Update Dockerfile
Browse files- Dockerfile +6 -89
Dockerfile
CHANGED
|
@@ -1,91 +1,8 @@
|
|
| 1 |
-
FROM python:3.10-slim
|
| 2 |
-
|
| 3 |
-
# Set environment variables
|
| 4 |
-
ENV PYTHONDONTWRITEBYTECODE=1
|
| 5 |
-
ENV PYTHONUNBUFFERED=1
|
| 6 |
-
ENV DEBIAN_FRONTEND=noninteractive
|
| 7 |
-
|
| 8 |
-
# Create non-root user for HF Spaces
|
| 9 |
-
RUN useradd -m -u 1000 user
|
| 10 |
-
ENV HOME=/home/user
|
| 11 |
-
ENV PATH=/home/user/.local/bin:$PATH
|
| 12 |
-
|
| 13 |
-
# Install system dependencies for pdf2htmlEX
|
| 14 |
RUN apt-get update && apt-get install -y --no-install-recommends \
|
| 15 |
-
wget \
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
cmake \
|
| 20 |
-
pkg-config \
|
| 21 |
-
libfontforge-dev \
|
| 22 |
-
libcairo2-dev \
|
| 23 |
-
libpango1.0-dev \
|
| 24 |
-
libfreetype6-dev \
|
| 25 |
-
libglib2.0-dev \
|
| 26 |
-
libjpeg-dev \
|
| 27 |
-
libpng-dev \
|
| 28 |
-
libopenjp2-7-dev \
|
| 29 |
-
libtiff-dev \
|
| 30 |
-
libxml2-dev \
|
| 31 |
-
libpoppler-dev \
|
| 32 |
-
libpoppler-cpp-dev \
|
| 33 |
-
libpoppler-glib-dev \
|
| 34 |
-
poppler-utils \
|
| 35 |
fontforge \
|
| 36 |
-
fonts-dejavu \
|
| 37 |
-
|
| 38 |
-
fonts-noto \
|
| 39 |
-
fonts-noto-cjk \
|
| 40 |
-
fonts-thai-tlwg \
|
| 41 |
-
ttf-mscorefonts-installer \
|
| 42 |
-
&& rm -rf /var/lib/apt/lists/* \
|
| 43 |
-
&& apt-get clean
|
| 44 |
-
|
| 45 |
-
# Install pdf2htmlEX from source
|
| 46 |
-
WORKDIR /tmp
|
| 47 |
-
|
| 48 |
-
# Clone and build pdf2htmlEX
|
| 49 |
-
RUN git clone --depth 1 https://github.com/pdf2htmlEX/pdf2htmlEX.git && \
|
| 50 |
-
cd pdf2htmlEX && \
|
| 51 |
-
mkdir build && cd build && \
|
| 52 |
-
cmake .. -DCMAKE_INSTALL_PREFIX=/usr && \
|
| 53 |
-
make -j$(nproc) && \
|
| 54 |
-
make install && \
|
| 55 |
-
cd /tmp && rm -rf pdf2htmlEX
|
| 56 |
-
|
| 57 |
-
# Verify pdf2htmlEX installation
|
| 58 |
-
RUN pdf2htmlEX --version || echo "pdf2htmlEX installed"
|
| 59 |
-
|
| 60 |
-
# Set working directory
|
| 61 |
-
WORKDIR $HOME/app
|
| 62 |
-
|
| 63 |
-
# Copy requirements first for caching
|
| 64 |
-
COPY --chown=user:user requirements.txt .
|
| 65 |
-
|
| 66 |
-
# Switch to non-root user
|
| 67 |
-
USER user
|
| 68 |
-
|
| 69 |
-
# Install Python dependencies
|
| 70 |
-
RUN pip install --no-cache-dir --upgrade pip && \
|
| 71 |
-
pip install --no-cache-dir -r requirements.txt
|
| 72 |
-
|
| 73 |
-
# Copy application code
|
| 74 |
-
COPY --chown=user:user . .
|
| 75 |
-
|
| 76 |
-
# Create necessary directories
|
| 77 |
-
RUN mkdir -p /home/user/app/temp /home/user/app/output
|
| 78 |
-
|
| 79 |
-
# Expose port for Gradio
|
| 80 |
-
EXPOSE 7860
|
| 81 |
-
|
| 82 |
-
# Set environment variables for Gradio
|
| 83 |
-
ENV GRADIO_SERVER_NAME=0.0.0.0
|
| 84 |
-
ENV GRADIO_SERVER_PORT=7860
|
| 85 |
-
|
| 86 |
-
# Health check
|
| 87 |
-
HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
|
| 88 |
-
CMD curl -f http://localhost:7860/ || exit 1
|
| 89 |
-
|
| 90 |
-
# Run the application
|
| 91 |
-
CMD ["python", "app.py"]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
RUN apt-get update && apt-get install -y --no-install-recommends \
|
| 2 |
+
wget curl git build-essential cmake pkg-config \
|
| 3 |
+
libcairo2-dev libpango1.0-dev libfreetype6-dev libglib2.0-dev \
|
| 4 |
+
libjpeg-dev libpng-dev libopenjp2-7-dev libtiff-dev libxml2-dev \
|
| 5 |
+
libpoppler-dev libpoppler-cpp-dev libpoppler-glib-dev poppler-utils \
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
fontforge \
|
| 7 |
+
fonts-dejavu fonts-liberation fonts-noto fonts-noto-cjk fonts-thai-tlwg \
|
| 8 |
+
&& rm -rf /var/lib/apt/lists/* && apt-get clean
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|