agents.memory_v2.kg_memory_agent¶

Knowledge Graph Memory Agent with Graph Database Integration.

This agent extends the existing KG transformer capabilities with: 1. Graph database upload and storage (Neo4j, Neptune, etc.) 2. Memory-specific knowledge graph construction 3. Time-weighted graph retrieval 4. Configurable storage backends

Based on existing ParallelKGTransformer but optimized for memory workflows.

Classes¶

`GraphDatabaseConnector`	Abstract connector for graph databases.
`GraphStorageBackend`	Supported graph database backends.
`KGMemoryAgent`	Knowledge Graph Memory Agent with database integration.
`KGMemoryConfig`	Configuration for KG Memory Agent.

Functions¶

`create_memory_kg_agent`([storage_backend, llm_config])	Factory function to create KG Memory Agent.
`create_neo4j_memory_agent`(uri, username, password[, ...])	Create KG Memory Agent with Neo4j backend.

Module Contents¶

class agents.memory_v2.kg_memory_agent.GraphDatabaseConnector(config)¶

Abstract connector for graph databases.

Initialize connector with configuration.

Parameters:: config (KGMemoryConfig)

async close()¶

Close database connection.

Return type:: None

async connect()¶

Connect to graph database.

Return type:: None

async retrieve_graph(graph_id)¶

Retrieve knowledge graph by ID.

Parameters:: graph_id (str)
Return type:: tuple[haive.agents.document_modifiers.kg.kg_map_merge.models.KnowledgeGraph, dict[str, Any]] | None

async store_knowledge_graph(graph, graph_id, metadata=None)¶

Store knowledge graph in configured backend.

Parameters:

graph (haive.agents.document_modifiers.kg.kg_map_merge.models.KnowledgeGraph)
graph_id (str)
metadata (dict[str, Any] | None)

Return type:

bool

class agents.memory_v2.kg_memory_agent.GraphStorageBackend¶

Bases: str, enum.Enum

Supported graph database backends.

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

class agents.memory_v2.kg_memory_agent.KGMemoryAgent(config)¶

Knowledge Graph Memory Agent with database integration.

Initialize KG Memory Agent.

Parameters:: config (KGMemoryConfig)

async close()¶

Close agent and database connections.

Return type:: None

async process_conversation_to_graph(messages, graph_id=None)¶

Process conversation messages into knowledge graph.

Parameters:

messages (list[langchain_core.messages.BaseMessage]) – Conversation messages
graph_id (str | None) – Optional graph ID

Returns:

Tuple of (graph_id, knowledge_graph)

Return type:

tuple[str, haive.agents.document_modifiers.kg.kg_map_merge.models.KnowledgeGraph]

async process_memories_to_graph(memories, graph_id=None)¶

Process memories into knowledge graph and store.

Parameters:

memories (list[agents.memory_v2.memory_state_original.EnhancedMemoryItem]) – List of memory items to process
graph_id (str | None) – Optional graph ID (generated if not provided)

Returns:

Tuple of (graph_id, knowledge_graph)

Return type:

tuple[str, haive.agents.document_modifiers.kg.kg_map_merge.models.KnowledgeGraph]

async query_graph_by_entity(entity_name)¶

Query graph database for entity and its relationships.

Parameters:: entity_name (str) – Name of entity to query
Returns:: List of related entities and relationships
Return type:: list[dict[str, Any]]

async retrieve_memory_graph(graph_id)¶

Retrieve stored knowledge graph.

Parameters:: graph_id (str) – ID of graph to retrieve
Returns:: KnowledgeGraph if found, None otherwise
Return type:: haive.agents.document_modifiers.kg.kg_map_merge.models.KnowledgeGraph | None

async setup()¶

Setup agent and connections.

Return type:: None

class agents.memory_v2.kg_memory_agent.KGMemoryConfig(/, **data)¶

Bases: pydantic.BaseModel

Configuration for KG Memory 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)

model_config¶: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

agents.memory_v2.kg_memory_agent.create_memory_kg_agent(storage_backend='memory', llm_config=None, **storage_kwargs)¶

Factory function to create KG Memory Agent.

Parameters:

storage_backend (str) – “memory”, “neo4j”, “file”
llm_config (haive.core.engine.aug_llm.AugLLMConfig) – LLM configuration
**storage_kwargs – Backend-specific settings

Returns:

Configured KGMemoryAgent

Return type:

KGMemoryAgent

agents.memory_v2.kg_memory_agent.create_neo4j_memory_agent(uri, username, password, database='neo4j', llm_config=None)¶

Create KG Memory Agent with Neo4j backend.

Parameters:

uri (str) – Neo4j connection URI
username (str) – Database username
password (str) – Database password
database (str) – Database name
llm_config (haive.core.engine.aug_llm.AugLLMConfig) – LLM configuration

Returns:

KGMemoryAgent configured for Neo4j

Return type:

KGMemoryAgent