# CPU-only ML Training Service Dockerfile
# Fallback for environments without NVIDIA Container Runtime

# =============================================================================
# Builder Stage
# =============================================================================
FROM rust:1.83-slim AS builder

# Install build dependencies
RUN apt-get update && apt-get install -y \
    build-essential \
    pkg-config \
    libssl-dev \
    protobuf-compiler \
    libpq-dev \
    ca-certificates \
    git \
    curl \
    && rm -rf /var/lib/apt/lists/*

# Set working directory
WORKDIR /build

# Enable git fetch via CLI for cargo
ENV CARGO_NET_GIT_FETCH_WITH_CLI=true

# Copy workspace manifests
COPY Cargo.toml Cargo.lock ./

# Copy workspace members
COPY trading_engine ./trading_engine
COPY risk ./risk
COPY risk-data ./risk-data
COPY trading-data ./trading-data
COPY tli ./tli
COPY ml ./ml
COPY ml-data ./ml-data
COPY data ./data
COPY backtesting ./backtesting
COPY adaptive-strategy ./adaptive-strategy
COPY common ./common
COPY storage ./storage
COPY model_loader ./model_loader
COPY market-data ./market-data
COPY database ./database
COPY config ./config
COPY services/backtesting_service ./services/backtesting_service
COPY services/trading_service ./services/trading_service
COPY services/ml_training_service ./services/ml_training_service
COPY services/api_gateway ./services/api_gateway
COPY services/load_tests ./services/load_tests
COPY services/stress_tests ./services/stress_tests
COPY services/integration_tests ./services/integration_tests
COPY tests ./tests
COPY migrations ./migrations

# Build without CUDA features
RUN cargo build --release \
    --package ml_training_service \
    --no-default-features

# =============================================================================
# Runtime Stage
# =============================================================================
FROM debian:bookworm-slim

# Install runtime dependencies
RUN apt-get update && apt-get install -y \
    ca-certificates \
    libssl3 \
    libpq5 \
    curl \
    && rm -rf /var/lib/apt/lists/*

# Download and install grpc_health_probe
RUN curl -sSL https://github.com/grpc-ecosystem/grpc-health-probe/releases/download/v0.4.25/grpc_health_probe-linux-amd64 \
    -o /usr/local/bin/grpc_health_probe && \
    chmod +x /usr/local/bin/grpc_health_probe

# Create non-root user
RUN groupadd --system --gid 1000 foxhunt && \
    useradd --system --uid 1000 --gid foxhunt --shell /bin/bash foxhunt

# Create application directories
RUN mkdir -p /app/config /app/logs /app/models /app/checkpoints && \
    chown -R foxhunt:foxhunt /app

# Set working directory
WORKDIR /app

# Copy binary from builder
COPY --from=builder /build/target/release/ml_training_service ./ml_training_service
RUN chmod +x ./ml_training_service

# Switch to non-root user
USER foxhunt

# Expose gRPC and metrics ports
EXPOSE 50053 9094

# Health check using grpc_health_probe
HEALTHCHECK --interval=10s --timeout=5s --start-period=30s --retries=3 \
    CMD /usr/local/bin/grpc_health_probe -addr=localhost:50053 || exit 1

# Run the application
ENTRYPOINT ["./ml_training_service"]
CMD ["serve"]