Align PandasCursor chunk processing with PolarsCursor implementation

[laughingman7743]

Summary

Align PandasCursor's chunk processing implementation and documentation with the recently merged PolarsCursor implementation (#637).

Background

PolarsCursor was updated with improved chunk processing that:

• Enables lazy loading for ALL data access methods (fetchone, fetchmany, fetchall) when chunksize is set
• Provides iter_chunks() method on both Cursor and ResultSet
• Has clear documentation explaining the lazy loading behavior

PandasCursor has the same underlying behavior but lacks:

• Consistent API (iter_chunks() on ResultSet)
• Clear documentation about lazy loading for fetch methods
• as_pandas() method on DataFrameIterator

Tasks

Code Changes

• Add iter_chunks() method to AthenaPandasResultSet
• Add as_pandas() method to DataFrameIterator (align with Polars)
• Update PandasCursor.iter_chunks() to delegate to ResultSet (like PolarsCursor)

Documentation Changes

• Update docs/pandas.rst to clarify that chunksize enables lazy loading for all fetch methods
• Add chunksize and iter_chunks() documentation to AsyncPandasCursor section

Import Fix

• Change import abc to from collections import abc (same fix as PolarsCursor)

Related

• PR Add chunk processing support for PolarsCursor #637: Add chunk processing support for PolarsCursor

🤖 Generated with Claude Code