Add string.convert.convert_ipv4 APIs to pylibcudf (#16994)

Contributes to #15162

Authors:
  - Matthew Roeschke (https://github.com/mroeschke)

Approvers:
  - Vyas Ramasubramani (https://github.com/vyasr)

URL: #16994

python/cudf/cudf/_lib/string_casting.pyx

@@ -22,11 +22,6 @@ from pylibcudf.libcudf.strings.convert.convert_integers cimport (
     is_hex as cpp_is_hex,
     to_integers as cpp_to_integers,
 )
-from pylibcudf.libcudf.strings.convert.convert_ipv4 cimport (
-    integers_to_ipv4 as cpp_integers_to_ipv4,
-    ipv4_to_integers as cpp_ipv4_to_integers,
-    is_ipv4 as cpp_is_ipv4,
-)
 from pylibcudf.libcudf.types cimport data_type, type_id
 
 from cudf._lib.types cimport underlying_type_t_type_id
@@ -569,14 +564,10 @@ def int2ip(Column input_col):
     A Column with integer represented in string ipv4 format
 
     """
-
-    cdef column_view input_column_view = input_col.view()
-    cdef unique_ptr[column] c_result
-    with nogil:
-        c_result = move(
-            cpp_integers_to_ipv4(input_column_view))
-
-    return Column.from_unique_ptr(move(c_result))
+    plc_column = plc.strings.convert.convert_ipv4.integers_to_ipv4(
+        input_col.to_pylibcudf(mode="read")
+    )
+    return Column.from_pylibcudf(plc_column)
 
 
 def ip2int(Column input_col):
@@ -592,14 +583,10 @@ def ip2int(Column input_col):
     A Column with ipv4 represented as integer
 
     """
-
-    cdef column_view input_column_view = input_col.view()
-    cdef unique_ptr[column] c_result
-    with nogil:
-        c_result = move(
-            cpp_ipv4_to_integers(input_column_view))
-
-    return Column.from_unique_ptr(move(c_result))
+    plc_column = plc.strings.convert.convert_ipv4.ipv4_to_integers(
+        input_col.to_pylibcudf(mode="read")
+    )
+    return Column.from_pylibcudf(plc_column)
 
 
 def is_ipv4(Column source_strings):
@@ -608,15 +595,10 @@ def is_ipv4(Column source_strings):
     that have strings in IPv4 format. This format is nnn.nnn.nnn.nnn
     where nnn is integer digits in [0,255].
     """
-    cdef unique_ptr[column] c_result
-    cdef column_view source_view = source_strings.view()
-
-    with nogil:
-        c_result = move(cpp_is_ipv4(
-            source_view
-        ))
-
-    return Column.from_unique_ptr(move(c_result))
+    plc_column = plc.strings.convert.convert_ipv4.is_ipv4(
+        source_strings.to_pylibcudf(mode="read")
+    )
+    return Column.from_pylibcudf(plc_column)
 
 
 def htoi(Column input_col, **kwargs):

python/pylibcudf/pylibcudf/libcudf/strings/convert/convert_ipv4.pxd

@@ -8,11 +8,11 @@ from pylibcudf.libcudf.column.column_view cimport column_view
 cdef extern from "cudf/strings/convert/convert_ipv4.hpp" namespace \
         "cudf::strings" nogil:
     cdef unique_ptr[column] ipv4_to_integers(
-        column_view input_col) except +
+        column_view input) except +
 
     cdef unique_ptr[column] integers_to_ipv4(
-        column_view input_col) except +
+        column_view integers) except +
 
     cdef unique_ptr[column] is_ipv4(
-        column_view source_strings
+        column_view input
     ) except +

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

@@ -13,7 +13,7 @@
 # =============================================================================
 
 set(cython_sources convert_booleans.pyx convert_datetime.pyx convert_durations.pyx
-                   convert_fixed_point.pyx
+                   convert_fixed_point.pyx convert_ipv4.pyx
 )
 
 set(linked_libraries cudf::cudf)

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

@@ -4,4 +4,5 @@ from . cimport (
     convert_datetime,
     convert_durations,
     convert_fixed_point,
+    convert_ipv4,
 )

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

@@ -4,4 +4,5 @@
     convert_datetime,
     convert_durations,
     convert_fixed_point,
+    convert_ipv4,
 )

python/pylibcudf/pylibcudf/strings/convert/convert_ipv4.pxd

@@ -0,0 +1,10 @@
+# Copyright (c) 2024, NVIDIA CORPORATION.
+
+from pylibcudf.column cimport Column
+
+
+cpdef Column ipv4_to_integers(Column input)
+
+cpdef Column integers_to_ipv4(Column integers)
+
+cpdef Column is_ipv4(Column input)

python/pylibcudf/pylibcudf/strings/convert/convert_ipv4.pyx

@@ -0,0 +1,92 @@
+# 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.strings.convert cimport convert_ipv4 as cpp_convert_ipv4
+
+
+cpdef Column ipv4_to_integers(Column input):
+    """
+    Converts IPv4 addresses into integers.
+
+    For details, see cpp:func:`cudf::strings::ipv4_to_integers`
+
+    Parameters
+    ----------
+    input : Column
+        Strings instance for this operation
+
+    Returns
+    -------
+    Column
+        New uint32 column converted from strings.
+    """
+    cdef unique_ptr[column] c_result
+
+    with nogil:
+        c_result = move(
+            cpp_convert_ipv4.ipv4_to_integers(
+                input.view()
+            )
+        )
+
+    return Column.from_libcudf(move(c_result))
+
+
+cpdef Column integers_to_ipv4(Column integers):
+    """
+    Converts integers into IPv4 addresses as strings.
+
+    For details, see cpp:func:`cudf::strings::integers_to_ipv4`
+
+    Parameters
+    ----------
+    integers : Column
+        Integer (uint32) column to convert.
+
+    Returns
+    -------
+    Column
+        New strings column.
+    """
+    cdef unique_ptr[column] c_result
+
+    with nogil:
+        c_result = move(
+            cpp_convert_ipv4.integers_to_ipv4(
+                integers.view()
+            )
+        )
+
+    return Column.from_libcudf(move(c_result))
+
+
+cpdef Column is_ipv4(Column input):
+    """
+    Returns a boolean column identifying strings in which all
+    characters are valid for conversion to integers from IPv4 format.
+
+    For details, see cpp:func:`cudf::strings::is_ipv4`
+
+    Parameters
+    ----------
+    input : Column
+        Strings instance for this operation.
+
+    Returns
+    -------
+    Column
+        New column of boolean results for each string.
+    """
+    cdef unique_ptr[column] c_result
+
+    with nogil:
+        c_result = move(
+            cpp_convert_ipv4.is_ipv4(
+                input.view()
+            )
+        )
+
+    return Column.from_libcudf(move(c_result))

python/pylibcudf/pylibcudf/tests/test_string_convert_ipv4.py

@@ -0,0 +1,31 @@
+# Copyright (c) 2024, NVIDIA CORPORATION.
+import pyarrow as pa
+import pylibcudf as plc
+from utils import assert_column_eq
+
+
+def test_ipv4_to_integers():
+    arr = pa.array(["123.45.67.890", None])
+    result = plc.strings.convert.convert_ipv4.ipv4_to_integers(
+        plc.interop.from_arrow(arr)
+    )
+    expected = pa.array([2066564730, None], type=pa.uint32())
+    assert_column_eq(result, expected)
+
+
+def test_integers_to_ipv4():
+    arr = pa.array([1, 0, None], type=pa.uint32())
+    result = plc.strings.convert.convert_ipv4.integers_to_ipv4(
+        plc.interop.from_arrow(arr)
+    )
+    expected = pa.array(["0.0.0.1", "0.0.0.0", None])
+    assert_column_eq(result, expected)
+
+
+def test_is_ipv4():
+    arr = pa.array(["0.0.0.1", "1.2.34", "A", None])
+    result = plc.strings.convert.convert_ipv4.is_ipv4(
+        plc.interop.from_arrow(arr)
+    )
+    expected = pa.array([True, False, False, None])
+    assert_column_eq(result, expected)