llm_extract_json.py

import json

def extract_json(text):
    """
    Extracts JSON objects or arrays from a given text, including nested structures.

    :param text: A string containing mixed content with JSON objects or arrays.
    :return: A list of extracted JSON objects or arrays.
    """
    json_objects = []
    stack = []
    start_idx = None

    for i, char in enumerate(text):
        if char in '{[':
            if not stack:
                start_idx = i  # Potential start of a JSON object/array
            stack.append(char)
        elif char in '}]':
            if stack:
                stack.pop()
                if not stack:
                    # Potential end of a JSON object/array
                    try:
                        json_obj = json.loads(text[start_idx:i+1])
                        json_objects.append(json_obj)
                    except json.JSONDecodeError:
                        pass  # Not a valid JSON, ignore

    return json_objects
  
  
response="""Sure! Here are the nutritional values for 100g of red cabbage in JSON format:
Sure! Here are the nutritional values for 100g of red cabbage in JSON format:

{
"ingredient": "red cabbage",
"protein": {
"value": 2.6,
"unit": "mg"
},
"fat": {
"value": 0.3,
"unit": "mg"
},
"oxalate": {
"value": 1.4,
"unit": "mg"
},
"sulfate": {
"value": 0.2,
"unit": "mg"
}
}
"""

print(extract_json(result))