self_query_mcp_agent

Self Query MCP Discovery Agent.

Enhanced RAG agent that uses Self Query methodology for structured querying with metadata filtering and parent document retrieval.

Classes

EnhancedMCPDocument

Enhanced document creation with chunking for parent retrieval.

MCPServerMetadata

Metadata schema for MCP servers.

SelfQueryMCPAgent

Enhanced MCP Discovery Agent with Self Query and Parent Document Retrieval.

Functions

test_enhanced_agent()

Test the enhanced MCP agent.

Module Contents

class self_query_mcp_agent.EnhancedMCPDocument

Enhanced document creation with chunking for parent retrieval.

static create_mcp_documents_with_chunks() → tuple[list[langchain.schema.Document], list[langchain.schema.Document]]

Create both parent documents and child chunks for retrieval.

class self_query_mcp_agent.MCPServerMetadata(/, **data: Any)

Bases: pydantic.BaseModel

Metadata schema for MCP servers.

category: str = None
has_install: bool = None
language: str = None
prompts_count: int = None
resources_count: int = None
server_name: str = None
stars: int = None
tools_count: int = None
total_features: int = None
class self_query_mcp_agent.SelfQueryMCPAgent(config: haive.core.engine.aug_llm.AugLLMConfig = None)

Enhanced MCP Discovery Agent with Self Query and Parent Document Retrieval.

analyze_query_intent(query: str) → str

Analyze query to determine best search method.

Perform all search methods and return results.

async search_similarity(query: str, k: int = 5) → list[langchain.schema.Document]

Fallback similarity search.

async search_with_parent_retriever(query: str, k: int = 5) → list[langchain.schema.Document]

Search using parent document retriever for full content.

async search_with_self_query(query: str, k: int = 5) → list[langchain.schema.Document]

Search using self-query retriever for structured queries.

setup_retrievers()

Set up both self-query and parent document retrievers.

config
embeddings
llm
async self_query_mcp_agent.test_enhanced_agent()

Test the enhanced MCP agent.