gpn23-recipes/2_to_markdown.py

import json
import sys
from pathlib import Path
from datetime import timedelta
import re
from pathvalidate import validate_filename
import os
from tqdm import tqdm

include_title = False
output_dir = "recipes-md"


def valid_duration(duration):
    if not duration:
        return False
    if duration == "0":
        return False
    return True

def valid_description(description):
    if not description:
        return False
    return True

def parse_duration(duration):
    if not duration:
        return ""
    match = re.match(r"PT(?:(\d+)H)?(?:(\d+)M)?", duration)
    if not match:
        return duration
    hours, minutes = match.groups()
    hours = f"{int(hours)}h " if hours else ""
    minutes = f"{int(minutes)}m" if minutes else ""
    return f"{hours}{minutes}".strip()

def format_recipe_to_markdown(recipe):
    md = []

    # Title (may be implicit by filename)
    if include_title:
        md.append(f"# {recipe.get('name', 'Untitled Recipe')}")

    # Description
    if valid_description(recipe.get('description')):
        md.append(f"\n_{recipe['description']}_\n")

    # Details
    details_parts = []
    if valid_duration(recipe.get('prepTime')):
        details_parts.append(f"Prep time: {parse_duration(recipe['prepTime'])}")
    if valid_duration(recipe.get('cookTime')):
        details_parts.append(f"Cook time: {parse_duration(recipe['cookTime'])}")
    if valid_duration(recipe.get('totalTime')):
        details_parts.append(f"Total time: {parse_duration(recipe['totalTime'])}")
    if 'recipeYield' in recipe:
        details_parts.append(f"Portionen: {recipe['recipeYield']}")
    if details_parts:
        for p in details_parts:
            md.append(f"* {p}\n")

    # Ingredients
    if 'recipeIngredient' in recipe:
        md.append("\n## Zutaten")
        for ingredient in recipe['recipeIngredient']:
            md.append(f"- {ingredient}")

    # Instructions
    if 'recipeInstructions' in recipe:
        md.append("\n## Zubereitung")
        instructions = recipe['recipeInstructions']
        if isinstance(instructions, list):
            for i, step in enumerate(instructions, 1):
                if isinstance(step, dict):
                    text = step.get('text', '').strip()
                else:
                    text = str(step).strip()
                md.append(f"{i}. {text.replace(chr(10), ' ')}")
        else:
            for i, line in enumerate(instructions.strip().split('\n'), 1):
                if line.strip():
                    md.append(f"{i}. {line.strip()}")

    # Nutrition
    if 'nutrition' in recipe:
        md.append("\n## Nutrition")
        for key, value in recipe['nutrition'].items():
            if key != "@type":
                md.append(f"- **{key.replace('_', ' ').capitalize()}**: {value}")

    return '\n'.join(md)

def main(json_file):
    with open(json_file, 'r', encoding='utf-8') as f:
        data = json.load(f)

    # If the JSON-LD is embedded in @graph or a list
    # if isinstance(data, list):
    #     recipe = next((item for item in data if item.get('@type') == 'Recipe'), data[0])
    # elif '@graph' in data:
    #     recipe = next((item for item in data['@graph'] if item.get('@type') == 'Recipe'), data['@graph'][0])
    if isinstance(data, list) and data[0].get('@type') == 'Recipe':
        recipes = data
    elif data.get('@type') == 'Recipe':
        recipes = [data]
    else:
        print("No Recipe object found in JSON.")
        return

    for recipe in tqdm(recipes):
        markdown_fn = f"{recipe.get('name', 'Untitled Recipe')}.md"
        validate_filename(markdown_fn) # XXX does this check directory traversal?
        markdown = format_recipe_to_markdown(recipe)
        os.makedirs(output_dir, exist_ok=True)
        with open(os.path.join(output_dir, markdown_fn), "w") as f:
            f.write(markdown)

if __name__ == "__main__":
    if len(sys.argv) != 2:
        print("Usage: python recipe_to_md.py <path_to_recipe_json>")
    else:
        main(sys.argv[1])