🛰️ smart_gui

smart_gui is a ROS 2 (ament_python) package focused on monitoring and testing ROS systems through a web interface. It combines a FastAPI backend and a Flutter frontend to inspect the ROS graph, stream live topic data, call services, and publish messages (including backend-managed loop publishing) without requiring custom debug scripts.

✨ Features

• Graph inspection:
  • GET /topics
  • GET /nodes
  • GET /services
• Service tooling:
  • GET /service-schema
  • POST /service-call
• Topic tooling:
  • Create a publisher for a topic/type.
  • Publish one message.
  • Start/stop backend publish loops (no per-message HTTP traffic from frontend).
• Live topic monitor via WebSocket:
  • WS /ws/topics/{topic}
• Special handling for sensor_msgs/msg/Image:
  • Backend JPEG compression to reduce payload size.

🗂️ Repository Layout

• smart_gui/ros2_inspector_api.py: main ROS2 + FastAPI backend node.
• frontend/: Flutter web application.
• launch/smart_gui_api.launch.py: launch backend and optional frontend process.
• smart_gui/random_int8_topics_node.py: test node that publishes random std_msgs/msg/Int8 on random topics.

📦 Requirements

• ROS 2 Jazzy (or compatible environment with rclpy).
• colcon.
• Python 3.
• Flutter (only if you want to run the web frontend with flutter run).

🛠️ Build

From your ROS 2 workspace root:

cd ~/colcon_ws
source /opt/ros/jazzy/setup.bash
colcon build --packages-select smart_gui --symlink-install
source install/setup.bash

▶️ Run

1) Backend + frontend together (recommended for development)

ros2 launch smart_gui smart_gui_api.launch.py \
  frontend_command:='NO_PROXY=localhost,127.0.0.1 no_proxy=localhost,127.0.0.1 flutter run -d web-server --web-hostname 0.0.0.0 --web-port 3000 --profile'

Open:

• Local machine: http://localhost:3000
• Another device in LAN: http://<YOUR_MACHINE_IP>:3000

2) Backend only

ros2 launch smart_gui smart_gui_api.launch.py run_frontend:=false

3) Backend only (direct run)

ros2 run smart_gui smart_gui_api --host 0.0.0.0 --port 8000

🔌 API Endpoints

Basic

• GET /health
• GET /topics
• GET /nodes
• GET /services

Topic message types/templates

• GET /topic-message-types
• GET /topic-message-template?message_type=<pkg/msg/Type>

Topic publish control

• POST /topic-publisher
• POST /topic-publish
• POST /topic-publish-loop/start
• POST /topic-publish-loop/stop

Services

• GET /service-schema?name=<service>&service_type=<pkg/srv/Type>
• POST /service-call

WebSocket

• WS /ws/topics/{topic}

🔁 Backend-Managed Topic Loop Publishing

Loop publishing is handled by the backend (thread per topic/type loop), so the frontend does not send one HTTP request per message.

Example:

curl -X POST http://127.0.0.1:8000/topic-publish-loop/start \
  -H 'Content-Type: application/json' \
  -d '{
    "name": "/demo_int",
    "message_type": "std_msgs/msg/Int32",
    "message": {"data": 10},
    "frequency_hz": 5.0
  }'

Stop loop:

curl -X POST http://127.0.0.1:8000/topic-publish-loop/stop \
  -H 'Content-Type: application/json' \
  -d '{
    "name": "/demo_int",
    "message_type": "std_msgs/msg/Int32"
  }'

🧪 Random Test Publisher Node

Run a helper node that creates random topic names and publishes random Int8 values:

ros2 run smart_gui random_int8_topics

Optional arguments:

ros2 run smart_gui random_int8_topics --topic-count 5 --hz 5.0

📱 Notes for Mobile/LAN Access

• The backend usually runs on port 8000, frontend on 3000.
• If running inside WSL2, make sure Windows forwards ports (3000/8000) to WSL and firewall rules allow inbound connections.
• Access from phone should use Windows LAN IP, not WSL internal IP.

🧭 Troubleshooting

• Backend health check:

curl http://127.0.0.1:8000/health

• Confirm loop endpoints are available:

curl -s http://127.0.0.1:8000/openapi.json | rg 'topic-publish-loop/start|topic-publish-loop/stop'

• If frontend looks stale on mobile browser, force refresh or open in private/incognito window.