haive.agents.rag.step_back.agent
================================

.. py:module:: haive.agents.rag.step_back.agent

.. autoapi-nested-parse::

   Step-Back Prompting RAG Agents.

   from typing import Any
   Implementation of step-back prompting for abstract reasoning.
   Generates broader conceptual queries for enhanced context retrieval.



Classes
-------

.. autoapisummary::

   haive.agents.rag.step_back.agent.DualRetrievalAgent
   haive.agents.rag.step_back.agent.StepBackQuery
   haive.agents.rag.step_back.agent.StepBackQueryGeneratorAgent
   haive.agents.rag.step_back.agent.StepBackRAGAgent
   haive.agents.rag.step_back.agent.StepBackResult


Functions
---------

.. autoapisummary::

   haive.agents.rag.step_back.agent.create_step_back_rag_agent
   haive.agents.rag.step_back.agent.get_step_back_rag_io_schema


Module Contents
---------------

.. py:class:: DualRetrievalAgent

   Bases: :py:obj:`haive.agents.base.agent.Agent`


   Agent that performs both original and step-back retrieval.


   .. py:method:: build_graph()

      Build dual retrieval graph.



   .. py:method:: model_post_init(__context)

      Initialize the mixin with state tracking attributes after Pydantic validation.



.. py:class:: StepBackQuery(/, **data)

   Bases: :py:obj:`pydantic.BaseModel`


   Step-back query generation result.

   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.


.. py:class:: StepBackQueryGeneratorAgent

   Bases: :py:obj:`haive.agents.base.agent.Agent`


   Agent that generates step-back queries for abstract reasoning.


   .. py:method:: build_graph()

      Build step-back query generation graph.



.. py:class:: StepBackRAGAgent

   Bases: :py:obj:`haive.agents.multi.base.SequentialAgent`


   Complete Step-Back RAG agent with abstract reasoning.


   .. py:method:: from_documents(documents, llm_config = None, embedding_model = None, abstraction_level = 'moderate', max_docs_each = 5, **kwargs)
      :classmethod:


      Create Step-Back RAG agent from documents.

      :param documents: Documents to index
      :param llm_config: LLM configuration
      :param embedding_model: Embedding model for retrieval
      :param abstraction_level: Level of abstraction for step-back queries
      :param max_docs_each: Max docs to retrieve for each query type
      :param \*\*kwargs: Additional arguments

      :returns: StepBackRAGAgent instance



.. py:class:: StepBackResult(/, **data)

   Bases: :py:obj:`pydantic.BaseModel`


   Combined results from step-back retrieval.

   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.


.. py:function:: create_step_back_rag_agent(documents, llm_config = None, reasoning_depth = 'moderate', **kwargs)

   Create a Step-Back RAG agent.

   :param documents: Documents for retrieval
   :param llm_config: LLM configuration
   :param reasoning_depth: Depth of reasoning ("shallow", "moderate", "deep")
   :param \*\*kwargs: Additional arguments

   :returns: Configured Step-Back RAG agent


.. py:function:: get_step_back_rag_io_schema()

   Get I/O schema for Step-Back RAG agents.