Dockerfile

FROM python:3.11-slim-bullseye

# Install system dependencies with clean up in single RUN layer
RUN sed -i 's/main/main contrib non-free/' /etc/apt/sources.list && \
    apt-get update && \
    apt-get install -y --no-install-recommends \
    unrar \
    libgl1 \
    libglib2.0-0 \
    # ffmpeg for video processing
    ffmpeg \
    # git for transformers dependencies
    git \
    && rm -rf /var/lib/apt/lists/*

WORKDIR /app

# Upgrade pip and install core dependencies first
RUN pip install --no-cache-dir --upgrade pip setuptools wheel packaging

# Install optimized CPU-only PyTorch stack
RUN pip install --no-cache-dir \
    "numpy<2" \
    torch==2.1.2+cpu \
    torchvision==0.16.2+cpu \
    torchaudio==2.1.2+cpu \
    --index-url https://download.pytorch.org/whl/cpu

# Copy requirements and install with special handling
COPY requirements.txt .
RUN pip install --no-cache-dir \
    -r requirements.txt \
    --find-links https://download.pytorch.org/whl/cpu \
    --extra-index-url https://pypi.org/simple

# Install transformers with specific version that works with BLIP-2
RUN pip install --no-cache-dir \
    accelerate==0.27.2 \
    transformers==4.36.2 \
    timm==0.9.12 \
    einops==0.7.0 \
    bitsandbytes==0.42.0  # Helps with memory optimization

# For BLIP-2 specifically, we need these
RUN pip install --no-cache-dir \
    protobuf==3.20.3 \
    sentencepiece==0.2.0

# Copy application code with proper permissions
COPY --chown=1000:1000 . .

# Create non-root user and set permissions
RUN useradd -m -u 1000 user && \
    chown -R user:user /app && \
    mkdir -p /app/downloads /app/extracted /app/extracted_frames /app/analysis_results && \
    chown -R user:user /app/*

USER user

# Environment variables for better performance and less logging
ENV HF_HUB_DISABLE_PROGRESS=1
ENV TF_CPP_MIN_LOG_LEVEL=3
ENV PYTORCH_NO_CUDA_MEMORY_CACHING=1
ENV TOKENIZERS_PARALLELISM=false
ENV PYTHONUNBUFFERED=1

# Set this if you're only using CPU
ENV CUDA_VISIBLE_DEVICES=""

# Pre-download the BLIP-2 model during build (optional)
# RUN python -c "from transformers import AutoModel; AutoModel.from_pretrained('Salesforce/blip2-opt-2.7b', trust_remote_code=True)"

CMD ["uvicorn", "download_api:app", "--host", "0.0.0.0", "--port", "7860", "--workers", "1"]