RAG Node
The Onyx RAG Node provides you with the flexibilty needed to create robust and complete RAG pipelines from scratch.
Creation
Choose your Vector Database and corresponding RAG metrics
RAG metrics include the embedding model, Vector Space, Similarity Metric, and metadata values to be included
Connect your Data Sources and click submit
Once Complete the OCE will build all the infrastructure you will need to leverage this RAG pipeline within Onyx
Usage
The RAG node will be responsible for Embedding incoming queries and performing simalirty search to retrieve relevant documenation. This can be passed to an LLM node for further processing/inference.
Choose your Knowledge Base
From the dropdown select the appropriate KB
Input Query
This can be paramaterized (for example if you want to pass in a query from your front end:)
- Denote Variable: < ?input_query? >
- Input Query will be embedded and semantic search will take place
- Relevant documents returned
Pass to LLM Node
Pipe doc information to LLM node (ensure you are reffering the RAG Node ID: < ?rag_node_id? >)
Rag Implementation