TESTING.md

Testing Guide for Duplicate Detection API

Quick Test Methods

1. Using the Test Script (Recommended)

Test the Hugging Face Space deployment:

python3 test.py --base-url https://LogicGoInfotechSpaces-duplicate-transaction-detection.hf.space

Test locally (if running on your machine):

python3 test.py --base-url http://127.0.0.1:8000

2. Using cURL Commands

Health Check:

curl https://LogicGoInfotechSpaces-duplicate-transaction-detection.hf.space/health

Get Suggestions:

curl https://LogicGoInfotechSpaces-duplicate-transaction-detection.hf.space/suggestions?limit=10

With pretty JSON output:

curl -s https://LogicGoInfotechSpaces-duplicate-transaction-detection.hf.space/suggestions | python3 -m json.tool

3. Using Browser

Simply open these URLs in your browser:

• Health Check:

  https://LogicGoInfotechSpaces-duplicate-transaction-detection.hf.space/health
  

• Get Suggestions:

  https://LogicGoInfotechSpaces-duplicate-transaction-detection.hf.space/suggestions?limit=5
  

4. Using Python Requests

Create a simple test script:

import requests
import json

BASE_URL = "https://LogicGoInfotechSpaces-duplicate-transaction-detection.hf.space"

# Test health endpoint
response = requests.get(f"{BASE_URL}/health")
print("Health Status:", response.status_code)
print("Response:", response.json())

# Test suggestions endpoint
response = requests.get(f"{BASE_URL}/suggestions", params={"limit": 5})
print("\nSuggestions Status:", response.status_code)
print("Suggestions Count:", len(response.json()))
print("\nFirst Suggestion:")
print(json.dumps(response.json()[0] if response.json() else {}, indent=2))

5. Using Postman or Insomnia

Health Endpoint:

• Method: GET
• URL: https://LogicGoInfotechSpaces-duplicate-transaction-detection.hf.space/health

Suggestions Endpoint:

• Method: GET
• URL: https://LogicGoInfotechSpaces-duplicate-transaction-detection.hf.space/suggestions
• Query Parameters:
  • limit: 10 (optional, default: 50, max: 500)

Expected Responses

Health Endpoint Response:

{
  "status": "ok"
}

Suggestions Endpoint Response:

[
  {
    "_id": "6923ec37a48c1900950d7e7a",
    "candidate_ids": ["6923ebe7dfb90e344a8a8289", "6923ebf3dfb90e344a8a82cf"],
    "message": "These seem similar. Would you like to merge them?",
    "details": {
      "amount_delta_pct": 0.16,
      "time_delta_minutes": 0.0,
      "merchant_match_rule": "exact"
    },
    "audit": {
      "generated_by": "duplicate-detector",
      "generated_at": "2025-11-24T05:25:11.966000",
      "rule_version": "v1.0"
    },
    "status": "pending"
  }
]

Testing the Scheduler

The scheduler runs automatically in the background. To verify it's working:

1. Check the logs - The scheduler logs when it runs duplicate detection
2. Monitor suggestions - New suggestions should appear periodically (default: every 60 seconds)
3. Check timestamps - The audit.generated_at field shows when each suggestion was created

Local Testing

To test locally, first start the server:

python3 -m uvicorn src.api:app --host 127.0.0.1 --port 8000 --reload

Then in another terminal:

python3 test.py --base-url http://127.0.0.1:8000

Troubleshooting

• 404 Error: Make sure the Space is running on Hugging Face
• Timeout: The suggestions endpoint might take time if there's a lot of data
• Empty suggestions: The scheduler might not have found duplicates yet, or the lookback window needs adjustment