gpn23-recipes/1_clean_json.py

32 lines
1 KiB
Python
Raw Normal View History

2025-06-01 23:11:38 +02:00
import json
input_file = 'export_2025-06-01.json'
output_file = 'export_2025-06-01_clean.json'
with open(input_file, 'r', encoding='utf-8') as f:
data = json.load(f)
def normalize_ingredients(recipe):
if 'recipeIngredient' in recipe:
recipe['recipeIngredient'] = [
ingredient.replace('g / Gramm', 'g') for ingredient in recipe['recipeIngredient']
]
recipe['recipeIngredient'] = [
ingredient.replace('kg / Kilogramm', 'kg') for ingredient in recipe['recipeIngredient']
]
recipe['recipeIngredient'] = [
ingredient.replace('.0 ', ' ') for ingredient in recipe['recipeIngredient']
]
return recipe
if isinstance(data, list):
cleaned_data = [normalize_ingredients(recipe) for recipe in data]
else:
cleaned_data = normalize_ingredients(data)
# Write the cleaned JSON data to a new file
with open(output_file, 'w', encoding='utf-8') as f:
json.dump(cleaned_data, f, ensure_ascii=False, indent=2)
print(f"Cleaned data written to '{output_file}'")