DevBcn Social Assistant

A conversational AI assistant for the DevBcn conference that helps with creating social media content, newsletters, and providing information about speakers and sessions.

Project Description

The DevBcn Social Assistant is a Spring Boot application that leverages AI models (either OpenAI or Ollama) to provide information about the DevBcn conference, generate social media posts, and create newsletter content. It features a chat interface built with Vaadin where users can interact with the assistant.

The assistant can:

• Provide information about the conference (dates, venue, tracks, etc.)
• Generate social media posts for different platforms (Twitter, Instagram, LinkedIn, Bluesky)
• Create Mailchimp newsletter content
• Retrieve and present information about speakers and sessions

Technologies Used

• Java 21: The application is built using Java 21
• Spring Boot 3.5.3: Framework for building the application
• Spring AI 1.0.0: Integration with AI models
  • Support for OpenAI models (default)
  • Support for Ollama models (alternative)
• Vaadin 24.8.0: Web framework for building the user interface
• Docker Compose: For running Ollama locally
• TestContainers: For testing with containerized dependencies

Project Structure

• src/main/java/com/devbcn/socialassistant/
  • SocialAssistantApplication.java: Main application class
  • config/: Configuration classes
  • dto/: Data transfer objects for sessions, speakers, etc.
  • service/: Service classes for business logic
    • ChatService.java: Handles interactions with AI models
    • SessionizeService.java: Provides tools for the AI to access conference data
    • SpeakerService.java: Manages speaker information
    • SessionService.java: Manages session information
  • views/: Vaadin UI components
    • HomeView.java: Main chat interface
• src/main/resources/
  • application.properties: Application configuration
  • system_message.md: Instructions for the AI model

Setup and Installation

Prerequisites

• Java 21 JDK
• Maven
• Docker (for Ollama profile)

Configuration

1. Create a .env.local file in the project root with your OpenAI API key:

   API_KEY=your_openai_api_key
   

2. Choose your AI provider profile:

   • OpenAI (default): Uses OpenAI models (requires API key)
   • Ollama: Uses local Ollama models (requires Docker)

Building and Running

Using Maven

# Build the project
mvn clean install

# Run with OpenAI profile (default)
mvn spring-boot:run

# Run with Ollama profile
mvn spring-boot:run -Pollama

Using Docker Compose (for Ollama profile)

# Start Ollama container
docker-compose up -d

# Run the application with Ollama profile
mvn spring-boot:run -Pollama

Usage

1. Access the application at http://localhost:8080
2. Use the chat interface to interact with the assistant
3. Ask questions about the DevBcn conference
4. Request social media posts or newsletter content

Examples

• "Tell me about the DevBcn conference"
• "Create a Twitter post announcing speaker John Doe"
• "Generate a Mailchimp newsletter about early bird tickets"
• "What sessions are available in the Java track?"

Development

Running Tests

mvn test

Profiles

• production: Optimizes the application for production
• openai: Uses OpenAI models (default)
• ollama: Uses Ollama models with Docker Compose integration

License

MIT

Contributors

• Anyul Rivas