[Smart Traffic Agent] RI Dependencies and Deployment setup (open-edge-platform#1537)

Signed-off-by: Krishna Murti <krishna.murti@intel.com>
Co-authored-by: Madhuri Kumari <madhuri.rai07@gmail.com>

Author: krish918

.gitmodules

@@ -37,3 +37,8 @@
 [submodule "robotics-ai-suite/pipelines/mpc-demo/ocs2/ocs2"]
 	path = robotics-ai-suite/pipelines/mpc-demo/ocs2/ocs2
 	url = https://github.com/leggedrobotics/ocs2.git
+[submodule "metro-ai-suite/smart-traffic-intersection-agent/deps/metro-vision"]
+	path = metro-ai-suite/smart-traffic-intersection-agent/deps/metro-vision
+	url = https://github.com/open-edge-platform/edge-ai-suites.git
+	branch = exp/smart-traffic-adapter
+  	shallow = true

metro-ai-suite/smart-traffic-intersection-agent/.gitignore

@@ -86,4 +86,4 @@ share/python-wheels/
 *.egg
 MANIFEST
 
-**/edge-ai-suites/
+**/docker/ri-compose.yaml

metro-ai-suite/smart-traffic-intersection-agent/deps/metro-vision

@@ -1,25 +1,22 @@
 # Copyright (C) 2025 Intel Corporation
 # SPDX-License-Identifier: Apache-2.0
 
-name: scene-intelligence
-
 secrets:
   root-cert:
-    file: ../edge-ai-suites/metro-ai-suite/metro-vision-ai-app-recipe/smart-intersection/src/secrets/certs/scenescape-ca.pem
+    file: ${RI_DIR}/src/secrets/certs/scenescape-ca.pem
 
 services:
-  # VLM OpenVINO Serving Service
   vlm-openvino-serving:
-    image: intel/vlm-openvino-serving:1.3.1
+    image: intel/vlm-openvino-serving:1.3.2
     ports:
-      - "${VLM_SERVICE_PORT:-9764}:8000"
+      - "8000"
     ipc: host
     networks:
       - scenescape
     environment:
       # Proxy settings
-      no_proxy_env: ${no_proxy_env},${HOST_IP}
-      no_proxy: ${no_proxy_env},${HOST_IP}
+      no_proxy_env: ${no_proxy:-${NO_PROXY:-}},${HOST_IP}
+      no_proxy: ${no_proxy:-${NO_PROXY:-}},${HOST_IP}
       http_proxy: ${http_proxy}
       https_proxy: ${https_proxy}
       # VLM Model configuration
@@ -30,7 +27,7 @@ services:
       WORKERS: ${VLM_WORKERS:-1}
       VLM_MAX_COMPLETION_TOKENS: ${VLM_MAX_COMPLETION_TOKENS:-1500}
       HUGGINGFACE_TOKEN: ${HUGGINGFACE_TOKEN}
-      OV_CONFIG: ${OV_CONFIG}
+      OV_CONFIG: ${OV_CONFIG:-}
       VLM_LOG_LEVEL: ${VLM_LOG_LEVEL:-info}
       OPENVINO_LOG_LEVEL: ${VLM_OPENVINO_LOG_LEVEL:-1}
       VLM_ACCESS_LOG_FILE: ${VLM_ACCESS_LOG_FILE:-/dev/null}
@@ -51,62 +48,59 @@ services:
       - ov-models:/root/.cache/huggingface
       - ov-models:/app/ov-model
 
-  # Smart Traffic Intersection Agent Service
-  traffic-intersection-agent:
+  traffic-agent:
     build:
-      context: ../src
+      context: ${APP_DIR:-..}/src
       dockerfile: Dockerfile
-    image: ${REGISTRY:-}traffic-intersection-agent:${TAG:-latest}
+    image: ${REGISTRY:-}smart-traffic-intersection-agent:${TAG:-latest}
     ports:
-      - "${TRAFFIC_INTERSECTION_AGENT_PORT:-8081}:8081"
-      - "${TRAFFIC_INTERSECTION_AGENT_UI_PORT:-7860}:7860"
+      - "${AGENT_BACKEND_PORT:-}:8081"
+      - "${AGENT_UI_PORT:-}:7860"
     networks:
       - scenescape
     depends_on:
       vlm-openvino-serving:
         condition: service_healthy
     environment:
-      # Service configuration
-      TRAFFIC_INTERSECTION_AGENT_HOST: 0.0.0.0
-      TRAFFIC_INTERSECTION_AGENT_PORT: ${TRAFFIC_INTERSECTION_AGENT_PORT:-8081}
       LOG_LEVEL: ${LOG_LEVEL:-INFO}
-      
-      # UI configuration
-      TRAFFIC_INTERSECTION_AGENT_UI_PORT: ${TRAFFIC_INTERSECTION_AGENT_UI_PORT:-7860}
       USE_API: "true"
-      API_URL: http://localhost:${TRAFFIC_INTERSECTION_AGENT_PORT:-8081}/api/v1/traffic/current
       REFRESH_INTERVAL: ${REFRESH_INTERVAL:-15}
 
       # Proxy settings
       http_proxy: ${http_proxy}
-      https_proxy: ${https_proxy}
-      no_proxy: ${no_proxy_env},.scenescape.intel.com,${HOST_IP},broker.scenescape.intel.com,vlm-openvino-serving
-      NO_PROXY: ${no_proxy_env},.scenescape.intel.com,${HOST_IP},broker.scenescape.intel.com,vlm-openvino-serving
       HTTP_PROXY: ${http_proxy}
+      https_proxy: ${https_proxy}
       HTTPS_PROXY: ${https_proxy}
+      no_proxy: ${no_proxy:-${NO_PROXY:-}},.scenescape.intel.com,${HOST_IP},broker.scenescape.intel.com,vlm-openvino-serving
+      NO_PROXY: ${no_proxy:-${NO_PROXY:-}},.scenescape.intel.com,${HOST_IP},broker.scenescape.intel.com,vlm-openvino-serving
 
       # MQTT configuration (connects to broker service)
       MQTT_HOST: ${MQTT_HOST:-broker.scenescape.intel.com}
       MQTT_PORT: ${MQTT_PORT:-1883}
 
-      # Intersection configuration (overrides config/traffic_agent.json)
-      INTERSECTION_NAME: ${INTERSECTION_NAME:-Intersection-1}
-      INTERSECTION_LATITUDE: ${INTERSECTION_LATITUDE:-37.7049108}
-      INTERSECTION_LONGITUDE: ${INTERSECTION_LONGITUDE:--121.9096158}
+      # Intersection configuration (overrides src/config/deployment_instance.json)
+      INTERSECTION_NAME: ${INTERSECTION_NAME:-}
+      INTERSECTION_LATITUDE: ${INTERSECTION_LATITUDE:-}
+      INTERSECTION_LONGITUDE: ${INTERSECTION_LONGITUDE:-}
 
-      # Weather configuration (overrides config/traffic_agent.json)
+      # Weather configuration (overrides src/config/traffic_agent.json)
       WEATHER_MOCK: ${WEATHER_MOCK:-false}
 
-      # VLM configuration (overrides config/traffic_agent.json)
-      VLM_MODEL: ${VLM_MODEL_NAME:-Qwen/Qwen2.5-VL-3B-Instruct}
+      # VLM configuration (overrides src/config/traffic_agent.json)
+      VLM_MODEL_NAME: ${VLM_MODEL_NAME:-Qwen/Qwen2.5-VL-3B-Instruct}
+      VLM_TIMEOUT_SECONDS: ${VLM_TIMEOUT_SECONDS:-}
+      VLM_MAX_COMPLETION_TOKENS: ${VLM_MAX_COMPLETION_TOKENS:-}
+      VLM_TEMPERATURE: ${VLM_TEMPERATURE:-}
+      VLM_TOP_P: ${VLM_TOP_P:-}
 
-      # Traffic analysis configuration (overrides config/traffic_agent.json)
+      # Traffic analysis configuration (overrides src/config/traffic_agent.json)
       HIGH_DENSITY_THRESHOLD: ${HIGH_DENSITY_THRESHOLD:-10}
-      MODERATE_DENSITY_THRESHOLD: ${MODERATE_DENSITY_THRESHOLD:-5}
+      MODERATE_DENSITY_THRESHOLD: ${MODERATE_DENSITY_THRESHOLD:-}
+      TRAFFIC_BUFFER_DURATION: ${TRAFFIC_BUFFER_DURATION:-}
 
     volumes:
-      - ../src/config:/app/config:ro
-      - traffic-intelligence-data:/app/data
+      - ${APP_DIR:-..}/src/config:/app/config:ro
+      - traffic-agent-data:/app/data
     secrets:
       - source: root-cert
         target: /app/secrets/certs/scenescape-ca.pem
@@ -120,9 +114,7 @@ services:
 
 volumes:
   ov-models:
-  traffic-intelligence-data:
+  traffic-agent-data:
 
 networks:
   scenescape:
-    external: true
-    name: metro-vision-ai-app-recipe_scenescape

…c-intersection-agent/docker/compose.yaml …rsection-agent/docker/agent-compose.yamlmetro-ai-suite/smart-traffic-intersection-agent/docker/compose.yaml renamed to metro-ai-suite/smart-traffic-intersection-agent/docker/agent-compose.yaml metro-ai-suite/smart-traffic-intersection-agent/docker/compose.yaml renamed to metro-ai-suite/smart-traffic-intersection-agent/docker/agent-compose.yaml

@@ -1,15 +1,46 @@
-# Scene Intelligence microservice
+# Smart Traffic Intersection Agent
 
-The Scene Intelligence microservice is a comprehensive traffic analysis service that provides real-time intersection monitoring, directional traffic density analysis, and Vision Language Model (VLM) powered traffic insights. It processes MQTT traffic data, manages camera images, and delivers intelligent traffic analysis through RESTful APIs.
+The Smart Traffic Intersection Agent is a comprehensive traffic analysis service that provides real-time intersection monitoring, directional traffic density analysis, and Vision Language Model (VLM) powered traffic insights. It processes MQTT traffic data, manages camera images, and delivers intelligent traffic analysis through RESTful APIs.
 
 ## Overview
 
 The microservice processes real-time traffic data from MQTT streams and provides advanced analytics including directional traffic density monitoring, VLM-powered traffic scene analysis, and comprehensive traffic summaries. It supports sliding window analysis, sustained traffic detection, and intelligent camera image management for enhanced traffic insights.
 
+## How It Works
+
+The diagram below illustrates the high-level architecture of the Smart Traffic Intelligence Agent, showcasing its core components and their interactions with external systems.
+
 <p align="center">
     <img src="./_images/ITT_architecture.png" alt="Architecture" />
 </p>
 
+The Smart Traffic UI below shows how traffic, weather data is analyzed and summary, alerts are shown to the user.
+
+<p align="center">
+    <img src="./_images/traffic_agent_ui.png" alt="Traffic Intersection Agent UI" />
+</p>
+
+## Components
+
+The Smart Traffic Intelligence stack includes these containerized services:
+
+- **MQTT Broker** (Eclipse Mosquitto) - Message broker for traffic data
+- **DL Streamer Pipeline Server** - Video analytics and AI inference
+- **SceneScape Database** - Configuration and metadata storage
+- **SceneScape Web Server** - Management interface
+- **SceneScape Controller** - Orchestration service
+- **VLM OpenVINO Serving** - Vision Language Model inference
+- **Traffic Intelligence** - Real-time traffic analysis with dual interface (API + UI)
+
+### Key Integration Points
+
+- **MQTT Communication**: All services communicate via the shared MQTT broker
+- **Docker Network**: Services discover each other via Docker service names
+- **Shared Secrets**: TLS certificates and auth files mounted from `src/secrets/`
+- **Persistent Storage**: Traffic data stored in Docker volume `traffic-intelligence-data`
+- **Health Monitoring**: All services include health check endpoints
+
+
 ## Supporting Resources
 
 - [Get Started Guide](get-started.md)