test_kmeans_format.py

#!/usr/bin/env python3
"""
Quick test to verify K-Means returns proper format
"""
import sys
sys.path.insert(0, '/home/deepu/Downloads/code2')

from risk_discovery import UnsupervisedRiskDiscovery

# Sample clauses
sample_clauses = [
    "The Company shall not be liable for any indirect damages.",
    "Either party may terminate this Agreement upon 30 days notice.",
    "All intellectual property rights remain with the Company.",
    "Customer agrees to pay monthly fees of $10,000.",
    "Each party shall keep confidential all information disclosed.",
    "The Service Provider agrees to indemnify the Client.",
    "This Agreement shall automatically terminate upon bankruptcy.",
    "Vendor shall maintain compliance with ISO 27001 standards.",
]

print("="*80)
print("Testing K-Means Risk Discovery Return Format")
print("="*80)

# Initialize and run discovery
discovery = UnsupervisedRiskDiscovery(n_clusters=3, random_state=42)
results = discovery.discover_risk_patterns(sample_clauses)

# Check result structure
print("\n" + "="*80)
print("Checking Result Structure")
print("="*80)

required_keys = ['method', 'n_clusters', 'discovered_patterns', 'cluster_labels', 'quality_metrics']
for key in required_keys:
    if key in results:
        print(f"✅ '{key}' present")
        if key == 'method':
            print(f"   Value: {results[key]}")
        elif key == 'n_clusters':
            print(f"   Value: {results[key]}")
        elif key == 'quality_metrics':
            print(f"   Metrics: {list(results[key].keys())}")
    else:
        print(f"❌ '{key}' MISSING")

print("\n" + "="*80)
print("Discovered Patterns:")
print("="*80)
for pattern_name, details in results['discovered_patterns'].items():
    print(f"\n{pattern_name}:")
    print(f"  Clauses: {details['clause_count']}")
    print(f"  Key terms: {', '.join(details['key_terms'][:5])}")

print("\n" + "="*80)
if all(key in results for key in required_keys):
    print("✅ TEST PASSED: All required keys present")
else:
    print("❌ TEST FAILED: Missing required keys")
print("="*80)