Yuxi-Know - An AI knowledge graph question answering platform based on a large model RAG knowledge base
Yuxi-Know is an intelligent question-answering platform built on a large-scale RAG knowledge base and knowledge graph technology. Yuxi-Know supports various knowledge base file formats (such as PDF, TXT, MD, Docx) and supports converting file content...
What is Yuxi-Know?
Yuxi-Know is an intelligent question-answering platform built on a large-scale RAG knowledge base and knowledge graph technology. Yuxi-Know supports various knowledge base file formats (such as PDF, TXT, MD, and Docx) and allows conversion of file content into vector storage for rapid retrieval. Yuxi-Know integrates Neo4j-based knowledge graph question-answering capabilities, enabling it to handle complex knowledge relationship queries. The platform supports multi-model adaptation, including OpenAI, mainstream domestic large-scale models, and locally deployed VLLM and Ollama. Yuxi-Know features intelligent agent extension capabilities, allowing developers to write custom intelligent agent code, further enhancing the system's flexibility and functionality.
Main functions of Yuxi-Know
- Multi-model supportIt supports multiple large model platforms, including OpenAI, mainstream domestic large model platforms, and locally deployed VLLM and OLLAM.
- Flexible knowledge base managementSupports uploading documents in multiple formats such as PDF, TXT, MD, and Docx.
- Knowledge Graph IntegrationBased on Neo4j's knowledge graph question answering capabilities, it supports users uploading knowledge graph files in JSONL format.
- Intelligent Agent ExpansionIt supports users in writing custom intelligent agent code, further expanding the system's functionality.
- Inference model supportSupports inference models such as Deepseek-R1, providing a more intelligent question-answering experience.
- Web SearchSupports webpage content retrieval and display.
- Visual configurationIt offers rich visual configuration features, allowing users to easily manage and optimize system settings.
The technical principle of Yuxi-Know
- RAG architectureThe system uses a vector model to convert text in the knowledge base into vectors and stores them in a vector database. When a user asks a question, the system retrieves the most relevant document fragment from the vector database. This retrieved document fragment, along with the user's question, is then input into a large language model to generate an accurate answer.
- Knowledge graph technologyNeo4j serves as a storage and query engine for knowledge graphs. Knowledge graphs store knowledge in the form of nodes and relationships, allowing users to perform complex knowledge relationship queries based on the graph. The construction and management of knowledge graphs are completed through a web interface, enabling users to easily upload and update graph data.
- Multi-model adaptationBased on configuration files (such as models.yaml), the system supports API calls from various large model platforms. Users can select different models as needed, and the system will automatically adapt and call the corresponding APIs.
- Vector Model and Reordering ModelVector models convert text into vectors and support various vector models such as BAAI/bge-m3. Reordering models optimize the order of search results, improving the accuracy and efficiency of question answering.
- Front-end and back-end technology stackThe front-end is based on VueJS, providing a user-friendly interface. The back-end is based on FastAPI, providing efficient server-side processing capabilities. Milvus-Standalone is used as the vector database, and Neo4j is used as the knowledge graph database.
- Docker containerized deploymentDocker enables containerized deployment, facilitating the rapid setup and management of development and production environments. It supports automatic updates and background operation of the development environment.
Yuxi-Know's project address
- GitHub repository:https://github.com/xerrors/Yuxi-Know
Application scenarios of Yuxi-Know
- Enterprise knowledge managementBuild an enterprise knowledge base so that employees can quickly look up information and improve work efficiency.
- EducationIt assists online learning by providing students with intelligent tutoring and knowledge answers.
- Customer ServiceAs an intelligent customer service system, it quickly answers customer inquiries and improves customer satisfaction.
- HealthcareIt provides access to medical knowledge and initial patient consultations to assist in medical decision-making.
- Scientific research and academicIt supports literature retrieval and knowledge graph construction, thus assisting scientific research.