JupyterDebug 0.1.1

A package for debugging Jupyter notebooks using OpenAI.

Documentation for JupyterDebug

Package Overview

JupyterDebug is a Python package designed to help users debug Python code in Jupyter notebooks using OpenAI's language models. It provides two main methods:

1. init(): Initializes the package by prompting the user for their OpenAI API key and selecting a model.
2. debug(max_iter=1): Automatically debugs the code in the previous cell by calling the selected language model iteratively to fix errors.

---

Installation

To install JupyterDebug, use pip:

pip install JupyterDebug

---

Usage

1. init()

The init() method initializes the package by prompting the user for their OpenAI API key and selecting a model. It must be called before using the debug() method.

Syntax:

import JupyterDebug as jd
jd.init()

Behavior:

• Prompts the user to enter their OpenAI API key.
• Asks the user to select a model from the following options:
  • gpt-4o-mini (default)
  • gpt-4o
  • o1
  • o1-mini
• Stores the API key and model selection in environment variables for use in the debug() method.

---

2. debug(max_iter=1)

The debug() method automatically debugs the code in the previous cell by calling the selected language model iteratively to fix errors. It should be called in a new cell immediately after the cell that generated the error.

Syntax:

jd.debug(max_iter=1)

Parameters:

• max_iter (optional, default=1): The maximum number of times the language model is called to iteratively debug the code. If the first solution does not fix all errors, the method will retry up to max_iter times.

Behavior:

• Captures the code from the previous cell and the error trace.
• Calls the selected language model to generate a fix for the code.
• Replaces the code in the previous cell with the fixed version.
• If max_iter > 1, the process is repeated until all errors are fixed or the maximum number of iterations is reached.

Example:

# Cell 1: Code with an error
x = 10
y = "20"
z = x + y  # This will raise a TypeError

# Cell 2: Debug the error
jd.debug()

---

Example Workflow

1. Install the package:

   pip install JupyterDebug
   

2. Initialize the package:

   import JupyterDebug as jd
   jd.init()
   

3. Write and run code in a Jupyter notebook cell:

   # Cell 1: Code with an error
   x = 10
   y = "20"
   z = x + y  # This will raise a TypeError
   

4. Debug the error in the next cell:

   # Cell 2: Debug the error
   jd.debug(max_iter=2)