llama-stack-agentic-sample/Containerfile at main · redhat-ai-dev/llama-stack-agentic-sample · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
# Llama Stack Agentic Application Container
# Multi-stage build for smaller, more secure image

# =============================================================================
# BUILDER STAGE
# =============================================================================
FROM registry.access.redhat.com/ubi9/python-312:9.7@sha256:92c71d1e64cf84b9aa6e8e81555397175b9367298b456d24eac5b55ab41fdab9 AS builder

USER root

# Allow uv to download Python 3.13 (required by pyproject.toml)
ENV UV_COMPILE_BYTECODE=0 \
    UV_LINK_MODE=copy

WORKDIR /app-root

# Install build dependencies (including mesa-libGL for RapidOCR/OpenCV model pre-download)
RUN dnf install -y gcc python3-devel make mesa-libGL && \
    dnf clean all && \
    pip3.12 install uv

# Copy dependency specification
COPY ./pyproject.toml ./uv.lock ./

# Install Python dependencies (uv will download Python 3.13 automatically)
RUN uv sync --frozen --no-dev

# Pre-download RapidOCR models during build (required for OpenShift where site-packages is read-only)
# Use torch engine for all components (Det, Cls, Rec) since onnxruntime is not in dependencies
RUN .venv/bin/python -c "from rapidocr import RapidOCR, EngineType; RapidOCR(params={'Det.engine_type': EngineType.TORCH, 'Cls.engine_type': EngineType.TORCH, 'Rec.engine_type': EngineType.TORCH})"

# =============================================================================
# RUNTIME STAGE
# =============================================================================
FROM registry.access.redhat.com/ubi9/python-312-minimal:9.7@sha256:2ac60c655288a88ec55df5e2154b9654629491e3c58b5c54450fb3d27a575cb6

ARG APP_ROOT=/app-root
WORKDIR /app-root

ENV PYTHONDONTWRITEBYTECODE=1 \
    PYTHONUNBUFFERED=1 \
    PYTHONCOERCECLOCALE=0 \
    PYTHONUTF8=1 \
    PYTHONIOENCODING=UTF-8 \
    LANG=en_US.UTF-8 \
    PATH="/app-root/.venv/bin:$PATH"

# Copy virtual environment from builder
COPY --from=builder --chown=1001:1001 /app-root/.venv ./.venv

# Copy application source code
COPY --chown=1001:1001 ./src/ ./src/
COPY --chown=1001:1001 ./streamlit_app.py ./

# Copy configuration (baked in - single source of truth)
COPY --chown=1001:1001 ./config/ ./config/

# Install runtime dependencies for docling/opencv (libGL for PDF processing)
USER root
RUN microdnf install -y mesa-libGL && \
    microdnf clean all

# Create directories for runtime data (don't chmod -R the whole app-root, it doubles image size!)
RUN mkdir -p /app-root/.llama && \
    chgrp -R 0 /app-root/.llama && \
    chmod -R g=u /app-root/.llama

# License
RUN mkdir -p /licenses
COPY LICENSE /licenses/

# Expose Streamlit port
EXPOSE 8501

USER 1001

# Default entrypoint runs the Streamlit app
CMD ["streamlit", "run", "streamlit_app.py", "--server.address=0.0.0.0"]

# Labels
LABEL com.redhat.component=llama-stack-agentic-app \
      description="Llama Stack Agentic AI Workflow Application" \
      io.k8s.description="Llama Stack Agentic AI Workflow Application" \
      io.k8s.display-name="Llama Stack Agentic App" \
      io.openshift.tags="ai,llama-stack,streamlit,agents,rag" \
      name=llama-stack-agentic-app \
      summary="Streamlit application for agentic AI workflows"