agents.memory_v2.memory_state¶

Memory state models for Memory V2 system using original Haive memory models.

This module integrates the proven memory models from haive.agents.memory.models and haive.agents.ltm.memory_schemas with our V2 enhancements for token tracking, graph integration, and advanced memory management.

Classes¶

`MemoryEntry`	A single memory entry with content and metadata.
`MemoryMetadata`	Metadata for a single memory entry.
`MemoryState`	State schema for memory agent operations.
`MemoryStats`	Statistics about memory operations and performance.

Module Contents¶

class agents.memory_v2.memory_state.MemoryEntry(/, **data)¶

Bases: pydantic.BaseModel

A single memory entry with content and metadata.

Represents a complete memory item that can be stored, retrieved, and analyzed by memory agents.

Parameters:: data (Any)

id¶: Unique identifier for the memory

content¶: The actual memory content

metadata¶: Structured metadata about the memory

embedding¶: Optional vector embedding for similarity search

similarity_score¶: Similarity score when retrieved (populated during retrieval)

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.

class agents.memory_v2.memory_state.MemoryMetadata(/, **data)¶

Bases: pydantic.BaseModel

Metadata for a single memory entry.

Tracks essential information about stored memories including type classification, importance, timestamps, and source information.

Parameters:: data (Any)

memory_type¶: Type of memory (semantic, episodic, procedural)

importance¶: Importance level (critical, high, medium, low, transient)

confidence¶: Confidence score for the memory (0.0-1.0)

timestamp¶: When the memory was created

source¶: Source of the memory (user_input, agent_inference, system)

tags¶: List of tags for categorization

entities¶: Named entities mentioned in the memory

relationships¶: Relationships extracted from the memory

context_id¶: ID linking related memories

retrieval_count¶: How many times this memory has been retrieved

last_accessed¶: When the memory was last accessed

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.

class agents.memory_v2.memory_state.MemoryState(/, **data)¶

Bases: haive.core.schema.prebuilt.messages_state.MessagesState

State schema for memory agent operations.

Extends MessagesState with memory-specific fields for tracking current memories, metadata, statistics, and operation results.

This state schema is used by all memory agents to maintain consistent state management and enable proper coordination between different memory strategies.

Parameters:: data (Any)

current_memories¶: List of memory entries currently being processed

retrieved_memories¶: List of memories retrieved in the last operation

memory_metadata¶: General metadata about the memory session

memory_stats¶: Performance and usage statistics

token_usage¶: Token usage tracking for memory operations

last_operation¶: Information about the last memory operation performed

memory_context¶: Context information for memory operations

active_filters¶: Currently active filters for memory search/retrieval

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.

add_memory(memory)¶

Add a memory entry to current memories.

Parameters:: memory (MemoryEntry)
Return type:: None

get_memory_summary()¶

Get a comprehensive summary of the current memory state.

Return type:: dict[str, Any]

update_retrieval_stats(memories, retrieval_time)¶

Update statistics after memory retrieval.

Parameters:

memories (list[MemoryEntry])
retrieval_time (float)

Return type:

None

class agents.memory_v2.memory_state.MemoryStats(/, **data)¶

Bases: pydantic.BaseModel

Statistics about memory operations and performance.

Tracks memory system performance, usage patterns, and optimization metrics.

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)