haive.agents.rag.multi_query.agent

Multi-Query RAG Agent.

Improves recall through query diversification. Generates multiple query variations and retrieves from all.

Classes

MultiQueryRAGAgent	Multi-Query RAG with query expansion for improved recall.
MultiRetrievalAgent	Agent that performs parallel retrieval with multiple queries.
QueryVariations	Structured output for query variations.

Module Contents

class haive.agents.rag.multi_query.agent.MultiQueryRAGAgent

Bases: haive.agents.multi.base.SequentialAgent

Multi-Query RAG with query expansion for improved recall.

classmethod from_documents(documents, llm_config=None, embedding_model=None, **kwargs)

Create Multi-Query RAG from documents.

Parameters:

• documents (list[langchain_core.documents.Document]) – Documents to index

• llm_config (haive.core.models.llm.base.LLMConfig | None) – Optional LLM configuration

• embedding_model (str | None) – Optional embedding model for vector store

• **kwargs – Additional arguments

Returns:

MultiQueryRAGAgent instance

class haive.agents.rag.multi_query.agent.MultiRetrievalAgent

Bases: haive.agents.base.agent.Agent

Agent that performs parallel retrieval with multiple queries.

build_graph()

Build graph that retrieves with multiple queries in parallel.

Return type:

haive.core.graph.state_graph.base_graph2.BaseGraph

class haive.agents.rag.multi_query.agent.QueryVariations(/, **data)

Bases: pydantic.BaseModel

Structured output for query variations.

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)