test(typing): Narrow Constructor, ConstructorEager aliases

[dangotbanned]

Will close #2346

Description (copied from issue)

When I initially tried to fix this in (#2232 (comment)) - most of the 795 errors were from tests using IntoFrame/IntoDataFrame.

narwhals/tests/utils.py

Lines 45 to 46 in d2a9f42

	Constructor: TypeAlias = Callable[[Any], IntoFrame]
	ConstructorEager: TypeAlias = Callable[[Any], IntoDataFrame]

AFAICT, none of these constructors should need to use a TypeAlias that mentions nw.DataFrame or nw.LazyFrame.
They all return a native object

conftest.py

narwhals/tests/conftest.py

Lines 72 to 222 in d2a9f42

	def pandas_constructor(obj: dict[str, list[Any]]) -> IntoDataFrame:
	import pandas as pd
	return pd.DataFrame(obj)
	def pandas_nullable_constructor(obj: dict[str, list[Any]]) -> IntoDataFrame:
	import pandas as pd
	return pd.DataFrame(obj).convert_dtypes(dtype_backend="numpy_nullable")
	def pandas_pyarrow_constructor(obj: dict[str, list[Any]]) -> IntoDataFrame:
	import pandas as pd
	return pd.DataFrame(obj).convert_dtypes(dtype_backend="pyarrow")
	def modin_constructor(obj: dict[str, list[Any]]) -> IntoDataFrame: # pragma: no cover
	import modin.pandas as mpd
	import pandas as pd
	return mpd.DataFrame(pd.DataFrame(obj)) # type: ignore[no-any-return]
	def modin_pyarrow_constructor(
	obj: dict[str, list[Any]],
	) -> IntoDataFrame: # pragma: no cover
	import modin.pandas as mpd
	import pandas as pd
	return mpd.DataFrame(pd.DataFrame(obj)).convert_dtypes(dtype_backend="pyarrow") # type: ignore[no-any-return]
	def cudf_constructor(obj: dict[str, list[Any]]) -> IntoDataFrame: # pragma: no cover
	import cudf
	return cudf.DataFrame(obj) # type: ignore[no-any-return]
	def polars_eager_constructor(obj: dict[str, list[Any]]) -> IntoDataFrame:
	import polars as pl
	return pl.DataFrame(obj)
	def polars_lazy_constructor(obj: dict[str, list[Any]]) -> pl.LazyFrame:
	import polars as pl
	return pl.LazyFrame(obj)
	def duckdb_lazy_constructor(obj: dict[str, list[Any]]) -> duckdb.DuckDBPyRelation:
	import duckdb
	import polars as pl
	_df = pl.LazyFrame(obj)
	return duckdb.table("_df")
	def dask_lazy_p1_constructor(obj: dict[str, list[Any]]) -> IntoFrame: # pragma: no cover
	import dask.dataframe as dd
	return dd.from_dict(obj, npartitions=1) # type: ignore[no-any-return]
	def dask_lazy_p2_constructor(obj: dict[str, list[Any]]) -> IntoFrame: # pragma: no cover
	import dask.dataframe as dd
	return dd.from_dict(obj, npartitions=2) # type: ignore[no-any-return]
	def pyarrow_table_constructor(obj: dict[str, Any]) -> IntoDataFrame:
	import pyarrow as pa
	return pa.table(obj)
	def pyspark_lazy_constructor() -> Callable[[Any], IntoFrame]: # pragma: no cover
	try:
	from pyspark.sql import SparkSession
	except ImportError: # pragma: no cover
	pytest.skip("pyspark is not installed")
	return None
	import warnings
	from atexit import register
	with warnings.catch_warnings():
	# The spark session seems to trigger a polars warning.
	# Polars is imported in the tests, but not used in the spark operations
	warnings.filterwarnings(
	"ignore", r"Using fork\(\) can cause Polars", category=RuntimeWarning
	)
	session = (
	SparkSession.builder.appName("unit-tests") # pyright: ignore[reportAttributeAccessIssue]
	.master("local[1]")
	.config("spark.ui.enabled", "false")
	# executing one task at a time makes the tests faster
	.config("spark.default.parallelism", "1")
	.config("spark.sql.shuffle.partitions", "2")
	# common timezone for all tests environments
	.config("spark.sql.session.timeZone", "UTC")
	.getOrCreate()
	)
	register(session.stop)
	def _constructor(obj: dict[str, list[Any]]) -> IntoFrame:
	_obj = deepcopy(obj)
	index_col_name = generate_temporary_column_name(n_bytes=8, columns=list(_obj))
	_obj[index_col_name] = list(range(len(_obj[next(iter(_obj))])))
	return ( # type: ignore[no-any-return]
	session.createDataFrame([*zip(*_obj.values())], schema=[*_obj.keys()])
	.repartition(2)
	.orderBy(index_col_name)
	.drop(index_col_name)
	)
	return _constructor
	def sqlframe_pyspark_lazy_constructor(
	obj: dict[str, Any],
	) -> IntoFrame: # pragma: no cover
	from sqlframe.duckdb import DuckDBSession
	session = DuckDBSession()
	return session.createDataFrame([*zip(*obj.values())], schema=[*obj.keys()])
	EAGER_CONSTRUCTORS: dict[str, Callable[[Any], IntoDataFrame]] = {
	"pandas": pandas_constructor,
	"pandas[nullable]": pandas_nullable_constructor,
	"pandas[pyarrow]": pandas_pyarrow_constructor,
	"pyarrow": pyarrow_table_constructor,
	"modin": modin_constructor,
	"modin[pyarrow]": modin_pyarrow_constructor,
	"cudf": cudf_constructor,
	"polars[eager]": polars_eager_constructor,
	}
	LAZY_CONSTRUCTORS: dict[str, Callable[[Any], IntoFrame]] = {
	"dask": dask_lazy_p2_constructor,
	"polars[lazy]": polars_lazy_constructor,
	"duckdb": duckdb_lazy_constructor,
	"pyspark": pyspark_lazy_constructor, # type: ignore[dict-item]
	"sqlframe": sqlframe_pyspark_lazy_constructor,
	}
	GPU_CONSTRUCTORS: dict[str, Callable[[Any], IntoFrame]] = {"cudf": cudf_constructor}

narwhals/narwhals/typing.py

Line 60 in d2a9f42

IntoDataFrame: TypeAlias = Union["NativeFrame", "DataFrame[Any]", "DataFrameLike"]

narwhals/narwhals/typing.py

Lines 75 to 77 in d2a9f42

	IntoFrame: TypeAlias = Union[
	"NativeFrame", "DataFrame[Any]", "LazyFrame[Any]", "DataFrameLike"
	]

[dangotbanned]

Not sure why the PR template didn't pull through 🤔

[dangotbanned]

After (#2344, #2345) the return types here can be reverted to IntoFrame, IntoDataFrame.

Like with (#2347) and IntoSeries - those aliases will be narrowed to what is seen here

[FBruzzesi]

Nice one Dan, thanks 🚀 No objections for having more precise hints

[dangotbanned]

Nice one Dan, thanks 🚀 No objections for having more precise hints

Thanks @FBruzzesi