🚀 Spring AI + Qdrant

🛠 Technologies

• Spring Boot 3.5.x — Java microservices framework
• Spring AI 1.x — unified AI abstraction
• Qdrant — vector database for semantic search
• Docker Compose — for local AI & DB orchestration
• JUnit 5 — unit & integration testing
• Lombok — concise Java code
• Gradle (Kotlin DSL) — build system

🏗 Architecture

⚡ Quick Start

1. Start dependencies:

   docker compose -f docker/docker-compose.yml up -d

   • Ollama: http://localhost:11434
   • Qdrant: http://localhost:6333
   • Qdrant Dashboard: http://localhost:6333/dashboard

2. Build & run:

   ./gradlew clean build
   ./gradlew bootRun

🔗 API

You can set model: phi3 (default) or use a more powerful model, e.g. deepseek-r1-distill-qwen-7b for better quality.

• POST /api/ai/chat — Regular LLM chat
• POST /api/ai/chat-rag — Chat with RAG
• POST /api/rag/save-text — Save texts to vector DB
• POST /api/rag/save-doc — Index document (PDF/TXT/JSON)
• POST /api/rag/search — Search in vector DB
• GET /api/rag/count — Documents count

🧪 Testing

Unit & integration tests: src/test/java

./gradlew test