test_payloads/README.md

Test Payloads for TDS-PROJ-1

This directory contains JSON payload files to test your TDS-PROJ-1 application with various scenarios.

📁 Available Test Payloads

1. payload-sales-summary.json

Task: Sum of Sales Calculator

• Tests CSV file handling from attachments
• Validates fetch() usage
• Checks dynamic calculation and display
• Attachment: sales.csv with product sales data

2. payload-calculator.json

Task: Basic Calculator

• Tests interactive UI creation
• Validates multiple button elements
• Checks for display element and math operations
• Attachment: None

3. payload-todo-list.json

Task: To-Do List Application

• Tests form input handling
• Validates Bootstrap integration
• Checks dynamic list manipulation
• Attachment: None

4. payload-weather-dashboard.json

Task: Weather Dashboard

• Tests search functionality
• Validates display areas
• Checks Bootstrap styling
• Attachment: None

5. payload-json-viewer.json

Task: JSON Data Viewer

• Tests JSON file loading from attachments
• Validates tree view rendering
• Checks search/filter functionality
• Attachment: data.json with sample user data

6. payload-github-user.json

Task: GitHub User Lookup

• Tests external API integration (GitHub API)
• Validates form handling
• Checks date formatting
• Attachment: None

7. payload-markdown-enhanced.json

Task: Enhanced Markdown Editor (Round 2)

• Tests round 2 functionality
• Validates tab switching UI
• Checks word count feature
• Attachment: input.md with sample markdown

🚀 How to Use

Option 1: Run All Tests (Automated)

# Make sure your server is running first:
# Terminal 1:
python -m uvicorn main:app --host 0.0.0.0 --port 8000

# Terminal 2:
./run_tests.sh

Option 2: Test Individual Payloads

# Test a specific payload
curl -X POST http://127.0.0.1:8000/build \
  -H "Content-Type: application/json" \
  -d @test_payloads/payload-sales-summary.json

# With pretty output
curl -X POST http://127.0.0.1:8000/build \
  -H "Content-Type: application/json" \
  -d @test_payloads/payload-calculator.json | jq '.'

Option 3: Test with Python

# Using the test_live_api.py script
python test_live_api.py

# Or use Python requests directly
python3 -c "
import requests
import json

with open('test_payloads/payload-todo-list.json') as f:
    payload = json.load(f)

response = requests.post(
    'http://127.0.0.1:8000/build',
    json=payload
)

print(f'Status: {response.status_code}')
print(f'Response: {response.json()}')
"

📊 Understanding Test Results

✅ Success Response (HTTP 200)

{
  "message": "Request received. Buildling Application..."
}

❌ Unauthorized (HTTP 401)

{
  "message": "Unauthorized: Invalid secret key."
}

Fix: Check that your .env file has API_SECRET=11032003

⚠️ Validation Error (HTTP 422)

{
  "detail": [...]
}

Fix: Check the payload structure matches the Pydantic model

🔍 What Happens After Testing

For each successful test request:

1. ✅ Request Accepted - Returns HTTP 200
2. 🤖 LLM Processing - Generates app based on brief
3. 📦 GitHub Repo Created - Creates repository with task name
4. 📝 Code Pushed - Pushes README.md, LICENSE, index.html
5. 🌐 Pages Enabled - Activates GitHub Pages
6. 📤 Evaluation Sent - Posts results to evaluation_url

📁 Created Repositories

After running tests, check your GitHub account for repositories named:

• sum-of-sales-test-001
• calculator-app-test-002
• todo-list-test-003
• weather-dashboard-test-004
• json-viewer-test-005
• github-user-lookup-test-006
• markdown-editor-test-007

Each will be live at: https://your-username.github.io/repo-name/

🛠️ Creating Your Own Test Payloads

Template structure:

{
  "email": "your-email@example.com",
  "secret": "11032003",
  "task": "unique-task-id",
  "round": 1,
  "nonce": "unique-nonce",
  "brief": "Description of what app should do...",
  "checks": ["JavaScript validation check 1", "JavaScript validation check 2"],
  "evaluation_url": "http://127.0.0.1:8000/_eval",
  "attachments": [
    {
      "name": "filename.ext",
      "url": "data:mime/type;base64,BASE64_ENCODED_CONTENT"
    }
  ]
}

Encoding Attachments

# Encode a file to base64
base64 -w 0 myfile.csv

# Or with data URI prefix
echo "data:text/csv;base64,$(base64 -w 0 myfile.csv)"

💡 Tips

1. Test Locally First - Always test on localhost:8000 before deploying
2. Check Server Logs - Monitor the uvicorn terminal for detailed processing info
3. One at a Time - Run tests one by one to avoid GitHub API rate limits
4. Clean Up - Delete test repositories from GitHub after testing
5. Secret Match - Ensure .env has API_SECRET=11032003

🐛 Troubleshooting

Server not responding?

# Check if server is running
curl http://127.0.0.1:8000/

# Restart server
python -m uvicorn main:app --host 0.0.0.0 --port 8000

401 Unauthorized?

# Check environment variable
cat .env | grep API_SECRET

# Should show: API_SECRET=11032003

JSON parsing errors?

• Check recent fixes to llm.py are applied
• Verify prompts in app/services/prompts/ are updated
• Fall back to mock service if aipipe fails

📚 Related Files

• /test_live_api.py - Tests live Hugging Face deployment
• /test_functionality.py - Runs comprehensive functionality tests
• /run_tests.sh - Automated test runner for all payloads
• /payload.markdown-to-html.json - Original example payload

Happy Testing! 🚀