Convert recipe PDFs into structured JSON

Here’s a clear approach and example for converting recipe PDFs into structured JSON format. This includes typical fields to extract and how the JSON might look.

Key Fields to Extract from Recipe PDFs:

• title: Name of the recipe

• description: Short summary or introduction

• ingredients: List of ingredients with quantity and unit

• instructions: Step-by-step cooking directions

• prep_time: Preparation time

• cook_time: Cooking time

• total_time: Total time (prep + cook)

• servings: Number of servings

• notes: Additional tips or notes (optional)

Example JSON Structure for a Recipe

json
CopyEdit
{
  "title": "Classic Pancakes",
  "description": "Fluffy and light pancakes perfect for breakfast.",
  "ingredients": [
    {
      "name": "all-purpose flour",
      "quantity": 1.5,
      "unit": "cups"
    },
    {
      "name": "baking powder",
      "quantity": 3.5,
      "unit": "teaspoons"
    },
    {
      "name": "salt",
      "quantity": 1,
      "unit": "teaspoon"
    },
    {
      "name": "white sugar",
      "quantity": 1,
      "unit": "tablespoon"
    },
    {
      "name": "milk",
      "quantity": 1.25,
      "unit": "cups"
    },
    {
      "name": "egg",
      "quantity": 1,
      "unit": "large"
    },
    {
      "name": "butter, melted",
      "quantity": 3,
      "unit": "tablespoons"
    }
  ],
  "instructions": [
    "In a large bowl, sift together the flour, baking powder, salt, and sugar.",
    "Make a well in the center and pour in the milk, egg, and melted butter; mix until smooth.",
    "Heat a lightly oiled griddle or frying pan over medium high heat.",
    "Pour or scoop the batter onto the griddle, using approximately 1/4 cup for each pancake.",
    "Brown on both sides and serve hot."
  ],
  "prep_time": "10 minutes",
  "cook_time": "15 minutes",
  "total_time": "25 minutes",
  "servings": 4,
  "notes": "Add blueberries or chocolate chips for a twist."
}

How to Implement PDF to JSON Conversion

1. Extract text from PDF
   Use libraries like PyMuPDF (fitz), pdfplumber, or pdfminer.six to extract raw text from the PDF.

2. Parse the text
   Identify and isolate recipe sections:

   • Title usually at the top or bolded

   • Ingredients often in a bulleted or numbered list with quantities

   • Instructions typically numbered or stepwise paragraphs

   • Times and servings may be listed under metadata or a separate section

3. Use Regex or NLP
   Apply regular expressions or NLP techniques to:

   • Extract quantities and units from ingredients

   • Split instructions into steps

   • Detect time formats (e.g., “10 minutes”, “1 hour”)

   • Extract servings (e.g., “Serves 4”)

4. Structure the data
   Assemble extracted data into a JSON object as shown above.

If you want, I can help create a Python script example that processes a recipe PDF and outputs JSON structured like this. Would you like that?