feat: Model Streamer initial support

.dockerignore

@@ -3,4 +3,13 @@
 
 .dockerignore
 .git/
-Dockerfile
+Dockerfile
+
+# Build artifacts
+target/
+
+# Python virtual environments
+.venv/
+venv/
+__pycache__/
+*.pyc

.github/workflows/ci.yml

@@ -213,6 +213,62 @@ jobs:
         files: lcov.info
         fail_ci_if_error: false
 
+  sidecar-test:
+    name: Sidecar Python Tests
+    runs-on: ubuntu-latest
+
+    steps:
+    - name: Checkout repository
+      uses: actions/checkout@v4
+
+    - name: Set up Python
+      uses: actions/setup-python@v5
+      with:
+        python-version: '3.12'
+
+    - name: Install sidecar dependencies
+      run: |
+        python -m pip install --upgrade pip
+        pip install -e "sidecar[dev]"
+
+    - name: Run sidecar unit tests (no MinIO)
+      run: |
+        cd sidecar
+        pytest tests/test_api.py -v
+
+    - name: Start MinIO for integration tests
+      run: |
+        docker run -d --name minio \
+          -p 9000:9000 -p 9001:9001 \
+          -e MINIO_ROOT_USER=minioadmin \
+          -e MINIO_ROOT_PASSWORD=minioadmin \
+          minio/minio:latest server /data --console-address ":9001"
+        # Wait for MinIO to be ready
+        sleep 5
+        # Create test bucket and upload test data
+        docker run --rm --network host --entrypoint /bin/sh minio/mc -c "
+          mc alias set myminio http://localhost:9000 minioadmin minioadmin &&
+          mc mb myminio/models --ignore-existing &&
+          echo '{\"model_type\": \"test\"}' > /tmp/config.json &&
+          mc cp /tmp/config.json myminio/models/test-model/config.json
+        "
+
+    - name: Run sidecar integration tests with MinIO
+      env:
+        AWS_ACCESS_KEY_ID: minioadmin
+        AWS_SECRET_ACCESS_KEY: minioadmin
+        AWS_ENDPOINT_URL: http://localhost:9000
+        AWS_REGION: us-east-1
+      run: |
+        cd sidecar
+        pytest tests/ -v
+
+    - name: Stop MinIO
+      if: always()
+      run: |
+        docker stop minio || true
+        docker rm minio || true
+
   docker-build:
     name: Docker Build
     runs-on: ubuntu-latest

.gitignore

@@ -51,4 +51,10 @@ logs/
 models.db
 **/__pycache__/
 docs/FEEDBACK.md
-docs/optimization-comparison.md
+
+# Python
+.venv/
+__pycache__/
+*.pyc
+*.pyo
+*.egg-info/

Cargo.toml

@@ -50,9 +50,11 @@ tonic-build = "0.13"
 tracing = "0.1"
 tracing-subscriber = { version = "0.3", features = ["env-filter"] }
 futures = "0.3"
+reqwest = { version = "0.12", features = ["json", "rustls-tls"], default-features = false }
 uuid = { version = "1.17", features = ["v4", "serde"] }
 thiserror = "2.0"
 redis = { version = "0.27", features = ["tokio-comp", "connection-manager"] }
+urlencoding = "2.1"
 
 # Kubernetes client (kube-rs) for CRD-based metadata backend
 kube = { version = "0.98", features = ["client", "runtime", "derive", "rustls-tls"], default-features = false }

Dockerfile

@@ -23,11 +23,23 @@ FROM nvcr.io/nvidia/base/ubuntu:noble-20250619 AS runtime
 
 WORKDIR /app
 
-# Install runtime dependencies
+# Install runtime dependencies including Python for the sidecar
 RUN apt-get update && \
-    apt-get install -y --no-install-recommends ca-certificates libssl-dev gpgv && \
+    apt-get install -y --no-install-recommends \
+        ca-certificates \
+        gpgv \
+        libssl-dev \
+        python3 \
+        python3-pip \
+        python3-venv && \
     rm -rf /var/lib/apt/lists/*
 
+# Create Python virtual environment and install sidecar package
+COPY sidecar/pyproject.toml /app/sidecar/
+COPY sidecar/src /app/sidecar/src
+RUN python3 -m venv /app/venv && \
+    /app/venv/bin/pip install --no-cache-dir /app/sidecar
+
 # Copy all built binaries
 COPY --from=builder /app/target/release/modelexpress-server .
 COPY --from=builder /app/target/release/modelexpress-cli .
@@ -37,15 +49,23 @@ COPY --from=builder /app/target/release/fallback_test .
 # Copy the Attribution files
 COPY ATTRIBUTIONS_Rust.md .
 
-# Expose the default port
+# Copy the entrypoint script
+COPY entrypoint.sh .
+RUN chmod +x /app/entrypoint.sh
+
+# Expose the default ports (gRPC server and sidecar)
 EXPOSE 8001
+EXPOSE 8002
 
 # Set default environment variables (can be overridden)
 ENV MODEL_EXPRESS_SERVER_PORT=8001
+ENV MODEL_EXPRESS_SIDECAR_PORT=8002
+ENV MODEL_EXPRESS_SIDECAR_ENDPOINT=http://127.0.0.1:8002
 ENV MODEL_EXPRESS_LOG_LEVEL=info
 ENV MODEL_EXPRESS_DATABASE_PATH=/app/models.db
 ENV MODEL_EXPRESS_CACHE_DIRECTORY=/app/cache
 ENV HF_HUB_CACHE=/app/cache
+ENV PYTHONPATH=/app/sidecar/src
 
-# Run the server by default
-CMD ["./modelexpress-server"]
+# Run both services via entrypoint
+ENTRYPOINT ["/app/entrypoint.sh"]

docker-compose.yml

@@ -10,7 +10,58 @@ services:
       dockerfile: Dockerfile
     ports:
       - "8001:8001"
+      - "8002:8002"
     environment:
       - MODEL_EXPRESS_SERVER_PORT=8001
+      - MODEL_EXPRESS_SIDECAR_PORT=8002
+      - MODEL_EXPRESS_SIDECAR_ENDPOINT=http://127.0.0.1:8002
       - MODEL_EXPRESS_LOG_LEVEL=info
+      - MODEL_EXPRESS_CACHE_DIRECTORY=/app/cache
+      # S3 credentials for MinIO (optional, can be passed at runtime)
+      - AWS_ACCESS_KEY_ID=${AWS_ACCESS_KEY_ID:-minioadmin}
+      - AWS_SECRET_ACCESS_KEY=${AWS_SECRET_ACCESS_KEY:-minioadmin}
+      - AWS_ENDPOINT_URL=http://minio:9000
+      - AWS_REGION=us-east-1
+    volumes:
+      - model-cache:/app/cache
+    depends_on:
+      - minio
     restart: unless-stopped
+
+  # MinIO for S3-compatible storage testing
+  minio:
+    image: minio/minio:latest
+    ports:
+      - "9000:9000"
+      - "9001:9001"
+    environment:
+      - MINIO_ROOT_USER=${MINIO_ROOT_USER:-minioadmin}
+      - MINIO_ROOT_PASSWORD=${MINIO_ROOT_PASSWORD:-minioadmin}
+    command: server /data --console-address ":9001"
+    volumes:
+      - minio-data:/data
+    healthcheck:
+      test: ["CMD", "curl", "-f", "http://localhost:9000/minio/health/live"]
+      interval: 30s
+      timeout: 20s
+      retries: 3
+    restart: unless-stopped
+
+  # MinIO client for bucket initialization (optional)
+  minio-init:
+    image: minio/mc:latest
+    depends_on:
+      - minio
+    entrypoint: >
+      /bin/sh -c "
+      sleep 5;
+      mc alias set myminio http://minio:9000 minioadmin minioadmin;
+      mc mb myminio/models --ignore-existing;
+      mc anonymous set download myminio/models;
+      echo 'MinIO initialization complete';
+      exit 0;
+      "
+
+volumes:
+  model-cache:
+  minio-data:

entrypoint.sh

@@ -0,0 +1,56 @@
+#!/bin/bash
+# SPDX-FileCopyrightText: Copyright (c) 2025-2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+
+# Entrypoint script that runs both the Model Streamer sidecar and the ModelExpress server.
+
+set -e
+
+# Configuration
+SIDECAR_PORT="${MODEL_EXPRESS_SIDECAR_PORT:-8002}"
+LOG_LEVEL="${MODEL_EXPRESS_LOG_LEVEL:-info}"
+
+echo "Starting ModelExpress with Model Streamer sidecar..."
+
+# Start the Python sidecar in the background
+echo "Starting Model Streamer sidecar on port ${SIDECAR_PORT}..."
+/app/venv/bin/python -m uvicorn \
+    modelexpress_sidecar.main:app \
+    --host 127.0.0.1 \
+    --port "${SIDECAR_PORT}" \
+    --log-level "${LOG_LEVEL}" &
+
+SIDECAR_PID=$!
+
+# Wait for sidecar to be ready
+echo "Waiting for sidecar to be ready..."
+MAX_RETRIES=30
+RETRY_COUNT=0
+while [ $RETRY_COUNT -lt $MAX_RETRIES ]; do
+    if /app/venv/bin/python -c "import httpx; httpx.get('http://127.0.0.1:${SIDECAR_PORT}/health', timeout=2)" > /dev/null 2>&1; then
+        echo "Sidecar is ready!"
+        break
+    fi
+    RETRY_COUNT=$((RETRY_COUNT + 1))
+    sleep 1
+done
+
+if [ $RETRY_COUNT -eq $MAX_RETRIES ]; then
+    echo "Warning: Sidecar health check timed out, proceeding anyway..."
+fi
+
+# Trap to ensure cleanup on exit
+cleanup() {
+    echo "Shutting down..."
+    if [ -n "$SIDECAR_PID" ]; then
+        kill "$SIDECAR_PID" 2>/dev/null || true
+    fi
+}
+trap cleanup EXIT INT TERM
+
+# Start the main server. Run it in the background and wait so the shell stays
+# alive to handle signals and the cleanup trap can fire to kill the sidecar.
+echo "Starting ModelExpress server..."
+/app/modelexpress-server "$@" &
+SERVER_PID=$!
+wait $SERVER_PID

modelexpress_common/Cargo.toml

@@ -28,6 +28,8 @@ chrono = { workspace = true }
 clap = { workspace = true }
 config = { workspace = true }
 jiff = { workspace = true }
+reqwest = { workspace = true }
+urlencoding = { workspace = true }
 
 [dev-dependencies]
 mockall = "0.13"

modelexpress_common/proto/model.proto

@@ -43,6 +43,7 @@ enum ModelStatus {
 // Enum representing model providers
 enum ModelProvider {
   HUGGING_FACE = 0;
+  MODEL_STREAMER = 1;
 }
 
 // Request for streaming model files