feat: add Spark schema converter

[Tomoscorbin]

Summary

• add delta_engine.databricks.to_spark_schema() as a lazy public Spark interop helper
• keep native PySpark conversion inside the Spark adapter while preserving column order, spelling, types, and modeled nullability
• document metadata and array/map nullability semantics and preserve the dependency-free declaration/import surface
• cover every modeled scalar type, nested schemas, unsupported types, public imports, and missing-PySpark behavior

Why

ETL code can now reuse a DeltaTable declaration as its authoritative DataFrame schema without adding PySpark to DeltaTable or leaking Spark types into the backend-neutral schema API.

Validation

• uv run pytest — 1,245 passed, 78 deselected; 97.15% coverage
• uv run ruff check src tests
• uv run ruff format --check src tests
• uv run mypy src
• uv run lint-imports — all 7 contracts kept
• uv run --group docs sphinx-build -W -b html docs <temporary-output>