Backend for the Apolo project (Yuno Hackathon) — A Spring Boot service that integrates Artificial Intelligence capabilities for risk profile analysis.
- Juan Sebastian Puentes Julio
- Daniel Patiño Mejia
- Tulio Riaño Sánchez
- Isaac Burgos
This section illustrates the system architecture. The images are stored in docs/diagrams/.
This high-level architecture diagram outlines the core components of our solution. We leverage external APIs for AI capabilities and integrations with platforms such as Slack and email. A non-relational database is utilized to efficiently handle high data volumes and flexible data structures.
This diagram details the specific services involved in the implementation. The central component is the Merchant Service, which orchestrates merchant creation, manages interactions, and coordinates calls to AI services.
Our data model is implemented using MongoDB to handle document-based storage. We utilize a combination of embedded and referenced documents to optimize performance and data integrity:
- Embedded: Merchant Context
- Referenced: Interactions
The class diagram illustrates the relationships between entities, enumerations, and the design patterns applied:
-
Builder Pattern: Used for constructing complex objects like
Merchant,MerchantContext, andInteractions, allowing for flexible object creation with varying attributes. -
Facade Pattern: Implemented to simplify file format conversions, providing a unified interface to delegate tasks to specific implementation classes.
This repository contains a Spring Boot backend (Java 17) that exposes REST endpoints and integrates with external AI services. It uses MongoDB for persistence and exposes API documentation via Swagger/OpenAPI.
- Java: 17 ☕
- Framework: Spring Boot 🍃
- Build Tool: Maven 🐘
- Database: MongoDB 🍃
- API Documentation: Swagger UI (springdoc) 📜
To run this project you need:
- Java 17 or newer.
- Maven 3.6+.
- A MongoDB instance (local or cloud).
The application reads configuration from src/main/resources/application.properties by default.
⚠️ IMPORTANT: Sensitive values (API Keys, DB credentials) should never be committed to the repository. It is recommended to use environment variables.
Key Properties:
spring.data.mongodb.uri: MongoDB connection URI.openai.api.key: OpenAI API Key.gladia.api.key: Gladia API Key.
export OPENAI_API_KEY="your-openai-key"
export GLADIA_API_KEY="your-gladia-key"
export MONGODB_URI="your-mongodb-uri"
mvn spring-boot:run -Dspring-boot.run.jvmArguments="-Dopenai.api.key=$OPENAI_API_KEY -Dgladia.api.key=$GLADIA_API_KEY -Dspring.data.mongodb.uri=$MONGODB_URI"
mvn clean package -DskipTests
java -Dopenai.api.key="$OPENAI_API_KEY" -Dgladia.api.key="$GLADIA_API_KEY" -Dspring.data.mongodb.uri="$MONGODB_URI" -jar target/ApoloBackend-0.0.1-SNAPSHOT.jar
We rely on Docker to support our development process, making it more efficient and optimizing resource usage. This tool allows us to standardize our development environment and avoid the overhead of installing heavy dependencies locally.
To start the services using Docker Compose:
docker-compose up -d- Build (with tests)
mvn clean package- Build (skip tests)
mvn clean package -DskipTests- Run Tests
mvn test