chore(engine): Create framework for query execution

[chaudum]

What this PR does / why we need it

This PR implements basic query execution using pull-based iterators. These iterators are defined by the Pipeline interface.

// Pipeline represents a data processing pipeline that can read Arrow records.
// It provides methods to read data, access the current record, and close resources.
type Pipeline interface {
	// Read reads the next value into its state.
	// It returns an error if reading fails or when the pipeline is exhausted. In this case, the function returns EOF.
	// The implementation must retain the returned error in its state and return it with subsequent Value() calls.
	Read() error
	// Value returns the current value in state.
	Value() (arrow.Record, error)
	// Close closes the resources of the pipeline.
	// The implementation must close all the of the pipeline's inputs.
	Close()
	// Inputs returns the inputs of the pipeline.
	Inputs() []Pipeline
	// Transport returns the type of transport of the implementation.
	Transport() Transport
}

Pipelines can be chained and the parent pulls results from its children by calling Read().

---

This PR supersedes #17110

---

TODO

• Implement iterator for DataObjScan
• Implement iterator for Filter
• Implement iterator for Projection
• Implement sorting by timestamp column, not by column index 2
• Correctly implement memory management for batches (arrow.Record)