Add TTL support and batch operations

[rogeeff]

Summary

This PR adds time-to-live (TTL) support to the cache, allowing entries to expire automatically after a configurable duration. It also introduces batch operations for improved efficiency when working with multiple entries at once.

Changes

• TTL expiration: Each entry can optionally have a TTL (in seconds). Expired entries are lazily evicted on access or explicitly via cleanupExpired().
• find() method: Returns a direct reference to cached values, avoiding the copy overhead of get() for large value types.
• batch_get() / batch_put(): Bulk operations that acquire the lock once for better performance than individual calls in a loop.
• Internal refactor: Extracted get_internal() and put_internal() helpers to support lock management for batch operations.
• Benchmark updates: Added TTL benchmarking scenario and averaged multi-run benchmarks to the Python benchmark script.

Testing

• Added unit tests for find(), TTL, batch operations, and cleanup
• Existing tests pass unchanged