#!/usr/bin/env python3
"""
Manual validation of Batch 12 library matches.

Reviews 4 fuzzy matches from batch12_library_query_results.json
Validates location accuracy and eliminates false positives.
"""

import json
from pathlib import Path

def validate_batch12():
    """Manual validation with reasoning."""
    
    input_file = Path("scripts/batch12_library_query_results.json")
    
    print("=" * 80)
    print("BATCH 12 LIBRARY MATCHES - MANUAL VALIDATION")
    print("=" * 80)
    print()
    
    # Load query results
    with open(input_file, 'r', encoding='utf-8') as f:
        data = json.load(f)
    
    matches = data['matches']
    
    print(f"Reviewing {len(matches)} matches from automated query:")
    print()
    
    # Manual validation
    validated = []
    rejected = []
    
    # Match 1: Biblioteca Pública Federico Varela (Chañaral) → Q135435755 (La Reina)
    print("1. Biblioteca Pública Federico Varela (Chañaral)")
    print("   → Q135435755: Biblioteca Pública de La Reina (La Reina)")
    print("   ❌ REJECT: Location mismatch")
    print("      Reason: Chañaral (Atacama Region) ≠ La Reina (Santiago Metro)")
    print("      Generic 'Biblioteca Pública' name caused false match")
    print()
    rejected.append({
        "library_name": "Biblioteca Pública Federico Varela",
        "city": "Chañaral",
        "q_number": "Q135435755",
        "wikidata_name": "Biblioteca Pública de La Reina",
        "reason": "Location mismatch: Chañaral (Atacama) ≠ La Reina (Santiago)"
    })
    
    # Match 2: Biblioteca Pública de Illapel (Los Vilos) → Q135435755 (La Reina)
    print("2. Biblioteca Pública de Illapel (Los Vilos)")
    print("   → Q135435755: Biblioteca Pública de La Reina (La Reina)")
    print("   ❌ REJECT: Same false positive")
    print("      Reason: Illapel/Los Vilos (Coquimbo Region) ≠ La Reina (Santiago)")
    print("      Generic 'Biblioteca Pública' name caused false match")
    print()
    rejected.append({
        "library_name": "Biblioteca Pública de Illapel",
        "city": "Los Vilos",
        "q_number": "Q135435755",
        "wikidata_name": "Biblioteca Pública de La Reina",
        "reason": "Location mismatch: Los Vilos (Coquimbo) ≠ La Reina (Santiago)"
    })
    
    # Match 3: Biblioteca Pública N°56 (Talagante) → Q135435755 (La Reina)
    print("3. Biblioteca Pública N°56 (Talagante)")
    print("   → Q135435755: Biblioteca Pública de La Reina (La Reina)")
    print("   ❌ REJECT: Same false positive")
    print("      Reason: Talagante ≠ La Reina (both Santiago Metro, but different communes)")
    print("      Generic 'Biblioteca Pública' name caused false match")
    print()
    rejected.append({
        "library_name": "Biblioteca Pública N°56",
        "city": "Talagante",
        "q_number": "Q135435755",
        "wikidata_name": "Biblioteca Pública de La Reina",
        "reason": "Location mismatch: Talagante ≠ La Reina"
    })
    
    # Match 4: Biblioteca Pública N° 244 (Pichilemu) → Q134891536 (Lebu)
    print("4. Biblioteca Pública N° 244 (Pichilemu)")
    print("   → Q134891536: Biblioteca Pública Municipal de Lebu (Lebu)")
    print("   ❌ REJECT: Location mismatch")
    print("      Reason: Pichilemu (O'Higgins) ≠ Lebu (Bío Bío)")
    print("      Generic 'Biblioteca Pública' name caused false match")
    print()
    rejected.append({
        "library_name": "Biblioteca Pública N° 244",
        "city": "Pichilemu",
        "q_number": "Q134891536",
        "wikidata_name": "Biblioteca Pública Municipal de Lebu",
        "reason": "Location mismatch: Pichilemu (O'Higgins) ≠ Lebu (Bío Bío)"
    })
    
    # Summary
    print("=" * 80)
    print("VALIDATION SUMMARY")
    print("=" * 80)
    print(f"Validated: {len(validated)}")
    print(f"Rejected: {len(rejected)}")
    print()
    
    # Analysis
    print("KEY FINDING:")
    print("  All 4 matches are FALSE POSITIVES caused by generic 'Biblioteca Pública' names")
    print("  Q135435755 matched 3 different libraries due to fuzzy name similarity")
    print("  Location information was ignored or insufficient for validation")
    print()
    
    print("ISSUE:")
    print("  Chilean public libraries often have generic names:")
    print("    - Biblioteca Pública N°56")
    print("    - Biblioteca Pública N° 244")
    print("    - Biblioteca Pública [City Name]")
    print("  These create high false positive rates in fuzzy matching")
    print()
    
    print("RECOMMENDATION:")
    print("  ❌ DO NOT apply any Batch 12 library matches")
    print("  🔍 Alternative strategy needed:")
    print("     1. Manual Wikidata search for specific named libraries")
    print("     2. Focus on well-documented museums instead")
    print("     3. Create Wikidata entries for missing libraries")
    print()
    
    # Save validation results
    output = {
        "batch": 12,
        "validation_date": "2025-11-09",
        "institution_type": "LIBRARY",
        "validated_matches": validated,
        "rejected_matches": rejected,
        "summary": {
            "validated": len(validated),
            "rejected": len(rejected),
            "false_positive_rate": f"{len(rejected)/len(matches)*100:.1f}%",
            "recommendation": "REJECT ALL - Generic library names cause false positives"
        }
    }
    
    output_file = Path("scripts/batch12_final_validation.json")
    with open(output_file, 'w', encoding='utf-8') as f:
        json.dump(output, f, indent=2, ensure_ascii=False)
    
    print(f"✅ Validation results saved to: {output_file}")

if __name__ == "__main__":
    validate_batch12()