PR-Story is a GitHub App that transforms the commit messages of a pull request into a visual narrative. It uses AI to generate a story and accompanying images, which are then posted as a comment on the pull request. This provides a unique and intuitive way to understand the changes introduced in a PR.

• Listens for Pull Request Events: The app is triggered when a new pull request is opened.
• AI-Powered Narrative Generation: It uses an OpenAI model to convert commit messages into a compelling story.
• Text-to-Image Generation: A Diffusion Pipeline from Hugging Face's diffusers library creates images based on the generated story.
• Automated PR Commenting: The generated images and their corresponding story captions are posted as a comment on the pull request.

1. Webhook Trigger: A webhook is configured to listen for pull_request events from GitHub.
2. Fetch Commit Messages: When a pull request is opened, the app fetches all the commit messages associated with it.
3. Generate Story and Prompts: The commit messages are sent to an OpenAI model (GPT-4o) to generate a narrative. This narrative is broken down into a series of prompts for the text-to-image model.
4. Generate Images: The prompts are fed into a pre-trained text-to-image model (playgroundai/playground-v2.5-1024px-aesthetic) to generate a sequence of images.
5. Upload Images: The generated images are uploaded to Imgur to get public URLs.
6. Post Comment: A comment is created on the pull request, containing the images and their corresponding captions, forming a visual story of the changes.

1. Clone the Repository:

   git clone https://github.com/BH3325/pr_story.git
   cd pr_story

2. Create a GitHub App:

   • Go to your GitHub settings and create a new GitHub App.
   • Set the Webhook URL to your application's public URL (e.g., https://your-domain.com/webhooks).
   • Create a Webhook Secret and save it.
   • Under Permissions, grant the following repository permissions:
     • Pull requests: Read & write
     • Contents: Read-only
   • Install the app on your target repositories.
   • Generate a private key for your app and download it.

3. Set Up Environment Variables: Create a .env file in the root of the project and add the following variables:

   APP_ID="your_github_app_id"
   WEBHOOK_SECRET="your_github_app_webhook_secret"
   PK="your_base64_encoded_private_key"
   OPENAI_API_KEY="your_openai_api_key"
   IMGUR_CLIENT_ID="your_imgur_client_id"
   

   • PK should be the base64 encoded version of your GitHub App's private key. You can encode it using:

     cat your-private-key.pem | base64

4. Install Dependencies: This project uses uv for package management.

   pip install uv
   uv sync

5. Run the Application: Use uvicorn to run the FastAPI application:

   uv run uvicorn main:app --host 0.0.0.0 --port 8000

Once the app is running and installed on a repository:

1. Open a new pull request in that repository.
2. The app will automatically process the commit messages and post a comment with the visual story.

• Fine tuning: Improve specificity of generated images to better match code context.
• Theme customisation: Allow users to select different art styles for image generation.
• Audio integration: Generate songs or soundscapes to accompany the visual story.