✨ LLM Web Search Agent

🚀 Overview

This project enables a Language Model (LLM) to search the web and scrape data from the search results to build a knowledge base. The LLM then uses this enriched knowledge base to answer user queries with greater accuracy and context.

The LLM Web Search Agent is built using Java (Spring Boot) and SearxNG as the web search engine. It combines advanced AI-powered semantic understanding with robust web search functionalities, creating a powerful and efficient solution for handling user queries.

---

✨ Features

• Seamless Language Model Integration:
  • Support for configurable LLM APIs (e.g., OpenAI).
  • Embedding-based processing for improved query understanding.
• Powerful Search Engine Support:
  • SearxNG ensures reliable, privacy-respecting web searches.
• Comprehensive Logging:
  • Track requests and responses for debugging and analysis.
  • Debug-level logs for detailed troubleshooting.

---

📜 Requirements

To get started, make sure you have the following:

1. Java 21 or later
2. Maven for dependency management
3. Docker and Docker Compose to run SearxNG

---

⚙️ Configuration

Application Properties

Prepare the application.properties file with the required configurations:

🌱 Spring Configuration

spring.application.name=smartAgent

🤖 LLM Configuration

• model-url: URL of the LLM API.
• model-api-key: Your API key for the LLM.
• embedding-api-key: API key for embeddings.
• model-name: The name of the LLM to use.
• embedding-model-name: Name of the embedding model.

🔍 Search Engine Configuration

• search-engine-url: The URL for the SearxNG search engine.

📜 Logging Configuration

Enable detailed request and response logging:

langchain4j.open-ai.chat-model.log-requests=true
langchain4j.open-ai.chat-model.log-responses=true
logging.level.dev.langchain4j=DEBUG
logging.level.dev.ai4j.openai4j=DEBUG

🛠️ SearxNG Configuration

1. Navigate to the searxng-docker directory.
2. Ensure the necessary environment variables are set in the Docker Compose file.
3. Start SearxNG using:

   docker-compose up -d

---

🛠️ Build and Run

1. Build the Application:

   mvn clean package

2. Run the Application:

   java -jar target/<your-jar-name>.jar

3. Test the Endpoint: Use the following example curl command to test:

   curl --location 'http://localhost:8080/web-search-agent?question=weatherin2025-1-1' \
   --data ''

---

📦 Dependencies

• Spring Boot: Framework for building modern web applications.
• LangChain4j: For seamless LLM and embedding integration.
• Lombok: Simplify Java development with powerful annotations.
• SearxNG: A metasearch engine respecting your privacy.

---

🔖 Logging

The application offers detailed logging capabilities:

• Requests & Responses: Logged by LangChain4j for transparency.
• Debugging: Enabled for in-depth troubleshooting.

---

🤝 Contributing

We welcome contributions from the open-source community! Here’s how you can help:

1. Fork the repository and create a new branch.
2. Submit feature requests, report bugs, or suggest enhancements.
3. Open pull requests with your improvements.

Together, we can make this project even better! 💪

---

📜 License

This project is licensed under the MIT License. Feel free to use, modify, and distribute it as you see fit.

---

❤️ Acknowledgments

We’d like to thank the creators of these amazing technologies:

• Spring Boot
• LangChain4j
• SearxNG

---

🌟 Join Us!

Help us grow the LLM Web Search Agent into the ultimate search and AI tool. Let’s build something incredible together!

---

🔦 Swagger API Documentation

1. After starting the Spring Boot application, visit http://localhost:8080/swagger-ui/index.html to access the Swagger API documentation.
2. Alternatively, the pre-generated Swagger documentation is located in the swagger folder. Open index.html(https://github.com/alanta335/llm_web_search_agent_java/blob/main/swagger/index.html) to view it locally.