agents.rag.hyde.enhanced_agent_v2¶

Enhanced HyDE RAG Agent v2 with Advanced Prompt Selection and Multi-Document Generation.

This version integrates the new enhanced prompt system with: - Automatic prompt type selection based on query analysis - Multi-document generation from different perspectives - Improved separation of generation from parsing - Domain-specific prompt templates - Ensemble retrieval using multiple hypothetical documents

Classes¶

`AdaptiveHyDEGenerator`	Generator that adapts its prompt based on query analysis.
`DomainAnalysisAgent`	Analyzes queries to determine relevant domains for multi-domain generation.
`EnhancedHyDERAGAgentV2`	Enhanced HyDE RAG Agent with advanced prompt selection and multi-document generation.
`EnhancedHyDERetrieverV2`	Enhanced retriever with better state handling and fallback mechanisms.
`EnsembleDocumentParser`	Parses ensemble document output into individual documents.
`EnsembleHyDERetriever`	Retriever that handles multiple documents for ensemble retrieval.
`HyDEAgentConfig`	Configuration for Enhanced HyDE RAG Agent.
`HyDEDocumentAnalyzer`	Analyzes generated hypothetical documents and extracts structured information.
`HyDEGenerationMode`	Different modes for HyDE document generation.
`MultiDomainHyDERetriever`	Retriever that handles documents from multiple domains.
`QueryAnalysisAgent`	Agent that analyzes queries and selects appropriate prompt types.

Functions¶

`create_enhanced_hyde_v2`(documents[, llm_config, ...])	Create Enhanced HyDE RAG Agent v2 with specified configuration.
`create_ensemble_hyde`(documents[, num_docs, llm_config])	Create HyDE agent with ensemble document generation.
`create_multi_perspective_hyde`(documents, perspectives)	Create HyDE agent with multi-perspective generation.

Module Contents¶

class agents.rag.hyde.enhanced_agent_v2.AdaptiveHyDEGenerator(llm_config, **kwargs)¶

Bases: haive.agents.simple.agent.SimpleAgent

Generator that adapts its prompt based on query analysis.

Init .

Parameters:: llm_config (haive.core.models.llm.base.LLMConfig) – [TODO: Add description]

run(input_data)¶

Generate document using adaptively selected prompt.

Parameters:: input_data (dict[str, Any])
Return type:: dict[str, Any]

class agents.rag.hyde.enhanced_agent_v2.DomainAnalysisAgent(llm_config, **kwargs)¶

Bases: haive.agents.simple.agent.SimpleAgent

Analyzes queries to determine relevant domains for multi-domain generation.

Init .

Parameters:: llm_config (haive.core.models.llm.base.LLMConfig) – [TODO: Add description]

class agents.rag.hyde.enhanced_agent_v2.EnhancedHyDERAGAgentV2(/, **data)¶

Bases: haive.agents.multi.enhanced_sequential_agent.SequentialAgent, haive.core.common.mixins.tool_route_mixin.ToolRouteMixin

Enhanced HyDE RAG Agent with advanced prompt selection and multi-document generation.

Key Features: - Automatic prompt type selection based on query analysis - Multi-document generation from different perspectives/domains - Ensemble retrieval using multiple hypothetical documents - Proper separation of generation from parsing - Configurable generation strategies - Enhanced error handling and fallback mechanisms

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)

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

Create Enhanced HyDE RAG Agent v2 from documents.

Parameters:

documents (list[langchain_core.documents.Document]) – Documents to index for retrieval
llm_config (haive.core.models.llm.base.LLMConfig | None) – LLM configuration
embedding_model (str | None) – Optional embedding model
config (HyDEAgentConfig | None) – HyDE agent configuration
**kwargs – Additional arguments

Returns:

Configured Enhanced HyDE RAG Agent v2

Return type:

EnhancedHyDERAGAgentV2

setup_hyde_agent()¶

Setup HyDE agent with enhanced prompts.

Return type:: EnhancedHyDERAGAgentV2

class agents.rag.hyde.enhanced_agent_v2.EnhancedHyDERetrieverV2¶

Bases: haive.agents.base.agent.Agent

Enhanced retriever with better state handling and fallback mechanisms.

build_graph()¶

Build Graph.

Returns:: Add return description]
Return type:: [TODO

class agents.rag.hyde.enhanced_agent_v2.EnsembleDocumentParser(llm_config, **kwargs)¶

Bases: haive.agents.simple.agent.SimpleAgent

Parses ensemble document output into individual documents.

Init .

Parameters:: llm_config (haive.core.models.llm.base.LLMConfig) – [TODO: Add description]

class agents.rag.hyde.enhanced_agent_v2.EnsembleHyDERetriever¶

Bases: haive.agents.base.agent.Agent

Retriever that handles multiple documents for ensemble retrieval.

build_graph()¶

Build Graph.

Returns:: Add return description]
Return type:: [TODO

class agents.rag.hyde.enhanced_agent_v2.HyDEAgentConfig(/, **data)¶

Bases: pydantic.BaseModel

Configuration for Enhanced HyDE RAG Agent.

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.hyde.enhanced_agent_v2.HyDEDocumentAnalyzer(llm_config, **kwargs)¶

Bases: haive.agents.simple.agent.SimpleAgent

Analyzes generated hypothetical documents and extracts structured information.

Init .

Parameters:: llm_config (haive.core.models.llm.base.LLMConfig) – [TODO: Add description]

class agents.rag.hyde.enhanced_agent_v2.HyDEGenerationMode¶

Bases: str, enum.Enum

Different modes for HyDE document generation.

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

class agents.rag.hyde.enhanced_agent_v2.MultiDomainHyDERetriever¶

Bases: haive.agents.base.agent.Agent

Retriever that handles documents from multiple domains.

build_graph()¶

Build Graph.

Returns:: Add return description]
Return type:: [TODO

class agents.rag.hyde.enhanced_agent_v2.QueryAnalysisAgent(llm_config, **kwargs)¶

Bases: haive.agents.simple.agent.SimpleAgent

Agent that analyzes queries and selects appropriate prompt types.

Init .

Parameters:: llm_config (haive.core.models.llm.base.LLMConfig) – [TODO: Add description]

run(input_data)¶

Run query analysis and add prompt type selection.

Parameters:: input_data (dict[str, Any])
Return type:: dict[str, Any]

agents.rag.hyde.enhanced_agent_v2.create_enhanced_hyde_v2(documents, llm_config=None, generation_mode=HyDEGenerationMode.SINGLE, auto_select_prompt=True, **kwargs)¶

Create Enhanced HyDE RAG Agent v2 with specified configuration.

Parameters:

documents (list[langchain_core.documents.Document]) – Documents for retrieval
llm_config (haive.core.models.llm.base.LLMConfig | None) – LLM configuration
generation_mode (HyDEGenerationMode) – Mode for document generation
auto_select_prompt (bool) – Whether to auto-select prompt types
**kwargs – Additional configuration options

Returns:

Configured Enhanced HyDE RAG Agent v2

Return type:

EnhancedHyDERAGAgentV2

agents.rag.hyde.enhanced_agent_v2.create_ensemble_hyde(documents, num_docs=3, llm_config=None, **kwargs)¶

Create HyDE agent with ensemble document generation.

Parameters:

documents (list[langchain_core.documents.Document]) – Documents for retrieval
num_docs (int) – Number of documents to generate
llm_config (haive.core.models.llm.base.LLMConfig | None) – LLM configuration
**kwargs – Additional options

Returns:

Ensemble HyDE agent

Return type:

EnhancedHyDERAGAgentV2

agents.rag.hyde.enhanced_agent_v2.create_multi_perspective_hyde(documents, perspectives, llm_config=None, **kwargs)¶

Create HyDE agent with multi-perspective generation.

Parameters:

documents (list[langchain_core.documents.Document]) – Documents for retrieval
perspectives (list[haive.agents.rag.common.query_constructors.hyde.enhanced_prompts.HyDEPerspective]) – List of perspectives to use
llm_config (haive.core.models.llm.base.LLMConfig | None) – LLM configuration
**kwargs – Additional options

Returns:

Multi-perspective HyDE agent

Return type:

EnhancedHyDERAGAgentV2