agents.rag.agentic.agentic_rag_agent¶

Agentic RAG Multi-Agent System.

This implements an advanced RAG system with document grading, query rewriting, and conditional routing between retrieval and web search.

Classes¶

`AgenticRAGAgent`	Advanced Agentic RAG system with document grading and query refinement.
`AgenticRAGState`	State for the Agentic RAG workflow.

Module Contents¶

class agents.rag.agentic.agentic_rag_agent.AgenticRAGAgent¶

Bases: haive.agents.simple.SimpleAgent

Advanced Agentic RAG system with document grading and query refinement.

This agent implements a sophisticated RAG workflow that includes: 1. Initial retrieval from vector store 2. Document relevance grading 3. Query rewriting if documents aren’t relevant 4. Web search fallback 5. Final answer generation

Example

# Create agentic RAG agent
agent = AgenticRAGAgent.create_default(
name="agentic_rag",
retriever_config=vector_store_config,
use_web_search=True
)

# Process a query
result = await agent.arun("What are the latest advances in quantum computing?")

# The agent will:
# 1. Retrieve documents from vector store
# 2. Grade them for relevance
# 3. Rewrite query if needed
# 4. Use web search if local docs aren't sufficient
# 5. Generate comprehensive answer

build_graph()¶

Build the Agentic RAG workflow graph.

Return type:: haive.core.graph.state_graph.base_graph2.BaseGraph

classmethod create_default(**kwargs)¶

Create a default Agentic RAG agent.

Parameters:: **kwargs – Configuration options - name: Agent name - retriever_config: Retriever or vector store config (required) - use_web_search: Whether to enable web search fallback - temperature: LLM temperature - engine: Custom AugLLMConfig if needed
Returns:: AgenticRAGAgent configured for advanced RAG
Return type:: AgenticRAGAgent

class agents.rag.agentic.agentic_rag_agent.AgenticRAGState(/, **data)¶

Bases: haive.core.schema.prebuilt.messages_state.MessagesState

State for the Agentic RAG workflow.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:: data (Any)