agents.rag.self_route.agent¶

Self-Route RAG Agents.

from typing import Any Implementation of self-routing RAG with dynamic strategy selection and iterative planning. Uses structured output models for complex routing decisions and preprocessing.

Classes¶

IterativePlan

Iterative processing plan with loop structure.

IterativePlannerAgent

Agent that creates iterative processing plans.

QueryAnalysis

Structured analysis of query for routing decisions.

QueryAnalyzerAgent

Agent that performs structured query analysis for routing.

QueryComplexity

Query complexity levels for routing decisions.

RoutingDecision

Final routing decision with execution plan.

RoutingDecisionAgent

Agent that makes final routing decisions.

RoutingStrategy

Available routing strategies.

SelfRouteRAGAgent

Complete Self-Route RAG agent with structured analysis and iterative planning.

Functions¶

create_self_route_rag_agent(documents[, llm_config, ...])

Create a Self-Route RAG agent.

get_self_route_rag_io_schema()

Get I/O schema for Self-Route RAG agents.

Module Contents¶

class agents.rag.self_route.agent.IterativePlan(/, **data)¶

Bases: pydantic.BaseModel

Iterative processing plan with loop structure.

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)

class agents.rag.self_route.agent.IterativePlannerAgent(llm_config=None, max_iterations=3, **kwargs)¶

Bases: haive.agents.base.agent.Agent

Agent that creates iterative processing plans.

Initialize iterative planner.

Parameters:

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

  • max_iterations (int) – Maximum number of iterations

  • **kwargs – Additional agent arguments

build_graph()¶

Build iterative planning graph.

Return type:

haive.core.graph.state_graph.base_graph2.BaseGraph

class agents.rag.self_route.agent.QueryAnalysis(/, **data)¶

Bases: pydantic.BaseModel

Structured analysis of query for routing decisions.

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)

class agents.rag.self_route.agent.QueryAnalyzerAgent(llm_config=None, analysis_depth='comprehensive', **kwargs)¶

Bases: haive.agents.base.agent.Agent

Agent that performs structured query analysis for routing.

Initialize query analyzer.

Parameters:

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

  • analysis_depth (str) – Depth of analysis (“basic”, “comprehensive”, “expert”)

  • **kwargs – Additional agent arguments

build_graph()¶

Build query analysis graph.

Return type:

haive.core.graph.state_graph.base_graph2.BaseGraph

class agents.rag.self_route.agent.QueryComplexity¶

Bases: str, enum.Enum

Query complexity levels for routing decisions.

Initialize self. See help(type(self)) for accurate signature.

class agents.rag.self_route.agent.RoutingDecision(/, **data)¶

Bases: pydantic.BaseModel

Final routing decision with execution plan.

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)

class agents.rag.self_route.agent.RoutingDecisionAgent(llm_config=None, enable_fallback=True, **kwargs)¶

Bases: haive.agents.base.agent.Agent

Agent that makes final routing decisions.

Initialize routing decision agent.

Parameters:

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

  • enable_fallback (bool) – Whether to enable fallback strategies

  • **kwargs – Additional agent arguments

build_graph()¶

Build routing decision graph.

Return type:

haive.core.graph.state_graph.base_graph2.BaseGraph

class agents.rag.self_route.agent.RoutingStrategy¶

Bases: str, enum.Enum

Available routing strategies.

Initialize self. See help(type(self)) for accurate signature.

class agents.rag.self_route.agent.SelfRouteRAGAgent(config)¶

Bases: haive.agents.multi.base.SequentialAgent

Complete Self-Route RAG agent with structured analysis and iterative planning.

Init .

Parameters:

config (SequentialAgentConfig) – [TODO: Add description]

classmethod from_documents(documents, llm_config=None, analysis_depth='comprehensive', max_iterations=3, enable_fallback=True, **kwargs)¶

Create Self-Route RAG agent from documents.

Parameters:

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

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

  • analysis_depth (str) – Depth of query analysis

  • max_iterations (int) – Maximum iterations for planning

  • enable_fallback (bool) – Whether to enable fallback routing

  • **kwargs – Additional arguments

Returns:

SelfRouteRAGAgent instance

agents.rag.self_route.agent.create_self_route_rag_agent(documents, llm_config=None, routing_mode='adaptive', **kwargs)¶

Create a Self-Route RAG agent.

Parameters:

  • documents (list[langchain_core.documents.Document]) – Documents for retrieval

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

  • routing_mode (str) – Routing mode (“conservative”, “adaptive”, “aggressive”)

  • **kwargs – Additional arguments

Returns:

Configured Self-Route RAG agent

Return type:

SelfRouteRAGAgent

agents.rag.self_route.agent.get_self_route_rag_io_schema()¶

Get I/O schema for Self-Route RAG agents.

Return type:

dict[str, list[str]]