feat(normalize): fail early on unsupported data type for file format

[AshwabhB]

Description

Runs the destination data-type/file-format compatibility check (verify_supported_data_types) during normalize, so an unsupported combination fails at the normalize step instead of surfacing later at load time. Reuses the existing load-time logic unchanged, so destinations that already declare exceptions (bigquery, redshift, athena, databricks, dremio, synapse, sqlalchemy) get the earlier check for free. model and reference jobs are skipped since they carry no typed data file.

Adds a destination_type field to NormalizeConfiguration so the error message can name the destination.

This covers task 3 of the issue. The Snowflake use_decfloat/parquet exception and remaining tasks will follow separately.

Related Issues

• Refs make the use of loader_file_format parameter consistent #815

Additional Context

Test: tests/normalize/test_normalize.py::test_normalize_fails_on_unsupported_data_type — bigquery cannot load a json column from parquet, and this now raises UnsupportedDataType during normalize. Ran make lint (mypy/ruff/flake8/bandit/docstrings) and the affected test-common paths locally; all green.