# =============================================================================
# Dockerfile.gpu — CUDA-accelerated build for DeepGuard AI
#
# This variant uses NVIDIA's CUDA 12.1 runtime image so PyTorch can run
# on the GPU, making model inference 5-10× faster than CPU.
#
# Prerequisites:
#   - NVIDIA GPU with driver >= 525.60.13 (Linux) / 528.33 (Windows WSL2)
#   - NVIDIA Container Toolkit: https://docs.nvidia.com/datacenter/cloud-native/container-toolkit
#
# Build:
#   docker build -f Dockerfile.gpu -t deepguard:gpu .
#
# Run:
#   docker run --gpus all -p 8501:8501 deepguard:gpu
#
# With docker-compose (add to your docker-compose.yml):
#   services:
#     streamlit:
#       image: deepguard:gpu
#       deploy:
#         resources:
#           reservations:
#             devices:
#               - driver: nvidia
#                 count: all
#                 capabilities: [gpu]
# =============================================================================

# ── Stage 1: System deps ───────────────────────────────────────────────────
FROM nvidia/cuda:12.1.1-runtime-ubuntu22.04 AS base

# Install system packages needed by OpenCV, MediaPipe, and PyTorch
RUN apt-get update && apt-get install -y --no-install-recommends \
    python3.11 \
    python3-pip \
    python3-dev \
    libgl1-mesa-glx \
    libglib2.0-0 \
    libsm6 \
    libxext6 \
    libxrender-dev \
    libgomp1 \
    libgles2-mesa \
    libegl1 \
    libomp-dev \
    && rm -rf /var/lib/apt/lists/*

# Symlink python3 → python for compatibility
RUN ln -sf /usr/bin/python3.11 /usr/bin/python

# ── Stage 2: Python deps ────────────────────────────────────────────────────
FROM base AS deps

WORKDIR /app
COPY requirements.txt ./

# Install PyTorch with CUDA 12.x support first (separate line for layer caching)
RUN pip install --no-cache-dir --upgrade pip
RUN pip install --no-cache-dir torch torchvision --index-url https://download.pytorch.org/whl/cu121

# Then the rest of the dependencies
RUN pip install --no-cache-dir \
    streamlit>=1.30 \
    opencv-python>=4.8 \
    pillow>=10.0 \
    numpy>=1.24 \
    transformers>=4.36 \
    mediapipe>=0.10.9 \
    huggingface-hub>=0.20 \
    fastapi>=0.100.0 \
    uvicorn[standard]>=0.23.0 \
    python-multipart>=0.0.6 \
    celery>=5.3.0 \
    redis>=5.0.0 \
    slowapi>=0.1.9 \
    pytest>=8.0.0 \
    gunicorn>=21.2.0

# ── Stage 3: Runtime ───────────────────────────────────────────────────────
FROM deps AS runtime

WORKDIR /app
COPY . .

RUN mkdir -p static/scans logs

# ── Ports ──────────────────────────────────────────────────────────────────
EXPOSE 8501
EXPOSE 8000

# ── Environment ────────────────────────────────────────────────────────────
ENV STREAMLIT_SERVER_PORT=8501
ENV STREAMLIT_SERVER_ADDRESS=0.0.0.0
ENV STREAMLIT_SERVER_HEADLESS=true
ENV DF_LOG_LEVEL=INFO
ENV CUDA_VISIBLE_DEVICES=all

# ── Default: Streamlit UI ──────────────────────────────────────────────────
CMD ["streamlit", "run", "app.py", \
     "--server.port=8501", \
     "--server.address=0.0.0.0", \
     "--server.headless=true"]

# ── Health check ──────────────────────────────────────────────────────────
HEALTHCHECK --interval=30s --timeout=10s --start-period=120s --retries=3 \
    CMD python -c "import urllib.request; urllib.request.urlopen('http://localhost:8501/_stcore/health')" || exit 1