Introduction

ai.FM is a web app that integrates generative AI into the antiquated ways of radio shows. Songs are chosen at random or submitted as wishes by the user and announced by an AI generated audio skit acting as one or multiple hosts.

Quick start

⚠️ Windows users:
We recommend using WSL 2 with Ubuntu 20.04+ to run the setup script.
To install WSL, follow Microsoft’s official guide.

To quickly get the application up and running, run this command on a machine running Ubuntu 20.04+ or a similar Linux distribution, with Docker installed:

curl -sS https://raw.githubusercontent.com/AET-DevOps25/team-in-the-midst-of-chaos-there-is-also-opportunity.-sun-tzu/main/setup.sh | bash

The script will:

• Create a project directory
• Prompt you for environment variables (without printing them to the screen)
• Start the application using Docker Compose

The application will become available at http://localhost:8080 (or whatever URL was set to). For more details, see Deployment and Setup.

Structure

Multiple Spring Boot microservices provide the base functionality for everything related to streaming songs and the handling of playlists utilizing a database and issuing the generation of audio skits on the fly.

Find more information here: server documentation

---

The GenAI service has its own separate REST API and is invoked every time an announcement needs to be generated.

Find more information here: genai documentation

---

The Angular frontend provides an easy to use UI for interacting with the backend. Users can insert song wishes to the radio's playlist and see the upcoming songs.

Find more information here: client documentation

---

Additionally, the entire project can be monitored via a Grafana dashboard that collects information on the Spring Boot microservices and the GenAI service using Prometheus.

Find more information here: monitoring documentation

Deployment and Setup

Prerequisites:

• Preferably Ubuntu 20.04+ or a similar Linux distribution. For Windows users, we recommend using WSL 2 with Ubuntu 20.04+ (see Microsoft’s official guide).
• Docker installed

Production Setup

The application can be deployed with Docker Compose as follows:

1. Create a directory aifm and navigate into it:

   mkdir aifm
   cd aifm

2. Download the compose.yml file from the GitHub repository into the folder:

   curl -sSo compose.yml https://raw.githubusercontent.com/AET-DevOps25/team-in-the-midst-of-chaos-there-is-also-opportunity.-sun-tzu/main/compose.yml

3. Create a .env file with the required environment variables:

   DOWNLOAD_PASS=your_password_here
   OPENAI_API_KEY=your_openai_key_here
   # URL=http://localhost:8080  # Optional; this should be the URL under which the client container will be reachable
   # VERSION=latest             # Optional

4. Pull the latest images and start the containers:

   docker compose up -d --pull always

5. Wait about a minute for the containers to fully initialize and connect.

6. The application is now running on port 8080. Open http://localhost:8080 (or whatever URL was set to) in your browser and enjoy some great music! 🎉

Development Setup

1. Clone the repository:

   git clone git@github.com:AET-DevOps25/team-in-the-midst-of-chaos-there-is-also-opportunity.-sun-tzu.git aifm

2. Navigate into the newly created directory:

   cd aifm

3. Create a .env file as in Production Setup step 3.

4. Build the images:

   docker compose build

5. Start the containers:

   docker compose up -d

6. Wait about a minute for the containers to fully initialize and connect.

7. Open http://localhost:8080 (or whatever URL was set to) in your browser to access the client container, which is the main entry point to the application. Alternatively, change the port to access the other containers:

   Port	Service
   8080	Web server that serves the Angular frontend and proxies API and Grafana requests. Main entry point to the application.
   8090	API Gateway
   8091	Stream service
   8092	Playlist service
   8093	Announcement service
   8000	GenAI service
   9090	Prometheus
   3000	Grafana
   3306	MySQL database

Software Engineering Process

The design process was thought out from the beginning with UML diagrams that set the relationships between objects, components and entities.

Analysis object model

Use case diagram

Component diagram of architecture