Tinysums is a Rails 8 app that periodically fetches the current top 30 stories from Hacker News and uses AI to generate summaries for both the articles and their comment threads. It features a React frontend embedded within the Rails application.

This README provides instructions for setting up and running the application locally.

• Ruby 3.4.2
• Rails 8.0.2
• Node.js 23.10.0
• Yarn
• PostgreSQL
• Redis
• Sidekiq
• React
• Tailwind CSS
• Gemini API
• Perplexity API

By default:

• Perplexity: Used for extracting structured information from story URLs.
• Gemini: Used for all other AI tasks, including generating summaries (for stories and comments) and performing classifications.

Before you begin, ensure you have the following installed:

• Ruby 3.4.2 (consider using a version manager like rbenv or rvm)
• Node.js 23.10.0 (consider using a version manager like nvm)
• Yarn package manager (npm install -g yarn)
• PostgreSQL database server
• Redis server

1. Clone the repository:

   git clone <repository-url> # Replace with the actual URL
   cd tinysums

2. Install Ruby dependencies:

   bundle install

3. Install Node.js dependencies:

   yarn install

4. Configure Environment Variables:

   Create a .env file in the root of the project. You will need to add API keys for Google AI (Gemini) and Perplexity, along with connection URLs for your database and Redis instance.

   # Example .env file
   DATABASE_URL=postgresql://<user>:<password>@<host>:<port>/tinysums_development
   REDIS_URL=redis://<host>:<port>/<db_number>
   GOOG_GEM_API_KEY=your_google_ai_gemini_api_key
   PERP_API_KEY=your_perplexity_api_key

   • Replace placeholders (<...>) with your actual database and Redis connection details. Ensure the user specified in DATABASE_URL exists and has privileges on the tinysums_development database.
   • Obtain API keys from Google AI Studio (for Gemini) and Perplexity Labs (ensure it's a key for their API, not just a Pro account).
   • API Keys are required for the core summarization functionality.

5. Setup the database:

   # Create the development and test databases
   rails db:create
   
   # Run database migrations
   rails db:migrate

1. Ensure PostgreSQL and Redis servers are running.

2. Start the application components (Web server, JS/CSS builders):

   Use the bin/dev command, which utilizes the Procfile.dev configuration:

   bin/dev

3. Access the application:

   Open your web browser and navigate to http://localhost:3000.

Important: The application will initially appear empty (no stories listed) when you first access it. You must run the TopStoriesSummaryJob (via Sidekiq) at least once to fetch and summarize the initial set of Hacker News stories. Subsequent runs of the job will refresh the stories database and regenerate summaries as needed.

1. Start Sidekiq:

   In a separate terminal window, run the following command from the project root to start the Sidekiq worker process:

   bundle exec sidekiq

2. Manually Triggering the Summary Job (Optional):

   The TopStoriesSummaryJob is responsible for fetching and summarizing stories. While it runs automatically hourly, you can trigger it manually for initial population or testing.

   Open a Rails console:

   bin/rails c

   Then, run the following command within the console to enqueue the job:

   # Enqueue the job to run asynchronously via Sidekiq
   TopStoriesSummaryJob.perform_async

   Note: This job fetches data and makes multiple AI API calls. It might take some time to complete depending on the number of stories and API response times. Monitor the Sidekiq process terminal for progress.

3. Automatic Hourly Job:

   As configured in config/schedule.yml, the TopStoriesSummaryJob is automatically scheduled to run via Sidekiq Cron at the top of every hour.

This project is released under the MIT License. See the LICENSE file for details.