File: base_engine.py

Type: CORE

Module: memorybolt/engines/

Overview:

base_engine.py defines the abstract interface that all memory engines in the MemoryBolt system must implement.

It enforces a unified memory contract for reading, writing, deleting, and querying memory objects — whether stored locally, in Redis, Postgres, S3, vector DBs, or hybrid formats.

By centralizing this contract, MemoryBolt ensures all engines can be swapped, composed, routed, or federated across multiple backends with zero breaking change.

This file is a critical spine of the architecture — designed for extension, not modification.

Key Concepts:

• BaseMemoryEngine is an abstract base class.
• All engines must inherit from it.
• Enforces the 5-core-method contract: .store(), .load(), .delete(), .query(), .healthcheck().
• Built-in support for capability introspection, so higher layers can route based on feature support (e.g., vector search, streaming, sharding).
• Designed for minimal overhead and clean subclassing.
• No hard dependencies — zero imports beyond abc, typing, and dataclasses unless needed.

Use Cases:

• Used by all memory engines (local, Redis, PG, vector, S3, semantic).
• Swappable via config.yaml → memory.engine.
• Called internally by the memory router (memory_core.py) and plugin loader.
• Enables dynamic orchestration (e.g., geo-sharding, fallback trees, hybrid memory flows).