#!/usr/bin/env python3
"""
Chilean Heritage Institutions - Batch 18 Enrichment
Push to 75% coverage by enriching 2 libraries with municipal parent organizations

Target: 66/90 → 68/90 (75.6% coverage)

Enrichment Strategy:
- Add Q-numbers for parent municipalities (Coquimbo, Pichilemu)
- Both are regional/tourist cities with well-documented communes in Wikidata
- Conservative approach: only add when municipality clearly documented

Date: 2025-11-09
"""

import yaml
from pathlib import Path

# Input/output paths
INPUT_FILE = Path(__file__).parent.parent / "data/instances/chile/chilean_institutions_batch17_enriched.yaml"
OUTPUT_FILE = Path(__file__).parent.parent / "data/instances/chile/chilean_institutions_batch18_enriched.yaml"

# Enrichment mappings: institution ID → (Q-number, enrichment type, reason)
ENRICHMENTS = {
    # Libraries with municipal parent organizations
    "https://w3id.org/heritage/custodian/cl/l-biblioteca-jorge-iribarren-cha-0018": {
        "q_number": "Q23660214",
        "type": "parent_organization",
        "reason": "Parent organization - Municipal library in Coquimbo commune"
    },
    "https://w3id.org/heritage/custodian/cl/l-biblioteca-p-blica-n-244-0045": {
        "q_number": "Q23660186",
        "type": "parent_organization",
        "reason": "Parent organization - Municipal library in Pichilemu commune"
    }
}

def enrich_institution(inst: dict) -> bool:
    """
    Enrich institution with Wikidata Q-number if applicable.
    Returns True if enriched, False otherwise.
    """
    inst_id = inst.get('id')
    
    if inst_id not in ENRICHMENTS:
        return False
    
    enrichment = ENRICHMENTS[inst_id]
    q_number = enrichment['q_number']
    enrich_type = enrichment['type']
    reason = enrichment['reason']
    
    # Check if already has this Wikidata identifier
    if inst.get('identifiers'):
        for ident in inst['identifiers']:
            if ident.get('identifier_scheme') == 'Wikidata' and ident.get('identifier_value') == q_number:
                print(f"⚠ Already enriched: {inst.get('name')} (Q{q_number})")
                return False
    
    # Add Wikidata identifier
    if not inst.get('identifiers'):
        inst['identifiers'] = []
    
    inst['identifiers'].append({
        'identifier_scheme': 'Wikidata',
        'identifier_value': q_number,
        'identifier_url': f'https://www.wikidata.org/wiki/{q_number}'
    })
    
    # Update provenance notes
    if not inst.get('provenance'):
        inst['provenance'] = {}
    
    existing_notes = inst['provenance'].get('notes', '')
    enrichment_note = f"Wikidata enrichment (Batch 18): {enrich_type} - {reason}"
    
    if existing_notes:
        inst['provenance']['notes'] = f"{existing_notes} | {enrichment_note}"
    else:
        inst['provenance']['notes'] = enrichment_note
    
    print(f"✓ Enriched: {inst.get('name')}")
    print(f"  Q-number: {q_number} ({enrich_type})")
    print(f"  Reason: {reason}")
    print()
    
    return True

def main():
    # Load institutions
    with open(INPUT_FILE, 'r', encoding='utf-8') as f:
        institutions = yaml.safe_load(f)
    
    # Enrich institutions
    enriched_count = 0
    for inst in institutions:
        if enrich_institution(inst):
            enriched_count += 1
    
    # Save enriched data
    with open(OUTPUT_FILE, 'w', encoding='utf-8') as f:
        yaml.dump(institutions, f, allow_unicode=True, sort_keys=False, width=120)
    
    print("=" * 60)
    print("Batch 18 Enrichment Complete")
    print("=" * 60)
    print(f"Institutions enriched: {enriched_count}")
    print(f"Output: {OUTPUT_FILE}")
    print()
    print("Next: Calculate coverage statistics to verify 75% target reached")

if __name__ == '__main__':
    main()