chore: Use pass_through instead of strict in narwhals.from_native, use stable.v1 API more (#3693)

* chore: use `pass_through` instead of `strict` in `narwhals.from_native`, use stable.v1 API more

* unrelated(?) ruff fixes

* fixup, increase coverage

* use to_native method

* add flake8-tidy-imports rule

* take check out of loop

Author: MarcoGorelli

altair/_magics.py

@@ -7,8 +7,8 @@
 from importlib.util import find_spec
 from typing import Any
 
+import narwhals.stable.v1 as nw
 from IPython.core import magic_arguments
-from narwhals.dependencies import is_pandas_dataframe as _is_pandas_dataframe
 
 from altair.vegalite import v5 as vegalite_v5
 
@@ -32,7 +32,7 @@ def _prepare_data(data, data_transformers):
     """Convert input data to data for use within schema."""
     if data is None or isinstance(data, dict):
         return data
-    elif _is_pandas_dataframe(data):
+    elif nw.dependencies.is_pandas_dataframe(data):
         if func := data_transformers.get():
             data = func(data)
         return data

altair/utils/_vegafusion_data.py

@@ -5,7 +5,7 @@
 from typing import TYPE_CHECKING, Any, Callable, Final, TypedDict, Union, overload
 from weakref import WeakValueDictionary
 
-from narwhals.dependencies import is_into_dataframe
+import narwhals.stable.v1 as nw
 from packaging.version import Version
 
 from altair.utils._importers import import_vegafusion
@@ -22,7 +22,7 @@
     import sys
     from collections.abc import MutableMapping
 
-    from narwhals.typing import IntoDataFrame
+    from narwhals.stable.v1.typing import IntoDataFrame
 
     from vegafusion.runtime import ChartState
 
@@ -54,7 +54,9 @@
     def is_supported_by_vf(data: Any) -> TypeIs[DataFrameLike]:
         # Test whether VegaFusion supports the data type
         # VegaFusion v2 support narwhals-compatible DataFrames
-        return isinstance(data, DataFrameLike) or is_into_dataframe(data)
+        return isinstance(data, DataFrameLike) or nw.dependencies.is_into_dataframe(
+            data
+        )
 
 else:
 

altair/utils/core.py

@@ -16,8 +16,7 @@
 
 import jsonschema
 import narwhals.stable.v1 as nw
-from narwhals.dependencies import get_polars, is_pandas_dataframe
-from narwhals.typing import IntoDataFrame
+from narwhals.stable.v1.typing import IntoDataFrame
 
 from altair.utils.schemapi import SchemaBase, SchemaLike, Undefined
 
@@ -35,7 +34,7 @@
     import typing as t
 
     import pandas as pd
-    from narwhals.typing import IntoExpr
+    from narwhals.stable.v1.typing import IntoExpr
 
     from altair.utils._dfi_types import DataFrame as DfiDataFrame
     from altair.vegalite.v5.schema._typing import StandardType_T as InferredVegaLiteType
@@ -470,8 +469,9 @@ def sanitize_narwhals_dataframe(
     columns: list[IntoExpr] = []
     # See https://github.com/vega/altair/issues/1027 for why this is necessary.
     local_iso_fmt_string = "%Y-%m-%dT%H:%M:%S"
+    is_polars_dataframe = nw.dependencies.is_polars_dataframe(data.to_native())
     for name, dtype in schema.items():
-        if dtype == nw.Date and nw.get_native_namespace(data) is get_polars():
+        if dtype == nw.Date and is_polars_dataframe:
             # Polars doesn't allow formatting `Date` with time directives.
             # The date -> datetime cast is extremely fast compared with `to_string`
             columns.append(
@@ -673,8 +673,8 @@ def parse_shorthand(  # noqa: C901
             if schema[unescaped_field] in {
                 nw.Object,
                 nw.Unknown,
-            } and is_pandas_dataframe(nw.to_native(data_nw)):
-                attrs["type"] = infer_vegalite_type_for_pandas(nw.to_native(column))
+            } and nw.dependencies.is_pandas_dataframe(data_nw.to_native()):
+                attrs["type"] = infer_vegalite_type_for_pandas(column.to_native())
             else:
                 attrs["type"] = infer_vegalite_type_for_narwhals(column)
             if isinstance(attrs["type"], tuple):

altair/utils/data.py

@@ -19,8 +19,7 @@
 )
 
 import narwhals.stable.v1 as nw
-from narwhals.dependencies import is_pandas_dataframe as _is_pandas_dataframe
-from narwhals.typing import IntoDataFrame
+from narwhals.stable.v1.typing import IntoDataFrame
 
 from ._importers import import_pyarrow_interchange
 from .core import (
@@ -76,7 +75,7 @@ class SupportsGeoInterface(Protocol):
 
 def is_data_type(obj: Any) -> TypeIs[DataType]:
     return isinstance(obj, (dict, SupportsGeoInterface)) or isinstance(
-        nw.from_native(obj, eager_or_interchange_only=True, strict=False),
+        nw.from_native(obj, eager_or_interchange_only=True, pass_through=True),
         nw.DataFrame,
     )
 
@@ -188,7 +187,7 @@ def sample(
     if data is None:
         return partial(sample, n=n, frac=frac)
     check_data_type(data)
-    if _is_pandas_dataframe(data):
+    if nw.dependencies.is_pandas_dataframe(data):
         return data.sample(n=n, frac=frac)
     elif isinstance(data, dict):
         if "values" in data:
@@ -210,7 +209,7 @@ def sample(
             raise ValueError(msg)
         n = int(frac * len(data))
     indices = random.sample(range(len(data)), n)
-    return nw.to_native(data[indices])
+    return data[indices].to_native()
 
 
 _FormatType = Literal["csv", "json"]
@@ -319,11 +318,11 @@ def _to_text_kwds(prefix: str, extension: str, filename: str, urlpath: str, /) -
 def to_values(data: DataType) -> ToValuesReturnType:
     """Replace a DataFrame by a data model with values."""
     check_data_type(data)
-    # `strict=False` passes `data` through as-is if it is not a Narwhals object.
-    data_native = nw.to_native(data, strict=False)
+    # `pass_through=True` passes `data` through as-is if it is not a Narwhals object.
+    data_native = nw.to_native(data, pass_through=True)
     if isinstance(data_native, SupportsGeoInterface):
         return {"values": _from_geo_interface(data_native)}
-    elif _is_pandas_dataframe(data_native):
+    elif nw.dependencies.is_pandas_dataframe(data_native):
         data_native = sanitize_pandas_dataframe(data_native)
         return {"values": data_native.to_dict(orient="records")}
     elif isinstance(data_native, dict):
@@ -364,7 +363,7 @@ def _from_geo_interface(data: SupportsGeoInterface | Any) -> dict[str, Any]:
     - ``typing.TypeGuard``
     - ``pd.DataFrame.__getattr__``
     """
-    if _is_pandas_dataframe(data):
+    if nw.dependencies.is_pandas_dataframe(data):
         data = sanitize_pandas_dataframe(data)
     return sanitize_geo_interface(data.__geo_interface__)
 
@@ -374,7 +373,7 @@ def _data_to_json_string(data: DataType) -> str:
     check_data_type(data)
     if isinstance(data, SupportsGeoInterface):
         return json.dumps(_from_geo_interface(data))
-    elif _is_pandas_dataframe(data):
+    elif nw.dependencies.is_pandas_dataframe(data):
         data = sanitize_pandas_dataframe(data)
         return data.to_json(orient="records", double_precision=15)
     elif isinstance(data, dict):
@@ -401,7 +400,7 @@ def _data_to_csv_string(data: DataType) -> str:
             f"See https://github.com/vega/altair/issues/3441"
         )
         raise NotImplementedError(msg)
-    elif _is_pandas_dataframe(data):
+    elif nw.dependencies.is_pandas_dataframe(data):
         data = sanitize_pandas_dataframe(data)
         return data.to_csv(index=False)
     elif isinstance(data, dict):

altair/vegalite/v5/api.py

@@ -280,7 +280,7 @@ def _prepare_data(
     # convert dataframes  or objects with __geo_interface__ to dict
     elif not isinstance(data, dict) and _is_data_type(data):
         if func := data_transformers.get():
-            data = func(nw.to_native(data, strict=False))
+            data = func(nw.to_native(data, pass_through=True))
 
     # convert string input to a URLData
     elif isinstance(data, str):

pyproject.toml

@@ -34,7 +34,7 @@ keywords = [
 ]
 requires-python = ">=3.9"
 dynamic = ["version"]
-license-files = { paths = ["LICENSE"] }
+license = {file = "LICENSE"}
 classifiers = [
     "Development Status :: 5 - Production/Stable",
     "Environment :: Console",
@@ -398,6 +398,15 @@ Use `Union[T, None]` instead.
 which have a similar but different semantic meaning.
 See https://github.com/vega/altair/pull/3449
 """
+"narwhals.dependencies".msg = """
+Import `dependencies` from `narwhals.stable.v1` instead.
+"""
+"narwhals.typing".msg = """
+Import `typing` from `narwhals.stable.v1` instead.
+"""
+"narwhals.dtypes".msg = """
+Import `dtypes` from `narwhals.stable.v1` instead.
+"""
 
 [tool.ruff.lint.per-file-ignores]
 # Only enforce type annotation rules on public api

tests/utils/test_schemapi.py

@@ -37,7 +37,7 @@
 if TYPE_CHECKING:
     from collections.abc import Iterable, Sequence
 
-    from narwhals.typing import IntoDataFrame
+    from narwhals.stable.v1.typing import IntoDataFrame
 
 _JSON_SCHEMA_DRAFT_URL = load_schema()["$schema"]
 # Make tests inherit from _TestSchema, so that when we test from_dict it won't

tests/vegalite/v5/test_api.py

@@ -1723,6 +1723,18 @@ def test_polars_with_pandas_nor_pyarrow(monkeypatch: pytest.MonkeyPatch):
     assert "numpy" not in sys.modules
 
 
+def test_polars_date_32():
+    df = pl.DataFrame(
+        {"a": [1, 2, 3], "b": [date(2020, 1, 1), date(2020, 1, 2), date(2020, 1, 3)]}
+    )
+    result = alt.Chart(df).mark_line().encode(x="a", y="b").to_dict()
+    assert next(iter(result["datasets"].values())) == [
+        {"a": 1, "b": "2020-01-01T00:00:00"},
+        {"a": 2, "b": "2020-01-02T00:00:00"},
+        {"a": 3, "b": "2020-01-03T00:00:00"},
+    ]
+
+
 @skip_requires_pyarrow(requires_tzdata=True)
 def test_interchange_with_date_32():
     # Test that objects which Narwhals only supports at the interchange

tests/vegalite/v5/test_params.py

@@ -107,7 +107,7 @@ def test_parameter_naming():
     # test automatic naming which has the form such as param_5
     prm0, prm1, prm2 = (alt.param() for _ in range(3))
 
-    res = re.match("param_([0-9]+)", prm0.param.name)
+    res = re.match(r"param_([0-9]+)", prm0.param.name)
 
     assert res
 

tools/generate_schema_wrapper.py

@@ -751,7 +751,7 @@ def generate_vegalite_schema_wrapper(fp: Path, /) -> ModuleDef[str]:
         "from typing import Any, Literal, Union, Protocol, Sequence, List, Iterator, TYPE_CHECKING",
         "import pkgutil",
         "import json\n",
-        "from narwhals.dependencies import is_pandas_dataframe as _is_pandas_dataframe",
+        "import narwhals.stable.v1 as nw\n",
         "from altair.utils.schemapi import SchemaBase, Undefined, UndefinedType, _subclasses # noqa: F401\n",
         import_type_checking(
             "from datetime import date, datetime",