force_parse_json

force_parse_json is a Python function designed to handle incomplete or 'partial' JSON strings. It first tries to 'close' or complete the JSON, then attempts to parse it. If the parsing is successful, it optionally cleans up the resulting Python dictionary before returning it.

If parsing fails, the function either returns None or prints an error message, depending on the parameters you specify.

Usage

from gpt_stream_parser import force_parse_json

output = ""
for chunk in response:
    output += chunk.choices[0].delta.function_call.arguments
    data = force_parse_json(output)
    if data:
        print(data)

Parameters

• partial_json (str): The incomplete JSON string.
• report_error (bool, optional): If set to True, any exceptions that occur during parsing will be printed to the console. If False, they will be ignored. Default is False.
• clean_up (bool, optional): If set to True, the resulting Python dictionary will be 'cleaned up' before it is returned. If False, the function will return the dictionary as is. Default is True.

Returns

• dict or None: The function returns the parsed JSON as a Python dictionary if successful. If an error occurs during parsing, it returns None.

Example

This example uses the force_parse_json function in the context of streaming API responses. It uses the OpenAI GPT-3.5 model to obtain chat completions, which are streamed in chunks. The function attempts to parse the data contained in each chunk and print it to the console.

import openai
from gpt_stream_parser import force_parse_json

# create a chat completion
response = openai.ChatCompletion.create(
    model="gpt-3.5-turbo-0613",
    messages=[{"role": "user", "content": "What's the weather like in Boston?"}],
    functions=[
        {
            "name": "get_current_weather",
            "description": "Get the current weather in a given location",
            "parameters": {
                "type": "object",
                "properties": {
                    "location": {
                        "type": "string",
                        "description": "The city and state, e.g. San Francisco, CA",
                    },
                    "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]},
                },
                "required": ["location"],
            },
        }
    ],
    function_call="auto",
    stream=True,
)

# parse and print the data contained in each chunk
output = ""
for chunk in response:
    output += chunk.choices[0].delta.function_call.arguments
    data = force_parse_json(output)
    if data:
        print(data)

In this example, force_parse_json is used to parse the data contained in each chunk of the response. The parsed data is then printed to the console. Note that this function is useful for handling streaming API responses, where data may be received in incomplete 'chunks'.