Dockerfile.cpu

# CoDRAG Headless — CPU Image
# For CI/CD runners (GitHub Actions, GitLab CI) using BYOK cloud LLMs.
# No GPU required. ONNX embeddings run natively on CPU.
#
# Build:
#   docker build -f Dockerfile.cpu -t codrag/headless:cpu .
#
# Usage:
#   codrag sync-headless \
#     --repo-path /workspace \
#     --model-provider openai --model-name gpt-4.1-mini \
#     --embedder native

FROM ubuntu:22.04 AS base

ENV DEBIAN_FRONTEND=noninteractive \
    PYTHONUNBUFFERED=1 \
    PIP_NO_CACHE_DIR=1

# System dependencies
RUN apt-get update && apt-get install -y --no-install-recommends \
    python3.11 python3.11-venv python3-pip \
    git openssh-client curl ca-certificates \
    && rm -rf /var/lib/apt/lists/*

RUN update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.11 1 \
    && update-alternatives --install /usr/bin/python python /usr/bin/python3.11 1

# ── Install CoDRAG ───────────────────────────────────────────
WORKDIR /opt/codrag

# Copy the CoDRAG source and install
# In CI, this is replaced by a pip install from the published package
COPY pyproject.toml .
COPY src/ src/
COPY engine/ engine/

RUN pip install --no-cache-dir ".[headless]"

# ── Pre-download ONNX embedding model ────────────────────────
# nomic-embed-text-v1.5 quantized (~132 MB), cached in /root/.cache/huggingface
RUN python -c "from codrag.core.embedder import NativeEmbedder; NativeEmbedder().download_model()" \
    || echo "WARN: Model pre-download skipped (will download on first run)"

# ── S3 sync dependencies ─────────────────────────────────────
RUN pip install --no-cache-dir boto3 minio

# ── Default entrypoint ────────────────────────────────────────
ENTRYPOINT ["codrag"]
CMD ["sync-headless", "--help"]