Dockerfile

# Use NVIDIA CUDA base image with Python 3.10 and CUDA 11.6
# Use CUDA 11.6 base image
FROM nvidia/cuda:11.6.2-cudnn8-runtime-ubuntu20.04

# Set environment variables
ENV DEBIAN_FRONTEND=noninteractive \
    PYTHONUNBUFFERED=1 \
    PYTHONDONTWRITEBYTECODE=1 \
    PIP_NO_CACHE_DIR=1 \
    PIP_DISABLE_PIP_VERSION_CHECK=1 \
    PIP_DEFAULT_TIMEOUT=100 \
    LD_LIBRARY_PATH=/usr/local/cuda-11.6/lib64:$LD_LIBRARY_PATH \
    NVIDIA_VISIBLE_DEVICES=all \
    NVIDIA_DRIVER_CAPABILITIES=compute,utility \
    GRADIO_SERVER_PORT=7860

# Install system dependencies, Python 3.10, and build tools
RUN apt-get update && apt-get install -y --no-install-recommends \
    software-properties-common \
    && add-apt-repository ppa:deadsnakes/ppa -y \
    && apt-get update \
    && apt-get install -y --no-install-recommends \
    # Python 3.10 and tools
    python3.10 python3.10-venv python3.10-dev \
    # Build dependencies for Python packages and SQLite
    build-essential cmake gcc g++ git curl wget tcl \
    # Math libraries
    libopenblas-dev libopenmpi-dev libomp-dev \
    # For data processing
    libffi-dev libssl-dev \
    # Clean up
    && rm -rf /var/lib/apt/lists/* \
    && apt-get clean

# Download, compile, and install SQLite
RUN wget -O /tmp/sqlite.tar.gz https://www.sqlite.org/2022/sqlite-autoconf-3400100.tar.gz \
    && mkdir -p /tmp/sqlite \
    && tar xzf /tmp/sqlite.tar.gz -C /tmp/sqlite --strip-components=1 \
    && cd /tmp/sqlite \
    && ./configure --prefix=/usr/local \
    && make -j$(nproc) \
    && make install \
    && ldconfig \
    && rm -rf /tmp/sqlite /tmp/sqlite.tar.gz

# Set Python 3.10 as default and install pip
RUN update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.10 1 \
    && update-alternatives --install /usr/bin/python python /usr/bin/python3 1 \
    && curl -sS https://bootstrap.pypa.io/get-pip.py | python3.10

# Set environment variables for ChromaDB and pysqlite3
ENV LD_LIBRARY_PATH=/usr/local/lib:${LD_LIBRARY_PATH} \
    PYSQLITE3_BUNDLED=1 \
    CFLAGS="-DSQLITE_MAX_VARIABLE_NUMBER=250000 -DSQLITE_MAX_EXPR_DEPTH=10000"

# Install pysqlite3 from source and other base python packages
RUN python3 -m pip install --no-cache-dir --upgrade setuptools wheel \
    && python3 -m pip install --no-cache-dir --no-binary pysqlite3 pysqlite3


# Set working directory
WORKDIR /app

# Copy requirements first to leverage Docker cache
COPY requirements.txt .

# Install PyTorch with CUDA 11.6 support first
RUN python3 -m pip install --no-cache-dir \
    torch==1.13.1+cu116 \
    torchvision==0.14.1+cu116 \
    torchaudio==0.13.1 \
    --extra-index-url https://download.pytorch.org/whl/cu116

# Install remaining requirements
RUN python3 -m pip install --no-cache-dir -r requirements.txt

# Verify PyTorch installation with CUDA
RUN python3 -c "import torch; print(f'PyTorch version: {torch.__version__}'); print(f'CUDA available: {torch.cuda.is_available()}'); print(f'CUDA version: {torch.version.cuda}')"

# Copy application code
COPY . .

# Remove any existing database directory to ensure a fresh start
RUN rm -rf ./chroma_db

# Expose the port Gradio runs on
EXPOSE 7860

# Download the sentence transformer model during build
RUN python3 -c "from sentence_transformers import SentenceTransformer; SentenceTransformer('pritamdeka/S-Biomed-Roberta-snli-multinli-stsb')"

# Default command to run the app
CMD ["python3", "app.py"]