Add Redis backend for persistent memory storage

[vgudur-dev]

Description

Implement a Redis-backed MemoryStore that persists agent memory across restarts and enables distributed agent deployments.

Motivation

The current InMemoryStore is great for development and testing, but production deployments need persistent storage. Redis is the most common choice for fast key-value storage in production AI systems.

Proposed Implementation

Create src/agent_memory_guard/storage/redis_store.py with:

1. RedisMemoryStore implementing the MemoryStore protocol
2. Support for Redis connection pools and Sentinel
3. Key namespacing for multi-tenant deployments
4. TTL support for automatic memory expiration
5. Atomic operations for concurrent agent access

Example API

from agent_memory_guard import MemoryGuard, Policy
from agent_memory_guard.storage.redis_store import RedisMemoryStore

store = RedisMemoryStore(url="redis://localhost:6379", namespace="agent-1")
guard = MemoryGuard(store=store, policy=Policy.strict())

Acceptance Criteria

• Implements full MemoryStore protocol
• Connection pooling and reconnection handling
• Key namespacing for multi-agent isolation
• Integration tests (can use fakeredis for CI)
• Add redis optional dependency group to pyproject.toml
• Performance benchmarks vs InMemoryStore

[snakefood3232]

Thanks for outlining the need for a RedisMemoryStore to persist agent memory across restarts. I'd handle this by creating redis_store.py to utilize Redis connection pools and Sentinel, implement key namespacing for multi-tenant setups, and ensure atomic operations for concurrent access. I can have a PR ready in about 5 days. I've worked on similar Redis integration projects for distributed AI systems, so I'm confident in delivering a robust solution.