count_database_entries.py

import json
import os
import re

def count_schools():
    """Count the number of schools in school_data_final.json"""
    school_data_file = os.path.join('data', 'school_data_final.json')
    
    try:
        with open(school_data_file, 'r', encoding='utf-8') as f:
            data = json.load(f)
        
        schools = data.get('schools', [])
        school_count = len(schools)
        
        print(f"\nNumber of schools in database: {school_count}")
        
        # Count schools by type
        school_types = {}
        for school in schools:
            school_type = school.get('type', 'Unknown')
            school_types[school_type] = school_types.get(school_type, 0) + 1
        
        print("\nSchools by type:")
        for school_type, count in school_types.items():
            print(f"  {school_type}: {count}")
            
        return school_count
    except Exception as e:
        print(f"Error counting schools: {e}")
        return 0

def count_suburbs():
    """Count the number of suburbs in geocode-db.js"""
    geocode_db_file = os.path.join('js', 'geocode-db.js')
    
    try:
        with open(geocode_db_file, 'r', encoding='utf-8') as f:
            content = f.read()
        
        # Find the start and end of the victorianSuburbsDB object
        start_marker = "const victorianSuburbsDB = {"
        end_marker = "};\n\n// Function to get coordinates"
        
        start_idx = content.find(start_marker) + len(start_marker)
        end_idx = content.find(end_marker)
        
        if start_idx == -1 or end_idx == -1:
            print("Could not find victorianSuburbsDB in geocode-db.js")
            return 0
            
        db_content = content[start_idx:end_idx].strip()
        
        # Count the number of suburb entries
        suburb_entries = re.findall(r"'[^']+': \{ lat:", db_content)
        suburb_count = len(suburb_entries)
        
        print(f"\nNumber of suburbs in database: {suburb_count}")
        
        # Count suburbs by postcode prefix
        postcode_regions = {}
        postcode_pattern = re.compile(r"postcode: '(\d{4})'")
        for line in db_content.split('\n'):
            match = postcode_pattern.search(line)
            if match:
                postcode = match.group(1)
                prefix = postcode[:1]  # First digit represents the state
                postcode_regions[prefix] = postcode_regions.get(prefix, 0) + 1
        
        print("\nSuburbs by postcode region:")
        for prefix, count in postcode_regions.items():
            region_name = "Victoria" if prefix == "3" else f"Region {prefix}"
            print(f"  {region_name}: {count}")
        
        return suburb_count
    except Exception as e:
        print(f"Error counting suburbs: {e}")
        return 0

def main():
    print("=== Schoolify Database Statistics ===")
    
    school_count = count_schools()
    suburb_count = count_suburbs()
    
    print("\n=== Summary ===")
    print(f"Total schools: {school_count}")
    print(f"Total suburbs: {suburb_count}")
    print("\nThis count was performed by examining:")
    print("- data/school_data_final.json for schools")
    print("- js/geocode-db.js for suburbs")

if __name__ == "__main__":
    main()