GenAI PR Summarizer

A CLI tool to generate professional summaries for GitHub pull requests using LLMs (via Ollama) and OpenMP spec context.

---

Features

• Summarizes each file in a PR using a local LLM (Ollama, e.g., Llama3).
• Automatically references relevant OpenMP specification context for more accurate and professional summaries.
• Exports summaries as Markdown, plain text, or JSON.
• Works offline after LLM model is downloaded.
• CLI tool (pip install) and Standalone Windows executable available.
• Supports image attachments in summaries (see Exporting & Including Images).

---

Requirements

• Python 3.8+ (if using the pip version)
• Ollama LLM server (Download & Install Ollama)
  You must have Ollama installed and running on your machine for this tool to generate AI summaries.
• Llama3 model (or other supported models) pulled via ollama pull llama3
• GitHub Personal Access Token (for authenticated API access)
• Windows: For the .exe, no Python is needed. For source/pip: Python 3.8+

---

Installation

1. From Source (Recommended for Contributors)

Clone this repo and install locally:

git clone https://github.com/Manoj-Kumar-BV/GenAI-Powered-Pull-Request-Summaries.git
cd GenAI-Powered-Pull-Request-Summaries
pip install .

2. Editable Mode (for development)

pip install -e .

3. From GitHub (without cloning)

pip install git+https://github.com/Manoj-Kumar-BV/GenAI-Powered-Pull-Request-Summaries.git

Note:
pip install genai-pr-summarizer will not work unless the package is published on PyPI. If you want to install using that command, the maintainer must first publish the package to PyPI.

---

Ollama Installation & Setup

1. Download Ollama

• Go to: https://ollama.com/download
• Download and install for your platform (Windows, Mac, Linux).

2. Start the Ollama Server

Open a new terminal and run:

ollama serve

Make sure this terminal stays open while you use GenAI PR Summarizer.

3. Pull the LLM Model

If you have not already pulled a model (e.g. llama3):

ollama pull llama3

---

Usage

Option 1: Python CLI

1. Install the tool (see Installation above).
2. Set up your .env file with your GitHub token (cp .env.example .env and edit).
3. Start Ollama (see above).
4. Run the CLI:

   genai-pr-summarizer

Option 2: Windows Standalone (.exe)

1. Download and extract the genai-pr-summarizer.zip from GitHub Releases.
2. Copy .env.example to .env and add your GitHub token.
3. Start Ollama (ollama serve).
4. Open Command Prompt in the extracted folder.
5. Run:

   genai-pr-summarizer.exe

---

Exporting and Including Images

• When exporting summaries, you can attach images by placing them in the summaries/ folder (created on first export).
• Images will be referenced in the Markdown export using this format:

  ![image1](image1.png)
  

• Number images in reverse order of upload (last image is image1.png, second last is image2.png, etc.).

---

How OpenMP Specification Context is Used

For each file changed in a pull request, the tool:

• Extracts the code diff (patch),
• Uses a semantic search (via a helper script and FAISS) over a pre-indexed OpenMP specification,
• Retrieves the most relevant OpenMP spec sections,
• Feeds both the diff and spec context to the LLM,
• The summary references the spec, making reviews more accurate.

---

Troubleshooting

• Ollama not running:
  The CLI will check for Ollama. If not running, you'll see an error like:
  Error: Ollama server is not running at http://localhost:11434. Please start Ollama (see Requirements above).

• First-time model load:
  The first time a model is used, Ollama will download it. This may take several minutes.

• No AI summaries:
  If Ollama is not running, or the required model is missing, the tool cannot generate summaries.

• pip install error:
  If you see ERROR: Could not find a version that satisfies the requirement genai-pr-summarizer, it means the package is not published to PyPI. Use pip install . from your local directory instead.

---

Support

For issues, please open an issue in this repository.

---