COBOL Enhancer

Overview

COBOL Enhancer is a Python-based tool designed to optimize COBOL code. Using the power of AI, it refactors COBOL code to be more efficient and error-free while maintaining the original functionality and structure of the code.

Quick Start

Anything-LLM, DAnswer, Dify, FastGPT, GenAI-Stack, Quivr

Prerequisites

• Python 3.9+
• Poetry for dependency management
• Direnv for environment variable management

Installation

1. Clone the repository from GitHub:

   git clone https://github.com/thibaudbrg/cobol_enhancer.git

2. Navigate to the cloned repository directory:

   cd cobol_enhancer

   4. Install dependencies using Poetry:

   poetry install

Environment Setup

1. Create a .env file by copying the .env.example template:

   cp .env.example .env

2. Fill in your Langchain and OpenAI API keys in the .env file.
3. Load the environment variables:

   eval "$(direnv hook bash)" direnv allow

Running the Application

1. Start the Langchain server:

   langchain serve

2. Run the tests for the cobol_enhancer workflow:

   poetry add pytest --group dev poetry run python -m pytest packages/ubp-cobol/tests/test_chain.py::test_workflow -s

3. To display a side-by-side comparison of the old and new COBOL code:

   poetry run python -m pytest packages/ubp-cobol/tests/test_chain.py::test_print_workflow -s

Program Workflow

The cobol_enhancer application is designed to iteratively improve a directory of COBOL code files using an AI-based language model (LLM). The process flow is as follows:

1. Input: The program takes a directory containing COBOL code files as its input.

2. Enhancement: Each COBOL file is passed through the LLM, which refactors the code to improve efficiency, readability, and adherence to modern coding practices while preserving the original functionality and structure.

3. Human Review: After an enhancement is suggested by the LLM, a human reviewer is prompted to either accept the changes or provide specific feedback for further refinement.

4. Iterative Generation: Based on the human reviewer's feedback, the LLM re-attempts to enhance the code. This loop continues until the reviewer is satisfied with the enhancements.

5. Completion: Once the code is approved by the human reviewer, it is deemed finalized, and the program proceeds to the next COBOL file in the directory.

Below is a visual representation of the workflow within cobol_enhancer:

License

Distributed under the MIT License. See LICENSE for more information.