import json
import pandas as pd

with open('./input_files/Lachine_Geojson_Corrected.geojson', 'r') as f:
    geojson_data = json.load(f)

for feature in geojson_data['features']:
    properties = feature.get('properties', {})
    properties.pop('detailed_model_filename', None)

excel_df = pd.read_excel('./input_files/building_usage.xlsx')
excel_df = excel_df.set_index('order of the file in geojson')
excel_df = excel_df.sort_index()

usage_codes = {
    'Residential': "1000",
    'Commercial': "6111",
    'Industrial': "4413",
    'Institutional': "6812",
    'Office': "6591"
}

features = geojson_data['features']

assert len(features) == len(excel_df)

print("The files' ranges match.")

for idx, (feature, (excel_idx, row)) in enumerate(zip(features, excel_df.iterrows()), start=1):
    properties = feature.get('properties', {})

    properties.pop('detailed_model_filename', None)

    # Get the surface areas from the Excel row
    commercial_area = row.get('Sup. commerciale (m²)', 0) or 0
    industrial_area = row.get('Sup. indus.léger (m²)', 0) or 0
    institutional_area = row.get('Sup. institutionnelle (m²)', 0) or 0
    office_area = row.get('Sup. bureau (m²)', 0) or 0
    residential_area = row.get('Résidentiels (m²)', 0) or 0

    # Create a dictionary of areas with their corresponding usage codes
    areas = {
        usage_codes['Commercial']: commercial_area,
        usage_codes['Industrial']: industrial_area,
        usage_codes['Institutional']: institutional_area,
        usage_codes['Office']: office_area,
        usage_codes['Residential']: residential_area
    }

    # Remove entries with zero area
    areas = {code: area for code, area in areas.items() if area > 0}

    total_area = sum(areas.values())

    if total_area == 0:
        print(f"Total area is zero for feature at index {idx}, skipping modification.")
        continue

    # Compute the percentage for each usage
    usages_list = []
    for code, area in areas.items():
        percentage = (area / total_area) * 100
        usages_list.append({'usage': code, 'percentage': round(percentage)})

    # Update 'building_type' and 'usages' in properties
    if len(usages_list) > 1:
        properties['building_type'] = 'mixed use'
        properties['usages'] = usages_list
    elif len(usages_list) == 1:
        # Single usage
        properties['building_type'] = list(areas.keys())[0]
    else:
        # No usages, skip modifying this feature
        continue

    # Add 'area' attribute from Excel
    properties['area'] = str(row.get('Area', ''))

    # Replace 'id' with 'development' from Excel
    development = row.get('development', '')
    cleaned_development = str(development).replace('-', '').replace(' ', '')
    properties['id'] = "Building" + cleaned_development

with open('input_files/Lachine_Geojson_Mixed_Use.geojson', 'w') as f:
    json.dump(geojson_data, f, indent=2)