[Python] Start of DuckDB Spark API

[Tishj]

This PR introduces the start of the Spark DuckDB API, this API aims to be 1 to 1 compatible with PySpark, but internally it will use DuckDB.

Namely this PR introduces the following core classes:

SparkSession

This class wraps a DuckDBPyConnection and acts as a wrapper around the connection
Methods that are already implemented:

• sql
• builder (and its relevant methods)
• table

DataFrame

This class wraps a DuckDBPyRelation and acts as a wrapper around the relation
Methods that are already implemented:

• show
• write.saveAsTable

Catalog

Part of the SparkSession, also wraps DuckDBPyConnection but exposes methods related to querying the catalog of the database
Methods that are already implemented:

• listDatabases
• listTables
• listColumns

Different levels of not implemented

Even the code for our Python client is written in C++, so for this API we wanted to write it in Python to make it easier for the users of the API to be able contribute to the API.

Some methods are planned to be implemented but aren't yet, but a lot of the methods are not explicitly planned to be implemented yet.

To differentiate between the two classes of "not implemented" we introduce a ContributionsAcceptedError, which also inherits from NotImplementedError so they can be generically dealt with, but semantically it serves as a way to communicate that the method that threw this exception is not planned to be implemented by the core team.

[Mytherin]

Thanks! Looks great