perf: make to_py_scalar 3x faster when argument is str

narwhals/stable/v1/__init__.py

@@ -51,7 +51,7 @@
 from narwhals.translate import _from_native_impl
 from narwhals.translate import get_native_namespace as nw_get_native_namespace
 from narwhals.translate import to_native
-from narwhals.translate import to_py_scalar as nw_to_py_scalar
+from narwhals.translate import to_py_scalar
 from narwhals.typing import IntoDataFrameT
 from narwhals.typing import IntoFrameT
 from narwhals.typing import IntoSeriesT
@@ -956,28 +956,6 @@ def wrapper(*args: Any, **kwargs: Any) -> Any:
         return decorator(func)
 
 
-def to_py_scalar(scalar: Any) -> Any:
-    """If a scalar is not Python native, converts it to Python native.
-
-    Raises:
-        ValueError: If the object is not convertible to a scalar.
-
-    Examples:
-        >>> import narwhals.stable.v1 as nw
-        >>> import pandas as pd
-        >>> df = nw.from_native(pd.DataFrame({"a": [1, 2, 3]}))
-        >>> nw.to_py_scalar(df["a"].item(0))
-        1
-        >>> import pyarrow as pa
-        >>> df = nw.from_native(pa.table({"a": [1, 2, 3]}))
-        >>> nw.to_py_scalar(df["a"].item(0))
-        1
-        >>> nw.to_py_scalar(1)
-        1
-    """
-    return _stableify(nw_to_py_scalar(scalar))
-
-
 def all() -> Expr:
     """
     Instantiate an expression representing all columns.

narwhals/translate.py

@@ -1,6 +1,5 @@
 from __future__ import annotations
 
-import numbers
 from datetime import datetime
 from datetime import timedelta
 from functools import wraps
@@ -46,6 +45,15 @@
 
 T = TypeVar("T")
 
+NON_TEMPORAL_SCALAR_TYPES = (
+    bool,
+    bytes,
+    str,
+    int,
+    float,
+    complex,
+)
+
 
 @overload
 def to_native(
@@ -843,7 +851,8 @@ def to_py_scalar(scalar_like: Any) -> Any:
         >>> nw.to_py_scalar(1)
         1
     """
-
+    if isinstance(scalar_like, NON_TEMPORAL_SCALAR_TYPES):
+        return scalar_like
     pa = get_pyarrow()
     if pa and isinstance(scalar_like, pa.Scalar):
         return scalar_like.as_py()
@@ -864,18 +873,11 @@ def to_py_scalar(scalar_like: Any) -> Any:
     if pd and isinstance(scalar_like, pd.Timedelta):
         return scalar_like.to_pytimedelta()
 
-    all_scalar_types = (
-        int,
-        float,
-        complex,
-        bool,
-        bytes,
-        str,
-        datetime,
-        timedelta,
-        numbers.Number,
-    )
-    if isinstance(scalar_like, all_scalar_types):
+    # SIM101 suggests merging the tuples, but that would create a new tuple,
+    # so doesn't seem ideal perf-wise?
+    if isinstance(scalar_like, NON_TEMPORAL_SCALAR_TYPES) or isinstance(  # noqa: SIM101
+        scalar_like, (datetime, timedelta)
+    ):
         return scalar_like
 
     msg = (

tests/stable_api_test.py

@@ -87,6 +87,10 @@ def test_stable_api_docstrings() -> None:
         nw_doc = getattr(nw, item).__doc__
         if item == "from_native":
             v1_doc = v1_doc.replace("native_dataframe", "native_object")
+        if item in {"to_py_scalar"}:
+            # We don't overwrite the docstring for these
+            assert v1_doc == nw_doc
+            continue
         assert (
             v1_doc.replace("import narwhals.stable.v1 as nw", "import narwhals as nw")
             == nw_doc