[P0-Enhancement] Add embedding cache with Redis backend

[VirtualAgentics]

Description

Implement Redis-based embedding cache to reduce provider costs and improve performance by caching computed embeddings.

Background

This enhancement complements Issues #51 (persistence) and #52 (Redis) by adding a caching layer for embeddings, reducing API calls to external providers and improving response times.

Requirements

• Implement Redis-based embedding cache
• Cache key format: (model_id, sha256(content))
• Add cache TTL configuration
• Implement cache invalidation strategies
• Add cache hit/miss metrics
• Add cache size monitoring
• Implement cache warming strategies

Implementation Details

Files to modify:

• src/contextforge_memory/cache/ - New caching module
• src/contextforge_memory/embeddings/ - Add cache integration
• src/contextforge_memory/config.py - Add cache configuration
• tests/ - Add cache tests
• README.md - Document cache configuration

Technical approach:

• Create embedding cache service with Redis backend
• Implement cache-aside pattern for embeddings
• Add cache key generation with content hashing
• Implement cache metrics and monitoring
• Add cache configuration options

Acceptance Criteria

• Embeddings are cached correctly with proper keys
• Cache hit rate improves performance measurably
• Cache TTL and invalidation work correctly
• Cache metrics are tracked and exposed
• Configuration is flexible and documented

Testing Requirements

• Unit tests for cache operations
• Integration tests with Redis
• Performance tests for cache benefits
• Cache invalidation tests

Documentation Updates

• README.md - Document cache configuration
• Performance guide - Cache benefits
• Configuration examples - Cache setup
• Monitoring guide - Cache metrics

Related Issues

• Complements: [P0] Add pluggable vector index backend (FAISS/pgvector/Weaviate) #51 (persistence), [P0] Add persistent session store with Redis/Postgres backend and TTL #52 (Redis)
• Depends on: None
• Blocks: None