haive.agents.rag.adaptive_tools.agent

Adaptive RAG with Tools Integration Agents.

from typing import Any Implementation of adaptive RAG with tool integration and ReAct patterns. Includes Google Search integration, tool selection, and dynamic routing based on query needs.

Classes

AdaptiveToolsRAGAgent	Complete Adaptive RAG agent with tools integration and ReAct patterns.
AdaptiveToolsResult	Complete result from adaptive tools RAG.
QueryNeed	Query need categories for tool selection.
SearchIntegrationAgent	Agent that integrates external search tools.
SearchResult	Results from search tools.
ToolSelection	Tool selection analysis and recommendations.
ToolSelectionAgent	Agent that selects optimal tools based on query analysis.
ToolType	Available tool types for adaptive RAG.

Functions

create_adaptive_synthesis_callable(llm_config)	Create callable function for adaptive synthesis.
create_adaptive_tools_rag_agent(documents[, ...])	Create an Adaptive Tools RAG agent.
create_google_search_callable(llm_config)	Create callable function for Google search integration.
create_tool_selector_callable(llm_config)	Create callable function for tool selection.
get_adaptive_tools_rag_io_schema()	Get I/O schema for Adaptive Tools RAG agents.

Module Contents

class haive.agents.rag.adaptive_tools.agent.AdaptiveToolsRAGAgent

Bases: haive.agents.multi.base.SequentialAgent

Complete Adaptive RAG agent with tools integration and ReAct patterns.

classmethod from_documents(documents, llm_config=None, enable_google_search=True, enable_local_retrieval=True, **kwargs)

Create Adaptive Tools RAG agent from documents.

Parameters:

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

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

• enable_google_search (bool) – Whether to enable Google Search integration

• enable_local_retrieval (bool) – Whether to enable local document retrieval

• **kwargs – Additional arguments

Returns:

AdaptiveToolsRAGAgent instance

class haive.agents.rag.adaptive_tools.agent.AdaptiveToolsResult(/, **data)

Bases: pydantic.BaseModel

Complete result from adaptive tools RAG.

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 haive.agents.rag.adaptive_tools.agent.QueryNeed

Bases: str, enum.Enum

Query need categories for tool selection.

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

class haive.agents.rag.adaptive_tools.agent.SearchIntegrationAgent

Bases: haive.agents.base.agent.Agent

Agent that integrates external search tools.

build_graph()

Build search integration graph.

Return type:

haive.core.graph.state_graph.base_graph2.BaseGraph

class haive.agents.rag.adaptive_tools.agent.SearchResult(/, **data)

Bases: pydantic.BaseModel

Results from search tools.

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 haive.agents.rag.adaptive_tools.agent.ToolSelection(/, **data)

Bases: pydantic.BaseModel

Tool selection analysis and recommendations.

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 haive.agents.rag.adaptive_tools.agent.ToolSelectionAgent

Bases: haive.agents.base.agent.Agent

Agent that selects optimal tools based on query analysis.

build_graph()

Build tool selection graph.

Return type:

haive.core.graph.state_graph.base_graph2.BaseGraph

class haive.agents.rag.adaptive_tools.agent.ToolType

Bases: str, enum.Enum

Available tool types for adaptive RAG.

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

haive.agents.rag.adaptive_tools.agent.create_adaptive_synthesis_callable(llm_config)

Create callable function for adaptive synthesis.

Parameters:

llm_config (haive.core.models.llm.base.LLMConfig)

haive.agents.rag.adaptive_tools.agent.create_adaptive_tools_rag_agent(documents, llm_config=None, tools_mode='full', **kwargs)

Create an Adaptive Tools RAG agent.

Parameters:

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

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

• tools_mode (str) – Tools mode (“full”, “search_only”, “local_only”)

• **kwargs – Additional arguments

Returns:

Configured Adaptive Tools RAG agent

Return type:

AdaptiveToolsRAGAgent

haive.agents.rag.adaptive_tools.agent.create_google_search_callable(llm_config)

Create callable function for Google search integration.

Parameters:

llm_config (haive.core.models.llm.base.LLMConfig)

haive.agents.rag.adaptive_tools.agent.create_tool_selector_callable(llm_config)

Create callable function for tool selection.

Parameters:

llm_config (haive.core.models.llm.base.LLMConfig)

haive.agents.rag.adaptive_tools.agent.get_adaptive_tools_rag_io_schema()

Get I/O schema for Adaptive Tools RAG agents.

Return type:

dict[str, list[str]]