Memory system for conversation history management.

Provides conversation memory management functionality.

Core Concepts

• BaseMemory: Base trait for memory.
• ConversationBufferMemory: Simple conversation buffer.
• ConversationBufferWindowMemory: Conversation buffer with window.
• ConversationSummaryMemory: LLM-based summary compression.
• ConversationSummaryBufferMemory: Hybrid compression (summary + recent messages).
• PersistentMemory: Trait for persistent memory storage.
• ContextWindow: Fits messages within a token budget.

Feature Flags

• mongodb-persistence — Enables MongoPersistentMemory (MongoDB-backed persistent memory).
• vectorstore-memory — Enables VectorStoreRetrieverMemory (vector-store-backed memory).