glam/scripts/add_alternative_names_libraries.py

#!/usr/bin/env python3
"""
Add alternative_names to library entries where Google Maps shows branch name
but original entry shows library system name.
"""
import yaml
import os
from datetime import datetime

# Define library entries to update
LIBRARY_UPDATES = {
    "1391_kb_isil.yaml": {
        "alternative_names": [
            {"name": "Bibliotheek Kop van Overijssel", "source": "original_entry", "note": "Library system name"},
            {"name": "Bibliotheek Steenwijk", "source": "google_maps", "note": "Branch location name"}
        ],
        "custodian_name_update": {
            "claim_value": "Bibliotheek Kop van Overijssel",
            "source": "original_entry",
            "note": "Library system serving Steenwijkerland, Westerveld municipalities. Steenwijk is main branch."
        }
    },
    "1416_kb_isil.yaml": {
        "alternative_names": [
            {"name": "NOBB", "source": "original_entry", "note": "Library system abbreviation"},
            {"name": "Bibliotheek Oss", "source": "google_maps", "note": "Main branch location"}
        ],
        "custodian_name_update": {
            "claim_value": "NOBB",
            "source": "original_entry", 
            "note": "Noord Oost Brabantse Bibliotheken - regional library system. Oss is one of multiple branches."
        }
    },
    "1456_kb_isil.yaml": {
        "alternative_names": [
            {"name": "Bibliotheken Zuidoost Fryslân", "source": "original_entry", "note": "Library system name"},
            {"name": "Bibliotheek Wolvega", "source": "google_maps", "note": "Branch location name"}
        ],
        "custodian_name_update": {
            "claim_value": "Bibliotheken Zuidoost Fryslân",
            "source": "original_entry",
            "note": "Regional library system. Wolvega is one of multiple branch locations."
        }
    },
    "1459_kb_isil.yaml": {
        "alternative_names": [
            {"name": "Bibliotheek Saba", "source": "original_entry", "note": "Island library system"},
            {"name": "Queen Wilhelmina Library", "source": "google_maps", "note": "Historical/popular name"}
        ],
        "custodian_name_update": {
            "claim_value": "Bibliotheek Saba",
            "source": "original_entry",
            "note": "Public library of Saba island (Caribbean Netherlands). Queen Wilhelmina Library is historical name."
        }
    },
    "1460_kb_isil.yaml": {
        "alternative_names": [
            {"name": "Bibliotheek Curaçao", "source": "original_entry", "note": "Short name"},
            {"name": "Nationale Bibliotheek Curaçao", "source": "google_maps", "note": "Full official name"}
        ],
        "custodian_name_update": {
            "claim_value": "Nationale Bibliotheek Curaçao",
            "source": "google_maps",
            "note": "National library of Curaçao. Both names are valid."
        }
    },
    "1463_kb_isil.yaml": {
        "alternative_names": [
            {"name": "Bibliotheken Noord Fryslân", "source": "original_entry", "note": "Library system name"},
            {"name": "Bibliotheek Dokkum", "source": "google_maps", "note": "Main branch location"}
        ],
        "custodian_name_update": {
            "claim_value": "Bibliotheken Noord Fryslân",
            "source": "original_entry",
            "note": "Regional library system serving North Friesland. Dokkum is one of multiple branches."
        }
    },
    "1467_kb_isil.yaml": {
        "alternative_names": [
            {"name": "Bibliotheek Berkel & IJssel", "source": "original_entry", "note": "Library system name (pre-merger)"},
            {"name": "BIJ de bieb / Zutphen", "source": "google_maps", "note": "Current brand name"}
        ],
        "custodian_name_update": {
            "claim_value": "Bibliotheek Berkel & IJssel",
            "source": "original_entry",
            "note": "Library system. 'BIJ de bieb' is a branded service/location name."
        }
    },
    "1489_kb_isil.yaml": {
        "alternative_names": [
            {"name": "Bibliotheek Krimpenerwaard", "source": "original_entry", "note": "Library system name"},
            {"name": "Bibliotheek Schoonhoven", "source": "google_maps", "note": "Branch location name"}
        ],
        "custodian_name_update": {
            "claim_value": "Bibliotheek Krimpenerwaard",
            "source": "original_entry",
            "note": "Regional library system. Schoonhoven is one of multiple branch locations."
        }
    }
}

ENTRIES_DIR = "/Users/kempersc/apps/glam/data/nde/enriched/entries"

def update_entry(filename: str, updates: dict):
    """Update an entry with alternative names and fix custodian_name."""
    filepath = os.path.join(ENTRIES_DIR, filename)
    
    if not os.path.exists(filepath):
        print(f"  ⚠ File not found: {filename}")
        return False
    
    with open(filepath, 'r', encoding='utf-8') as f:
        data = yaml.safe_load(f)
    
    # Add alternative_names if not present
    if 'alternative_names' not in data:
        data['alternative_names'] = []
    
    # Add new alternative names (avoid duplicates)
    existing_names = {alt.get('name', '') for alt in data.get('alternative_names', [])}
    for alt in updates['alternative_names']:
        if alt['name'] not in existing_names:
            data['alternative_names'].append(alt)
            existing_names.add(alt['name'])
    
    # Update custodian_name
    if 'custodian_name' in data and 'custodian_name_update' in updates:
        update_info = updates['custodian_name_update']
        data['custodian_name']['claim_value'] = update_info['claim_value']
        data['custodian_name']['source'] = update_info['source']
        data['custodian_name']['confidence'] = 0.90
        data['custodian_name']['manual_override'] = True
        data['custodian_name']['manual_override_reason'] = update_info['note']
        data['custodian_name']['manual_override_timestamp'] = datetime.now().isoformat()
    
    with open(filepath, 'w', encoding='utf-8') as f:
        yaml.dump(data, f, default_flow_style=False, allow_unicode=True, sort_keys=False)
    
    return True

def main():
    print("Adding alternative_names to library system entries...")
    print("=" * 60)
    
    updated = 0
    for filename, updates in LIBRARY_UPDATES.items():
        print(f"Processing: {filename}")
        if update_entry(filename, updates):
            print(f"  ✓ Updated with {len(updates['alternative_names'])} alternative names")
            updated += 1
    
    print("=" * 60)
    print(f"Updated {updated}/{len(LIBRARY_UPDATES)} entries")

if __name__ == "__main__":
    main()