[FEAT] Add initial Spark Connect support (#3261)

Implements basic Spark Connect functionality in Daft with the following:

- Add `daft-connect` crate to handle Spark Connect protocol
- Implement configuration management via Spark Connect API
- Add Python bindings with `connect_start()` function
- Add integration tests for config operations

Currently supports:
- Basic session management
- Config operations (Set, Get, GetWithDefault, GetOption, GetAll, Unset)
- Error handling and status reporting

Notable changes:
- New dependency on `spark-connect` protocol
- Added `tracing` for debugging and monitoring
- Integration with existing Daft infrastructure

Some operations like execute_plan, analyze_plan, and artifact handling
are
currently unimplemented and will return appropriate error messages.