ci: Add integration tests to CI workflow

.github/workflows/redhat-distro-container.yml

@@ -24,6 +24,9 @@ env:
 jobs:
   build-test-push:
     runs-on: ubuntu-latest
+    env:
+      INFERENCE_MODEL: meta-llama/Llama-3.2-1B-Instruct
+      VLLM_URL: http://localhost:8000/v1
     strategy:
       matrix:
         platform: [linux/amd64] # TODO: enable other arch once all pip packages are available.
@@ -32,6 +35,12 @@ jobs:
       - name: Checkout repository
         uses: actions/checkout@08c6903cd8c0fde910a37f88322edcfb5dd907a8 # v5.0.0
 
+      - name: Install uv
+        uses: astral-sh/setup-uv@6b9c6063abd6010835644d4c2e1bef4cf5cd0fca # v6.0.1
+        with:
+          python-version: 3.12
+          version: 0.7.6
+
       - name: Set up QEMU
         uses: docker/setup-qemu-action@29109295f81e9208d7d86ff1c6c12d2833863392 # v3.6.0
 
@@ -55,14 +64,22 @@ jobs:
         id: vllm
         uses: ./.github/actions/setup-vllm
 
-      - name: Smoke test image
+      - name: Start and smoke test LLS distro image
         id: smoke-test
         shell: bash
-        env:
-          INFERENCE_MODEL: meta-llama/Llama-3.2-1B-Instruct
-          VLLM_URL: http://localhost:8000/v1
         run: ./tests/smoke.sh
 
+      - name: Integration tests
+        id: integration-tests
+        shell: bash
+        run: ./tests/run_integration_tests.sh
+
+      - name: cleanup
+        if: always()
+        shell: bash
+        run: |
+          docker rm -f vllm llama-stack
+
       - name: Log in to Quay.io
         id: login
         if: github.event_name == 'push'

tests/run_integration_tests.sh

@@ -0,0 +1,75 @@
+#!/usr/bin/env bash
+
+set -euo pipefail
+
+# Configuration
+LLAMA_STACK_REPO="https://github.com/meta-llama/llama-stack.git"
+WORK_DIR="/tmp/llama-stack-integration-tests"
+INFERENCE_MODEL="${INFERENCE_MODEL:-meta-llama/Llama-3.2-1B-Instruct}"
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+
+# Get version dynamically from Containerfile.in (look in parent directory)
+CONTAINERFILE_IN="$SCRIPT_DIR/../distribution/Containerfile.in"
+LLAMA_STACK_VERSION=$(grep -o 'llama-stack==[0-9]\+\.[0-9]\+\.[0-9]\+' "$CONTAINERFILE_IN" | cut -d'=' -f3)
+if [ -z "$LLAMA_STACK_VERSION" ]; then
+    echo "Error: Could not extract llama-stack version from Containerfile.in"
+    exit 1
+fi
+
+function clone_llama_stack() {
+    # Clone the repository if it doesn't exist
+    if [ ! -d "$WORK_DIR" ]; then
+        git clone "$LLAMA_STACK_REPO" "$WORK_DIR"
+    fi
+
+    # Checkout the specific tag
+    cd "$WORK_DIR"
+    # fetch origin incase we didn't clone a fresh repo
+    git fetch origin
+    if ! git checkout "v$LLAMA_STACK_VERSION"; then
+        echo "Error: Could not checkout tag v$LLAMA_STACK_VERSION"
+        echo "Available tags:"
+        git tag | grep "^v" | tail -10
+        exit 1
+    fi
+}
+
+function run_integration_tests() {
+    echo "Running integration tests..."
+
+    cd "$WORK_DIR"
+
+    # Test to skip
+    SKIP_TESTS="test_text_chat_completion_tool_calling_tools_not_in_request or test_inference_store_tool_calls"
+
+    # Dynamically determine the path to run.yaml from the original script directory
+    STACK_CONFIG_PATH="$SCRIPT_DIR/../distribution/run.yaml"
+    if [ ! -f "$STACK_CONFIG_PATH" ]; then
+        echo "Error: Could not find stack config at $STACK_CONFIG_PATH"
+        exit 1
+    fi
+
+    uv run pytest -s -v tests/integration/inference/ \
+        --stack-config=server:"$STACK_CONFIG_PATH" \
+        --text-model=vllm-inference/"$INFERENCE_MODEL" \
+        -k "not ($SKIP_TESTS)"
+}
+
+function main() {
+    echo "Starting llama-stack integration tests"
+    echo "Configuration:"
+    echo "  LLAMA_STACK_VERSION: $LLAMA_STACK_VERSION"
+    echo "  LLAMA_STACK_REPO: $LLAMA_STACK_REPO"
+    echo "  WORK_DIR: $WORK_DIR"
+    echo "  INFERENCE_MODEL: $INFERENCE_MODEL"
+
+    clone_llama_stack
+    run_integration_tests
+
+    echo "Integration tests completed successfully!"
+}
+
+
+main "$@"
+exit 0

tests/smoke.sh

@@ -69,6 +69,5 @@ main() {
   echo "===> Smoke test completed successfully!"
 }
 
-trap 'docker rm -f -v llama-stack >/dev/null 2>&1 || true' EXIT
 main "$@"
 exit 0