Expand ruff lint rules and rename make targets (chk→lint, fmt→format)

[laughingman7743]

Summary

• Add 11 new ruff lint rule categories: UP (pyupgrade), PIE, PERF, T20, FLY, ISC, RSE, RUF, PGH, G, PT
• Rename make chk → make lint, make fmt → make format across Makefile, CLAUDE.md, and docs
• Fix all violations across 138 files (net -173 lines), achieving 0 errors on make lint (ruff check + ruff format + mypy)

Key changes

• Type annotations: Modernize Union → X | Y, Optional → X | None (UP007)
• Logging: Convert f-string logging to lazy % formatting (G004/G003)
• Type safety: Replace blanket # type: ignore with specific error codes like [arg-type], [override] (PGH003)
• Class attributes: Add ClassVar for mutable class-level constants; use # noqa: RUF012 for SQLAlchemy-inherited attributes
• Test quality: Add match= to pytest.raises(ValueError) (PT011), extract setup from raises blocks (PT012), use tuple for parametrize names (PT006)
• Code modernization: yield from (UP028), collection unpacking (RUF005), remove unused imports (F401)
• Framework conventions: Per-file-ignores for SQLAlchemy naming (N802, N801), PEP 249 required names (B027, N818)

Logic changes (non-cosmetic)

Most changes are cosmetic (import cleanup, comment changes, annotation syntax). The following are the only changes that affect runtime or test behavior:

File	Change	Risk
pyathena/polars/result_set.py:683	for batch in ...: yield batch → yield from ... (UP028)	None — semantically identical
pyathena/arrow/result_set.py:193	[ISO8601] + self._timestamp_parsers → [ISO8601, *self._timestamp_parsers] (RUF005)	None — produces the same list
pyathena/pandas/util.py:292	Added type annotation list[concurrent.futures.Future[Any]] to futures variable	None — annotation only, no runtime effect
pyathena/pandas/result_set.py:371	Added # noqa: PERF203 to intentional try-except-in-loop	None — comment only
pyathena/s3fs/result_set.py:20	Union[type[...], type[...]] → type[...] | type[...] with from __future__ import annotations	None — string annotation, no runtime effect
9 files (G004/G003)	_logger.info(f"msg {var}") → _logger.info("msg %s", var)	None — same output; lazy formatting avoids string interpolation when log level is disabled
tests/pyathena/sqlalchemy/test_compiler.py	Moved stmt = select(...) before with pytest.raises(...) block (PT012)	None — same test logic, setup moved outside context manager
4 test files (PT011)	Added match= parameter to pytest.raises(ValueError)	Low — makes assertions stricter (tests will catch the right ValueError)
tests/pyathena/test_cursor.py	"query,description" → ("query", "description") in parametrize (PT006)	None — equivalent pytest syntax
tests/pyathena/aio/test_cursor.py, tests/pyathena/aio/s3fs/test_cursor.py	rows = []; async for row in cursor: rows.append(row) → rows = [row async for row in cursor] (PERF401)	None — semantically identical

Test plan

• make lint passes cleanly (ruff check, ruff format --check, mypy)
• make test passes
• make test-sqla passes

🤖 Generated with Claude Code