Struct rig::rag::RagAgent

pub struct RagAgent<M: CompletionModel, C: VectorStoreIndex, T: VectorStoreIndex> {
    pub tools: ToolSet,
    /* private fields */
}

Struct representing a RAG agent, i.e.: an agent enhanced with two collections of vector store indices, one for context documents and one for tools. The ragged context and tools are used to enhance the completion model at prompt-time. Note: The type of the VectorStoreIndex must be the same for all the dynamic context and tools indices (but can be different for context and tools). If you need to use a more complex combination of vector store indices, you should implement a custom agent.

Fields

tools: ToolSet

Actual tool implementations

Implementations

impl<M: CompletionModel, C: VectorStoreIndex, T: VectorStoreIndex> RagAgent<M, C, T>

pub async fn call_tool( &self, toolname: &str, args: &str, ) -> Result<String, ToolSetError>

Trait Implementations

impl<M: CompletionModel, C: VectorStoreIndex, T: VectorStoreIndex> Chat for RagAgent<M, C, T>

async fn chat( &self, prompt: &str, chat_history: Vec<Message>, ) -> Result<String, PromptError>

Send a prompt with optional chat history to the underlying completion model.

impl<M: CompletionModel, C: VectorStoreIndex, T: VectorStoreIndex> Completion<M> for RagAgent<M, C, T>

async fn completion( &self, prompt: &str, chat_history: Vec<Message>, ) -> Result<CompletionRequestBuilder<M>, CompletionError>

Generates a completion request builder for the given prompt and chat_history. This function is meant to be called by the user to further customize the request at prompt time before sending it.

impl<M: CompletionModel, C: VectorStoreIndex, T: VectorStoreIndex> Prompt for RagAgent<M, C, T>

async fn prompt(&self, prompt: &str) -> Result<String, PromptError>

Send a simple prompt to the underlying completion model.