Problem Statement:

Writing clear and detailed pull request (PR) descriptions for OpenMP in the LLVM project takes a lot of time and often lacks consistency. Developers need a simple way to automatically create good PR summaries that also follow the OpenMP specification.

Objectives:

• Build a Clang CLI tool to automatically generate structured PR descriptions.
• Use code diffs and OpenMP spec content to create accurate summaries.
• Integrate Generative AI to write clear and informative descriptions.
• Make it easy for developers to get spec-aware PR summaries with just a PR number.

Methodology

• Fetch PR Data: Retrieve pull request title, changed files, and code diffs from GitHub using the PR number.
• Extract Title & Patches: Separate and store the PR title and code changes for analysis.
• Keyword Extraction: Identify key terms from the title and patch by removing stopwords and irrelevant tokens.
• Spec Matching: Match extracted keywords to relevant sections of the OpenMP specification using text similarity.
• AI Summary Generation: Use a GenAI model (LLaMA3-70B) to generate a structured PR description using the patch and matched spec section.

How to Run

• Install python 3.10+
• Create a python environment and activate

python -m venv envname
source envname/bin/activate #for mac or linux
Script/activate #for windows

• Install the required dependencies

pip install -r requirements.txt

• Run app.py

python app.py

Images of the Solution