ht-docker-ai/Dockerfile_paddleocr_cpu

# PaddleOCR CPU Variant
# OCR processing optimized for CPU-only inference
FROM python:3.10-slim-bookworm

LABEL maintainer="Task Venture Capital GmbH <hello@task.vc>"
LABEL description="PaddleOCR PP-OCRv4 - CPU optimized"
LABEL org.opencontainers.image.source="https://code.foss.global/host.today/ht-docker-ai"

# Environment configuration for CPU-only mode
ENV OCR_LANGUAGE="en"
ENV SERVER_PORT="5000"
ENV SERVER_HOST="0.0.0.0"
ENV PYTHONUNBUFFERED=1
# Disable GPU usage for CPU-only variant
ENV CUDA_VISIBLE_DEVICES="-1"

# Set working directory
WORKDIR /app

# Install system dependencies
RUN apt-get update && apt-get install -y --no-install-recommends \
    libgl1-mesa-glx \
    libglib2.0-0 \
    libgomp1 \
    curl \
    && rm -rf /var/lib/apt/lists/*

# Install Python dependencies (CPU version of PaddlePaddle - using stable 2.x versions)
RUN pip install --no-cache-dir \
    paddlepaddle==2.6.2 \
    paddleocr==2.8.1 \
    fastapi \
    uvicorn[standard] \
    python-multipart \
    opencv-python-headless \
    pillow

# Copy server files
COPY image_support_files/paddleocr_server.py /app/paddleocr_server.py
COPY image_support_files/paddleocr-entrypoint.sh /usr/local/bin/paddleocr-entrypoint.sh
RUN chmod +x /usr/local/bin/paddleocr-entrypoint.sh

# Note: OCR models will be downloaded on first run
# This avoids build-time segfaults with certain CPU architectures

# Expose API port
EXPOSE 5000

# Health check (longer start-period for CPU variant)
HEALTHCHECK --interval=30s --timeout=10s --start-period=120s --retries=3 \
    CMD curl -f http://localhost:5000/health || exit 1

ENTRYPOINT ["/usr/local/bin/paddleocr-entrypoint.sh"]