Description

Added support for Amazon S3 Tables (managed Iceberg tables) within the Athena Engine Adapter.

Specific changes include:

Catalog Support: Updated AthenaEngineAdapter.catalog_support to CatalogSupport.FULL_SUPPORT to allow cross-catalog queries without issuing SET CATALOG commands. This prevents SQLMesh from stripping out custom catalog names (e.g., s3tablescatalog/...) from model names.
S3 Table Partitioning: For S3 tables, partition properties are now correctly generated as partitioning=ARRAY['...'] instead of the standard partitioned_by property format.
S3 Table Properties: S3 Table CTAS queries in Athena implicitly default to table_type='ICEBERG' and do not allow explicit specification of most table properties (except format='PARQUET'). The adapter now filters out unsupported TBLPROPERTIES and table_type when generating CREATE TABLE ... AS statements for S3 tables to prevent pyathena.error.OperationalError.
Dictionary Mutation Bug: Fixed a subtle bug in _build_table_properties_exp where table_properties was being mutated during AST parsing, causing properties like s3_table to drop out during intermediate schema table creation.

Test Plan

Tested creating S3 Iceberg tables via sqlmesh plan utilizing custom MODEL properties (s3_table true / catalog defined).

Verified partitioning is properly formulated as an unquoted ARRAY structure for S3 tables.

Verified that temporary intermediate schema generation properly inherits S3 properties without failure.

Ran make fast-test to ensure no regressions with standard Hive/Iceberg tables in the Athena adapter

Checklist

I have run make style and fixed any issues

I have added tests for my changes (if applicable)

All existing tests pass (make fast-test)

My commits are signed off (git commit -s) per the DCO