fix(python): improve ingest from numpy scalar values (pola-rs#12025)

py-polars/polars/datatypes/convert.py

@@ -297,6 +297,7 @@ def DTYPE_TO_PY_TYPE(self) -> dict[PolarsDataType, PythonDataType]:
     def NUMPY_KIND_AND_ITEMSIZE_TO_DTYPE(self) -> dict[tuple[str, int], PolarsDataType]:
         return {
             # (np.dtype().kind, np.dtype().itemsize)
+            ("b", 1): Boolean,
             ("i", 1): Int8,
             ("i", 2): Int16,
             ("i", 4): Int32,

py-polars/polars/dependencies.py

@@ -207,20 +207,28 @@ def _might_be(cls: type, type_: str) -> bool:
         return False
 
 
-def _check_for_numpy(obj: Any) -> bool:
-    return _NUMPY_AVAILABLE and _might_be(cast(Hashable, type(obj)), "numpy")
+def _check_for_numpy(obj: Any, *, check_type: bool = True) -> bool:
+    return _NUMPY_AVAILABLE and _might_be(
+        cast(Hashable, type(obj) if check_type else obj), "numpy"
+    )
 
 
-def _check_for_pandas(obj: Any) -> bool:
-    return _PANDAS_AVAILABLE and _might_be(cast(Hashable, type(obj)), "pandas")
+def _check_for_pandas(obj: Any, *, check_type: bool = True) -> bool:
+    return _PANDAS_AVAILABLE and _might_be(
+        cast(Hashable, type(obj) if check_type else obj), "pandas"
+    )
 
 
-def _check_for_pyarrow(obj: Any) -> bool:
-    return _PYARROW_AVAILABLE and _might_be(cast(Hashable, type(obj)), "pyarrow")
+def _check_for_pyarrow(obj: Any, *, check_type: bool = True) -> bool:
+    return _PYARROW_AVAILABLE and _might_be(
+        cast(Hashable, type(obj) if check_type else obj), "pyarrow"
+    )
 
 
-def _check_for_pydantic(obj: Any) -> bool:
-    return _PYDANTIC_AVAILABLE and _might_be(cast(Hashable, type(obj)), "pydantic")
+def _check_for_pydantic(obj: Any, *, check_type: bool = True) -> bool:
+    return _PYDANTIC_AVAILABLE and _might_be(
+        cast(Hashable, type(obj) if check_type else obj), "pydantic"
+    )
 
 
 __all__ = [

py-polars/polars/utils/_construction.py

@@ -43,6 +43,7 @@
     Utf8,
     dtype_to_py_type,
     is_polars_dtype,
+    numpy_char_code_to_dtype,
     py_type_to_dtype,
 )
 from polars.datatypes.constructor import (
@@ -554,9 +555,17 @@ def sequence_to_pyseries(
             constructor = py_type_to_constructor(python_dtype)
             if constructor == PySeries.new_object:
                 try:
-                    return PySeries.new_from_anyvalues(name, values, strict)
-                # raised if we cannot convert to Wrap<AnyValue>
+                    srs = PySeries.new_from_anyvalues(name, values, strict)
+                    if _check_for_numpy(python_dtype, check_type=False) and isinstance(
+                        np.bool_(True), np.generic
+                    ):
+                        dtype = numpy_char_code_to_dtype(np.dtype(python_dtype).char)
+                        return srs.cast(dtype, strict=strict)
+                    else:
+                        return srs
+
                 except RuntimeError:
+                    # raised if we cannot convert to Wrap<AnyValue>
                     return sequence_from_anyvalue_or_object(name, values)
 
             return _construct_series_with_fallbacks(

py-polars/tests/unit/test_constructors.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 
 import sys
-from collections import namedtuple
+from collections import OrderedDict, namedtuple
 from datetime import date, datetime, timedelta, timezone
 from decimal import Decimal
 from random import shuffle
@@ -611,6 +611,21 @@ def test_init_ndarray(monkeypatch: Any) -> None:
     assert df2.rows() == [(1.0, 4.0), (2.5, None), (None, 6.5)]
 
 
+def test_init_numpy_scalars() -> None:
+    df = pl.DataFrame(
+        {
+            "bool": [np.bool_(True), np.bool_(False)],
+            "i8": [np.int8(16), np.int8(64)],
+            "u32": [np.uint32(1234), np.uint32(9876)],
+        }
+    )
+    df_expected = pl.from_records(
+        data=[(True, 16, 1234), (False, 64, 9876)],
+        schema=OrderedDict([("bool", pl.Boolean), ("i8", pl.Int8), ("u32", pl.UInt32)]),
+    )
+    assert_frame_equal(df, df_expected)
+
+
 def test_null_array_print_format() -> None:
     pa_tbl_null = pa.table({"a": [None, None]})
     df_null = pl.from_arrow(pa_tbl_null)