Pluggable Scan Planning

### Feature Request / Improvement

Umbrella issue. Today we have tightly coupled query planning in Python. This works well but runs into limitations when it comes to concurrency and performance. Having pluggable query planning will allow for:

- **Python based query planning** which is working today and is proven.
- **Iceberg-Rust based query planning** as suggeted in https://github.com/apache/iceberg-python/issues/2159#issuecomment-3033339213.
- **REST Catalog server side planning** this leverages the REST Catalog to plan the query, rather than the client. First issue can be found here: https://github.com/apache/iceberg-python/issues/2302.

As an API, I think it makes sense to follow the Server Side planning. So a `PlanTableScanRequest` as an input, and the [ScanTasks](https://github.com/apache/iceberg/blob/68a241615e104d97a86c932c381d7dfac802d769/open-api/rest-catalog-open-api.py#L1290) as the output.

Start by reading this umbrella issue alongside issue 2302, which covers the first REST Catalog server-side planning work. Review the proposed PlanTableScanRequest input and ScanTasks output, then compare them with the existing Python query-planning path. Done means an agreed pluggable planning API that accommodates Python, Iceberg-Rust, and REST Catalog planning.

Written by the indexing model from the issue text.