agents.memory_v2.integrated_memory_system

Integrated Memory System combining Graph, Vector, and Time-based memory.

This system shows how to use multiple memory strategies together: 1. GraphMemoryAgent for structured knowledge and relationships 2. ReactMemoryAgent for flexible tool-based memory management 3. LongTermMemoryAgent for persistent cross-conversation memory

Classes

IntegratedMemorySystem

Combines multiple memory systems for comprehensive memory management.

MemorySystemMode

Modes for the integrated memory system.

Functions

create_research_assistant()

Create a research assistant with integrated memory.

demo_integrated_memory()

Demonstrate the integrated memory system.

Module Contents

class agents.memory_v2.integrated_memory_system.IntegratedMemorySystem(user_id='default_user', neo4j_config=None, vector_store_path=None, engine=None)

Combines multiple memory systems for comprehensive memory management.

This system intelligently routes memory operations to the most appropriate subsystem based on content type and requirements.

Init .

Parameters:

  • user_id (str) – [TODO: Add description]

  • neo4j_config (dict[str, Any] | None) – [TODO: Add description]

  • vector_store_path (str | None) – [TODO: Add description]

  • engine (haive.core.engine.aug_llm.AugLLMConfig | None) – [TODO: Add description]

async consolidate_all_memories()

Consolidate memories across all systems.

Return type:

dict[str, Any]

async get_memory_analytics()

Get analytics across all memory systems.

Return type:

dict[str, Any]

async query_memory(query, mode=MemorySystemMode.INTELLIGENT, combine_results=True)

Query memory using appropriate system(s).

Parameters:

  • query (str) – Query string

  • mode (MemorySystemMode) – Query mode

  • combine_results (bool) – Whether to combine results from multiple systems

Returns:

Query results

Return type:

dict[str, Any]

async store_memory(content, mode=MemorySystemMode.INTELLIGENT, metadata=None)

Store memory using the appropriate system(s).

Parameters:

  • content (str) – Memory content to store

  • mode (MemorySystemMode) – Storage mode

  • metadata (dict[str, Any] | None) – Optional metadata

Returns:

Storage results from all used systems

Return type:

dict[str, Any]

class agents.memory_v2.integrated_memory_system.MemorySystemMode

Bases: str, enum.Enum

Modes for the integrated memory system.

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

async agents.memory_v2.integrated_memory_system.create_research_assistant()

Create a research assistant with integrated memory.

async agents.memory_v2.integrated_memory_system.demo_integrated_memory()

Demonstrate the integrated memory system.