Incremental Append Scan

[smaheshwar-pltr]

Feature Request / Improvement

Related is #240 for Incremental Changelog Scan support. This issue tracks Incremental Append Scan support for PyIceberg.

PyIceberg lacks incremental read utilities. It's been asked for multiple times - a strength of PyIceberg is that small data, such as just data appended between two snapshots, can be processed theoretically much faster than e.g. Spark.

This is an incremental append scan, that the Java API and the Spark engine support.

[smaheshwar-pltr]

#533 was the V0 - the work has since been split up into the following smaller PRs for review-ability:

1. [Append Scan] Introduce an AbstractTableScan with default methods #2230
2. [Append Scan] Extract manifest group planning into separate class #2232
3. [Append Scan] Add __eq__ and __hash__ methods to ManifestFile #2233
4. [Append Scan] Introduce IncrementalAppendScan class (without integration tests) #2234 (Introduce IncrementalAppendScan class smaheshwar-pltr/iceberg-python#5 is smaller diff)
5. [Append Scan] Integration tests for IncrementalAppendScan #2235 (Incremental append scan integration tests smaheshwar-pltr/iceberg-python#6 is smaller diff)