Configure memory usage

I saw a pull-request opened a while back that introduced the possibility of analyzing bigger than memory datasets by saving some of the state on disk, I was wondering if there is a way for the user to configure the maximum amount of memory the process can use?

Like `duckdb` has `"SET max_memory='1GB';"`

Start by locating the earlier pull request that added disk-backed state for datasets larger than memory, then trace how users configure execution settings in the Python bindings. Define how a user-configurable memory limit should behave, including its interaction with disk-backed processing, and add coverage showing that the configured limit is honored.

索引モデルが issue の本文から書いたものです。