Merge branch 'main' into feat/intent-classification-lora-auto-detection

Author: rootfs

.github/workflows/integration-test-docker.yml

@@ -21,10 +21,10 @@ concurrency:
   cancel-in-progress: true
 
 jobs:
-  test-quickstart:
+  test-ci-compose:
     if: github.repository == 'vllm-project/semantic-router' && !github.event.pull_request.draft
     runs-on: ubuntu-latest
-    timeout-minutes: 30
+    timeout-minutes: 20  # Reduced from 30 - CI compose is faster
 
     steps:
       - name: Check out the repo
@@ -46,33 +46,93 @@ jobs:
         with:
           python-version: '3.11'
 
-      - name: Install system dependencies
+      - name: Install dependencies
         run: |
           sudo apt-get update
-          sudo apt-get install -y \
-            make \
-            curl \
-            docker-compose
+          sudo apt-get install -y make curl
+          pip install huggingface_hub[cli]
 
-      - name: Run quickstart script
-        id: quickstart
+      - name: Download models
         run: |
-          timeout 1200 bash scripts/quickstart.sh || {
-            exit_code=$?
-            if [ $exit_code -eq 124 ]; then
-              echo "::error::Quickstart script timed out after 20 minutes"
-            else
-              echo "::error::Quickstart script failed with exit code $exit_code"
-            fi
-            exit $exit_code
-          }
+          echo "Downloading minimal models for CI..."
+          make download-models
         env:
           CI: true
           CI_MINIMAL_MODELS: true
-          TERM: xterm
           HF_HUB_ENABLE_HF_TRANSFER: 1
           HF_HUB_DISABLE_TELEMETRY: 1
 
+      - name: Start CI services
+        run: |
+          echo "Starting minimal CI services (semantic-router, envoy, llm-katan)..."
+          make docker-compose-up-ci
+        env:
+          CI: true
+
+      - name: Wait for services to be healthy
+        run: |
+          echo "Waiting for services to be healthy..."
+          max_attempts=60
+          attempt=1
+          
+          while [ $attempt -le $max_attempts ]; do
+            echo "Attempt $attempt/$max_attempts: Checking service health..."
+            
+            # Check semantic-router health
+            if docker ps --filter "name=semantic-router" --filter "health=healthy" --format "{{.Names}}" | grep -q "semantic-router"; then
+              echo "✅ semantic-router is healthy"
+              
+              # Check envoy health
+              if docker ps --filter "name=envoy-proxy" --filter "health=healthy" --format "{{.Names}}" | grep -q "envoy-proxy"; then
+                echo "✅ envoy-proxy is healthy"
+                
+                # Check llm-katan health
+                if docker ps --filter "name=llm-katan" --filter "health=healthy" --format "{{.Names}}" | grep -q "llm-katan"; then
+                  echo "✅ llm-katan is healthy"
+                  echo "🎉 All services are healthy!"
+                  exit 0
+                fi
+              fi
+            fi
+            
+            # Show current status
+            docker ps --format "table {{.Names}}\t{{.Status}}" | grep -E "NAMES|semantic-router|envoy|llm-katan" || true
+            
+            sleep 5
+            ((attempt++))
+          done
+          
+          echo "❌ Timeout waiting for services to be healthy"
+          docker ps -a
+          exit 1
+
+      - name: Test semantic router health endpoint
+        run: |
+          echo "Testing semantic router health..."
+          curl -f http://localhost:8080/health || {
+            echo "❌ Health check failed"
+            exit 1
+          }
+          echo "✅ Health check passed"
+
+      - name: Test envoy proxy endpoint
+        run: |
+          echo "Testing envoy proxy..."
+          curl -f http://localhost:19000/ready || {
+            echo "❌ Envoy ready check failed"
+            exit 1
+          }
+          echo "✅ Envoy is ready"
+
+      - name: Test llm-katan endpoint
+        run: |
+          echo "Testing llm-katan..."
+          curl -f http://localhost:8002/health || {
+            echo "❌ LLM-Katan health check failed"
+            exit 1
+          }
+          echo "✅ LLM-Katan is healthy"
+
       - name: Test semantic routing functionality
         run: |
           echo "Testing semantic router with a sample query..."
@@ -85,24 +145,31 @@ jobs:
               "temperature": 0.7
             }')
 
-          echo "Full response: $response"
+          echo "Response: $response"
+          
+          # Verify we got a response
+          if echo "$response" | grep -q "choices"; then
+            echo "✅ Chat completions test passed"
+          else
+            echo "⚠️ Response may not contain expected fields, but request succeeded"
+          fi
 
       - name: Show service logs on failure
         if: failure()
         run: |
           echo "=== Docker Compose Logs ==="
-          docker compose -f deploy/docker-compose/docker-compose.yml logs
+          make docker-compose-logs-ci || docker compose -f deploy/docker-compose/docker-compose.ci.yml logs
           echo "=== Container Status ==="
           docker ps -a
           echo "=== Semantic Router Logs ==="
-          docker logs semantic-router || true
+          docker logs semantic-router 2>&1 | tail -100 || true
           echo "=== Envoy Logs ==="
-          docker logs envoy-proxy || true
-          echo "=== Dashboard Logs ==="
-          docker logs semantic-router-dashboard || true
+          docker logs envoy-proxy 2>&1 | tail -100 || true
+          echo "=== LLM-Katan Logs ==="
+          docker logs llm-katan 2>&1 | tail -100 || true
 
       - name: Clean up
         if: always()
         run: |
-          make docker-compose-down || true
+          make docker-compose-down-ci || true
           docker system prune -af --volumes || true

dashboard/Dockerfile

@@ -2,7 +2,7 @@
 FROM node:18-alpine AS frontend-builder
 WORKDIR /app/frontend
 COPY frontend/package*.json ./
-RUN npm ci
+RUN npm ci --include=dev   # Ensure devDependencies are installed regardless of NODE_ENV
 COPY frontend/ ./
 RUN npm run build
 

dashboard/backend/Dockerfile

@@ -5,7 +5,7 @@ COPY dashboard/frontend/package.json dashboard/frontend/package-lock.json dashbo
 COPY dashboard/frontend/src ./src
 COPY dashboard/frontend/public ./public
 COPY dashboard/frontend/index.html ./
-RUN npm install
+RUN npm install --include=dev  # Ensure TypeScript/Vite are installed regardless of NODE_ENV
 RUN npm run build
 
 # Stage 2: Build backend with Go

deploy/docker-compose/docker-compose.ci.yml

@@ -0,0 +1,92 @@
+# Minimal Docker Compose for CI testing
+# This file contains only essential services needed for integration testing.
+# Excludes: grafana, prometheus, jaeger, openwebui, chat-ui, pipelines, mongo, dashboard
+#
+# Usage:
+#   make docker-compose-up-ci
+#   # or directly:
+#   docker compose -f deploy/docker-compose/docker-compose.ci.yml up -d
+
+services:
+
+  # Semantic Router External Processor Service
+  semantic-router:
+    image: ghcr.io/vllm-project/semantic-router/extproc:latest
+    container_name: semantic-router
+    ports:
+      - "50051:50051"  # gRPC for ExtProc
+      - "8080:8080"    # HTTP API (health, classify, metrics)
+    volumes:
+      - ../../config:/app/config:ro,z
+      - ../../models:/app/models:ro,z
+      - ~/.cache/huggingface:/root/.cache/huggingface:z
+    environment:
+      - LD_LIBRARY_PATH=/app/lib
+      - CONFIG_FILE=${CONFIG_FILE:-/app/config/config.yaml}
+      - HUGGINGFACE_HUB_CACHE=/root/.cache/huggingface
+      - HF_HUB_ENABLE_HF_TRANSFER=1
+    networks:
+      - semantic-network
+    healthcheck:
+      test: ["CMD", "curl", "-f", "localhost:8080/health"]
+      interval: 10s
+      timeout: 5s
+      retries: 5
+      start_period: 30s
+
+  # Envoy Proxy Service
+  envoy:
+    image: envoyproxy/envoy:v1.31.7
+    container_name: envoy-proxy
+    security_opt:
+      - label=disable
+    ports:
+      - "8801:8801"  # Main proxy port
+      - "19000:19000"  # Admin interface
+    volumes:
+      - ./addons/envoy.yaml:/etc/envoy/envoy.yaml:ro,z
+    command: ["/usr/local/bin/envoy", "-c", "/etc/envoy/envoy.yaml", "--component-log-level", "ext_proc:debug,router:debug"]
+    depends_on:
+      semantic-router:
+        condition: service_healthy
+    networks:
+      - semantic-network
+    healthcheck:
+      test: ["CMD", "bash", "-c", "(echo -e 'GET /ready HTTP/1.1\\r\\nHost: localhost\\r\\n\\r\\n' >&3; timeout 2 cat <&3) 3<>/dev/tcp/localhost/19000 | grep -q LIVE"]
+      interval: 10s
+      timeout: 5s
+      retries: 5
+      start_period: 10s
+
+  # LLM Katan service - lightweight mock LLM for testing
+  llm-katan:
+    image: ghcr.io/vllm-project/semantic-router/llm-katan:latest
+    container_name: llm-katan
+    ports:
+      - "8002:8002"
+    environment:
+      - HF_HUB_ENABLE_HF_TRANSFER=1
+    volumes:
+      - ../../models:/app/models:ro,z
+      - hf-cache:/home/llmkatan/.cache/huggingface
+    networks:
+      semantic-network:
+        ipv4_address: 172.28.0.20
+    command: ["llm-katan", "--model", "/app/models/Qwen/Qwen3-0.6B", "--served-model-name", "qwen3", "--host", "0.0.0.0", "--port", "8002"]
+    healthcheck:
+      test: ["CMD", "curl", "-fsS", "http://localhost:8002/health"]
+      interval: 10s
+      timeout: 5s
+      retries: 5
+      start_period: 10s
+
+networks:
+  semantic-network:
+    driver: bridge
+    ipam:
+      config:
+        - subnet: 172.28.0.0/16
+
+volumes:
+  hf-cache:
+