Add string.convert.convert_datetime/convert_booleans APIs to pylibcudf

python/cudf/cudf/_lib/string_casting.pyx

@@ -3,9 +3,6 @@
 from cudf._lib.column cimport Column
 
 from cudf._lib.scalar import as_device_scalar
-
-from cudf._lib.scalar cimport DeviceScalar
-
 from cudf._lib.types import SUPPORTED_NUMPY_TO_LIBCUDF_TYPES
 
 from libcpp.memory cimport unique_ptr
@@ -14,14 +11,6 @@ from libcpp.utility cimport move
 
 from pylibcudf.libcudf.column.column cimport column
 from pylibcudf.libcudf.column.column_view cimport column_view
-from pylibcudf.libcudf.scalar.scalar cimport string_scalar
-from pylibcudf.libcudf.strings.convert.convert_booleans cimport (
-    from_booleans as cpp_from_booleans,
-    to_booleans as cpp_to_booleans,
-)
-from pylibcudf.libcudf.strings.convert.convert_datetime cimport (
-    is_timestamp as cpp_is_timestamp,
-)
 from pylibcudf.libcudf.strings.convert.convert_floats cimport (
     from_floats as cpp_from_floats,
     to_floats as cpp_to_floats,
@@ -427,77 +416,21 @@ def stoul(Column input_col):
     return string_to_integer(input_col, cudf.dtype("uint64"))
 
 
-def _to_booleans(Column input_col, object string_true="True"):
-    """
-    Converting/Casting input column of type string to boolean column
-
-    Parameters
-    ----------
-    input_col : input column of type string
-    string_true : string that represents True
-
-    Returns
-    -------
-    A Column with string values cast to boolean
-    """
-
-    cdef DeviceScalar str_true = as_device_scalar(string_true)
-    cdef column_view input_column_view = input_col.view()
-    cdef const string_scalar* string_scalar_true = <const string_scalar*>(
-        str_true.get_raw_ptr())
-    cdef unique_ptr[column] c_result
-    with nogil:
-        c_result = move(
-            cpp_to_booleans(
-                input_column_view,
-                string_scalar_true[0]))
-
-    return Column.from_unique_ptr(move(c_result))
-
-
 def to_booleans(Column input_col):
-
-    return _to_booleans(input_col)
-
-
-def _from_booleans(
-        Column input_col,
-        object string_true="True",
-        object string_false="False"):
-    """
-    Converting/Casting input column of type boolean to string column
-
-    Parameters
-    ----------
-    input_col : input column of type boolean
-    string_true : string that represents True
-    string_false : string that represents False
-
-    Returns
-    -------
-    A Column with boolean values cast to string
-    """
-
-    cdef DeviceScalar str_true = as_device_scalar(string_true)
-    cdef DeviceScalar str_false = as_device_scalar(string_false)
-    cdef column_view input_column_view = input_col.view()
-    cdef const string_scalar* string_scalar_true = <const string_scalar*>(
-        str_true.get_raw_ptr())
-    cdef const string_scalar* string_scalar_false = <const string_scalar*>(
-        str_false.get_raw_ptr())
-    cdef unique_ptr[column] c_result
-    with nogil:
-        c_result = move(
-            cpp_from_booleans(
-                input_column_view,
-                string_scalar_true[0],
-                string_scalar_false[0]))
-
-    return Column.from_unique_ptr(move(c_result))
+    plc_column = plc.strings.convert.convert_booleans.to_booleans(
+        input_col.to_pylibcudf(mode="read"),
+        as_device_scalar("True").c_value,
+    )
+    return Column.from_pylibcudf(plc_column)
 
 
 def from_booleans(Column input_col):
-    return _from_booleans(input_col)
+    plc_column = plc.strings.convert.convert_booleans.from_booleans(
+        input_col.to_pylibcudf(mode="read"),
+        as_device_scalar("True").c_value,
+        as_device_scalar("False").c_value,
+    )
+    return Column.from_pylibcudf(plc_column)
 
 
 def int2timestamp(
@@ -520,11 +453,10 @@ def int2timestamp(
     A Column with date-time represented in string format
 
     """
-    cdef string c_timestamp_format = format.encode("UTF-8")
     return Column.from_pylibcudf(
         plc.strings.convert.convert_datetime.from_timestamps(
             input_col.to_pylibcudf(mode="read"),
-            c_timestamp_format,
+            format,
             names.to_pylibcudf(mode="read")
         )
     )
@@ -545,12 +477,11 @@ def timestamp2int(Column input_col, dtype, format):
 
     """
     dtype = dtype_to_pylibcudf_type(dtype)
-    cdef string c_timestamp_format = format.encode('UTF-8')
     return Column.from_pylibcudf(
         plc.strings.convert.convert_datetime.to_timestamps(
             input_col.to_pylibcudf(mode="read"),
             dtype,
-            c_timestamp_format
+            format
         )
     )
 
@@ -572,16 +503,11 @@ def istimestamp(Column input_col, str format):
     """
     if input_col.size == 0:
         return cudf.core.column.column_empty(0, dtype=cudf.dtype("bool"))
-    cdef column_view input_column_view = input_col.view()
-    cdef string c_timestamp_format = <string>str(format).encode('UTF-8')
-    cdef unique_ptr[column] c_result
-    with nogil:
-        c_result = move(
-            cpp_is_timestamp(
-                input_column_view,
-                c_timestamp_format))
-
-    return Column.from_unique_ptr(move(c_result))
+    plc_column = plc.strings.convert.convert_datetime.is_timestamp(
+        input_col.to_pylibcudf(mode="read"),
+        format
+    )
+    return Column.from_pylibcudf(plc_column)
 
 
 def timedelta2int(Column input_col, dtype, format):

python/cudf_polars/cudf_polars/dsl/expr.py

@@ -914,7 +914,7 @@ def do_evaluate(
             col = self.children[0].evaluate(df, context=context, mapping=mapping)
 
             is_timestamps = plc.strings.convert.convert_datetime.is_timestamp(
-                col.obj, format.encode()
+                col.obj, format
             )
 
             if strict:
@@ -937,7 +937,7 @@ def do_evaluate(
                 )
                 return Column(
                     plc.strings.convert.convert_datetime.to_timestamps(
-                        res.columns()[0], self.dtype, format.encode()
+                        res.columns()[0], self.dtype, format
                     )
                 )
         elif self.name == pl_expr.StringFunction.Replace:

python/pylibcudf/pylibcudf/libcudf/strings/convert/convert_booleans.pxd

@@ -8,10 +8,10 @@ from pylibcudf.libcudf.scalar.scalar cimport string_scalar
 cdef extern from "cudf/strings/convert/convert_booleans.hpp" namespace \
         "cudf::strings" nogil:
     cdef unique_ptr[column] to_booleans(
-        column_view input_col,
+        column_view input,
         string_scalar true_string) except +
 
     cdef unique_ptr[column] from_booleans(
-        column_view input_col,
+        column_view booleans,
         string_scalar true_string,
         string_scalar false_string) except +

python/pylibcudf/pylibcudf/libcudf/strings/convert/convert_datetime.pxd

@@ -10,14 +10,14 @@ from pylibcudf.libcudf.types cimport data_type
 cdef extern from "cudf/strings/convert/convert_datetime.hpp" namespace \
         "cudf::strings" nogil:
     cdef unique_ptr[column] to_timestamps(
-        column_view input_col,
+        column_view input,
         data_type timestamp_type,
         string format) except +
 
     cdef unique_ptr[column] from_timestamps(
-        column_view input_col,
+        column_view timestamps,
         string format,
-        column_view input_strings_names) except +
+        column_view names) except +
 
     cdef unique_ptr[column] is_timestamp(
         column_view input_col,

python/pylibcudf/pylibcudf/strings/convert/CMakeLists.txt

@@ -12,7 +12,7 @@
 # the License.
 # =============================================================================
 
-set(cython_sources convert_durations.pyx convert_datetime.pyx)
+set(cython_sources convert_booleans.pyx convert_durations.pyx convert_datetime.pyx)
 
 set(linked_libraries cudf::cudf)
 rapids_cython_create_modules(

python/pylibcudf/pylibcudf/strings/convert/__init__.pxd

@@ -1,2 +1,2 @@
 # Copyright (c) 2024, NVIDIA CORPORATION.
-from . cimport convert_datetime, convert_durations
+from . cimport convert_booleans, convert_datetime, convert_durations

python/pylibcudf/pylibcudf/strings/convert/__init__.py

@@ -1,2 +1,2 @@
 # Copyright (c) 2024, NVIDIA CORPORATION.
-from . import convert_datetime, convert_durations
+from . import convert_booleans, convert_datetime, convert_durations

python/pylibcudf/pylibcudf/strings/convert/convert_booleans.pxd

@@ -0,0 +1,9 @@
+# Copyright (c) 2024, NVIDIA CORPORATION.
+
+from pylibcudf.column cimport Column
+from pylibcudf.scalar cimport Scalar
+
+
+cpdef Column to_booleans(Column input, Scalar true_string)
+
+cpdef Column from_booleans(Column booleans, Scalar true_string, Scalar false_string)

python/pylibcudf/pylibcudf/strings/convert/convert_booleans.pyx

@@ -0,0 +1,91 @@
+# Copyright (c) 2024, NVIDIA CORPORATION.
+
+from libcpp.memory cimport unique_ptr
+from libcpp.utility cimport move
+from pylibcudf.column cimport Column
+from pylibcudf.libcudf.column.column cimport column
+from pylibcudf.libcudf.scalar.scalar cimport string_scalar
+from pylibcudf.libcudf.strings.convert cimport (
+    convert_booleans as cpp_convert_booleans,
+)
+from pylibcudf.scalar cimport Scalar
+
+from cython.operator import dereference
+
+
+cpdef Column to_booleans(Column input, Scalar true_string):
+    """
+    Returns a new bool column by parsing boolean values from the strings
+    in the provided strings column.
+
+    For details, see :cpp:func:`cudf::strings::to_booleans`.
+
+    Parameters
+    ----------
+    input :  Column
+        Strings instance for this operation
+
+    true_string : Scalar
+        String to expect for true. Non-matching strings are false
+
+    Returns
+    -------
+    Column
+        New bool column converted from strings.
+    """
+    cdef unique_ptr[column] c_result
+    cdef const string_scalar* c_true_string = <const string_scalar*>(
+        true_string.c_obj.get()
+    )
+
+    with nogil:
+        c_result = move(
+            cpp_convert_booleans.to_booleans(
+                input.view(),
+                dereference(c_true_string)
+            )
+        )
+
+    return Column.from_libcudf(move(c_result))
+
+cpdef Column from_booleans(Column booleans, Scalar true_string, Scalar false_string):
+    """
+    Returns a new strings column converting the boolean values from the
+    provided column into strings.
+
+    For details, see :cpp:func:`cudf::strings::from_booleans`.
+
+    Parameters
+    ----------
+    booleans :  Column
+        Boolean column to convert.
+
+    true_string : Scalar
+        String to use for true in the output column.
+
+    false_string : Scalar
+        String to use for false in the output column.
+
+    Returns
+    -------
+    Column
+        New strings column.
+    """
+    cdef unique_ptr[column] c_result
+    cdef const string_scalar* c_true_string = <const string_scalar*>(
+        true_string.c_obj.get()
+    )
+    cdef const string_scalar* c_false_string = <const string_scalar*>(
+        false_string.c_obj.get()
+    )
+
+    with nogil:
+        c_result = move(
+            cpp_convert_booleans.from_booleans(
+                booleans.view(),
+                dereference(c_true_string),
+                dereference(c_false_string),
+            )
+        )
+
+    return Column.from_libcudf(move(c_result))

python/pylibcudf/pylibcudf/strings/convert/convert_datetime.pxd

@@ -8,11 +8,16 @@ from pylibcudf.types cimport DataType
 cpdef Column to_timestamps(
     Column input,
     DataType timestamp_type,
-    const string& format
+    str format
 )
 
 cpdef Column from_timestamps(
-    Column input,
-    const string& format,
+    Column timestamps,
+    str format,
     Column input_strings_names
 )
+
+cpdef Column is_timestamp(
+    Column input,
+    str format,
+)