Skip to content

QUICKSTART_GUI.md

Latest commit

 

History

History
170 lines (120 loc) · 3.41 KB

QUICKSTART_GUI.md

File metadata and controls

170 lines (120 loc) · 3.41 KB

🚀 Quick Start - T5 Training GUI

⚡ 3-Step Setup

Step 1: Install Frontend Dependency

Open PowerShell and run:

cd "c:\Users\suhas\Documents\GitHub\OntoBot\rasa-frontend"
npm install

This will install the new react-select package needed for the sensor dropdown.

Step 2: Start Backend (Microservices)

Open a new PowerShell terminal:

cd "c:\Users\suhas\Documents\GitHub\OntoBot\microservices"
python app.py

Expected Output:

 * Serving Flask app 'app'
 * Debug mode: on
 * Running on http://0.0.0.0:6000

Keep this terminal open.

Step 3: Start Frontend

Open another PowerShell terminal:

cd "c:\Users\suhas\Documents\GitHub\OntoBot\rasa-frontend"
npm start

Expected Output:

Compiled successfully!
The app is running at: http://localhost:3000

Browser should automatically open.

🎯 Access the GUI

  1. Browser opens automatically to http://localhost:3000
  2. Click "Settings" in the top navigation
  3. Click "T5 Model Training" tab (the 5th tab)
  4. You're ready! 🎉

📝 Add Your First Training Example

Example: Multi-Sensor Correlation

1. Question:

What is the correlation between temperature and humidity in zone 5.04?

2. Sensors Involved:

  • Start typing "Zone_Air_Temperature" and select Zone_Air_Temperature_Sensor_5.04
  • Start typing "Zone_Air_Humidity" and select Zone_Air_Humidity_Sensor_5.04

3. SPARQL Query:

SELECT ?sensor ?timeseriesId ?storedAt 
WHERE {
  VALUES ?sensor { 
    bldg:Zone_Air_Humidity_Sensor_5.04 
    bldg:Zone_Air_Temperature_Sensor_5.04 
  }
  ?sensor brick:hasLocation ?location .
  ?sensor brick:timeseries ?timeseries .
  ?timeseries ref:hasTimeseriesId ?timeseriesId .
  ?timeseries ref:storedAt ?storedAt .
}

4. Category: Multi-Sensor Correlation

5. Notes: Zone 5.04 temperature-humidity correlation

6. Click "Add Example"

🏋️ Train Your Model

  1. Set Epochs: 10 (recommended for quick testing)
  2. Click "Start Training"
  3. Confirm the dialog
  4. Wait ~5-10 minutes - watch the progress bar!
  5. Monitor logs - they update in real-time
  6. When complete, click "Deploy Model to Production"

🔄 Activate the New Model

  1. Go to "Action Server" tab
  2. Click "Restart Action Server"
  3. Wait for "Restart completed successfully"
  4. Done! Test your queries in the chatbot

✅ Verify It Works

Test in your chatbot:

What is the correlation between temperature and humidity in zone 5.04?

The model should now generate the correct SPARQL query with the VALUES clause!

🆘 Troubleshooting

Backend won't start

cd microservices
pip install Flask Flask-Cors
python app.py

Frontend won't start

cd rasa-frontend
npm install
npm start

Can't see sensor dropdown

  • Check backend is running (port 6000)
  • Verify rasa-bldg1/actions/sensor_list.txt exists
  • Refresh the browser page

Training fails

  • Check at least 1 example exists
  • Verify Python environment has dependencies
  • Check training logs for specific errors

📚 Need More Help?

Check these guides:

  • T5_GUI_SETUP.md - Detailed setup guide
  • GUI_TRAINING_GUIDE.md - Complete usage guide
  • T5_GUI_IMPLEMENTATION.md - Technical details

🎉 That's It!

You can now train your T5 model using the GUI!

No more command-line needed! 🚀