app/backend/storage/markdown \
|
|
/app/backend/storage/json \
|
|
/app/backend/storage/exports \
|
|
/app/backend/models/paddleocr \
|
|
/app/backend/logs \
|
|
/var/log/supervisor \
|
|
/var/log/nginx \
|
|
/var/cache/nginx \
|
|
/var/run \
|
|
&& chmod -R 755 /app \
|
|
&& chown -R www-data:www-data /var/log/nginx /var/cache/nginx
|
|
|
|
# Expose port (only one port needed!)
|
|
EXPOSE 12015
|
|
|
|
# Health check
|
|
HEALTHCHECK --interval=30s --timeout=10s --start-period=40s --retries=3 \
|
|
CMD curl -f http://localhost:12015/health || exit 1
|
|
|
|
# Set working directory to backend for Python app
|
|
WORKDIR /app/backend
|
|
|
|
# Use entrypoint script to start supervisor
|
|
ENTRYPOINT ["/entrypoint.sh"]
|