The swagger-rag project is an innovative open-source solution built with Kotlin and Spring Boot, designed to enhance the accessibility of API documentation through advanced natural language processing. This tool parses Swagger (OpenAPI) JSON files, transforms API descriptions into natural language document chunks, and facilitates Retrieval-Augmented Generation (RAG) capabilities via a command-line interface (CLI) chatbot. Additionally, it offers extensibility for integration with other platforms, such as SlackBot.

Key Features

• Swagger (OpenAPI) JSON Parsing: Efficiently extracts and processes API descriptions for improved usability.
• Natural Language Conversion: Each API is segmented into digestible language chunks, optimizing them for embedding and search functionalities.
• CLI Chatbot Interface: Engages users with a Minimal Viable Product (MVP) chatbot, allowing for intuitive Q&A sessions.
• Extensibility: Future-proof design enabling integration with various interfaces, including SlackBot.
• Vector Similarity Search: Ensures accurate retrieval of API documentation based on user queries.
• Enhanced Query Understanding: Leverages natural language processing for improved interpretation of user inquiries.

Example Usage

Interact with the CLI chatbot to uncover information about APIs effortlessly. For example:

> How do I create a new user?
Found 3 relevant APIs:

1. [POST] /users  
   Register New User  
   Registers a new user in the system.  
   **Responses:**  
   [201] User successfully created  
   Relevance score: 0.81  

2. [POST] /orders  
   Create New Order  
   Creates a new order.  
   **Responses:**  
   [201] Order successfully created  
   Relevance score: 0.72  

3. [POST] /products  
   Register New Product  
   Registers a new product in the system.  
   **Responses:**  
   [201] Product successfully created  
   Relevance score: 0.72

System Requirements

• JDK: Version 21 or higher is required.
• Build Tool: Gradle (wrapper included within the project).
• OpenAI API Key: Set the environment variable OPENAI_API_KEY for embedding functionality.
• Docker: Required to run the Qdrant Vector DB.

Contributions to this project are welcome! Collaborative efforts can be made through issues or pull requests aimed at fixing bugs, introducing new features, or enhancing functionality. For noteworthy changes, it is advisable to discuss proposals via issues prior to implementation.