Feature: allow cache to be shared between multiple table instances

### SDK

Python

### Description

Every `Table` has a cache for index & metadata. Right now that cache is newly created for each table instance. This can lead to memory issues. However, it should be possible for multiple table instances to share the same cache. This would make it easy, for example, to create a bunch of tables to handle concurrent queries.

Start at the Python SDK's Table implementation and trace how its index and metadata cache is created for each instance. Determine how multiple Table instances could receive and share one cache, including its lifecycle, then add coverage showing that sharing works without duplicating the cache.

Written by the indexing model from the issue text.