Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Migrate string capitalize APIs to pylibcudf #15503

Merged
Merged
Show file tree
Hide file tree
Changes from 19 commits
Commits
Show all changes
23 commits
Select commit Hold shift + click to select a range
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions python/cudf/cudf/_lib/pylibcudf/libcudf/CMakeLists.txt
Original file line number Diff line number Diff line change
Expand Up @@ -15,6 +15,7 @@
set(cython_sources aggregation.pyx binaryop.pyx copying.pyx replace.pyx reduce.pxd
stream_compaction.pyx types.pyx unary.pyx
)
add_subdirectory(strings)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

For consistency with other files I'd suggest putting this after the rapids_cython_create_modules call, but it doesn't really matter.


set(linked_libraries cudf::cudf)

Expand Down
Original file line number Diff line number Diff line change
@@ -0,0 +1,10 @@
# Copyright (c) 2024, NVIDIA CORPORATION.

from libcpp.memory cimport unique_ptr
from libcpp.string cimport string

from cudf._lib.pylibcudf.libcudf.scalar.scalar cimport scalar


cdef extern from "cudf/scalar/scalar_factories.hpp" namespace "cudf" nogil:
cdef unique_ptr[scalar] make_string_scalar(const string & _string) except +
23 changes: 23 additions & 0 deletions python/cudf/cudf/_lib/pylibcudf/libcudf/strings/CMakeLists.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,23 @@
# =============================================================================
# Copyright (c) 2024, NVIDIA CORPORATION.
#
# Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except
# in compliance with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software distributed under the License
# is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express
# or implied. See the License for the specific language governing permissions and limitations under
# the License.
# =============================================================================

set(cython_sources char_types.pyx)

set(linked_libraries cudf::cudf)

rapids_cython_create_modules(
CXX
SOURCE_FILES "${cython_sources}"
LINKED_LIBRARIES "${linked_libraries}" ASSOCIATED_TARGETS cudf MODULE_PREFIX cpp_strings
)
Original file line number Diff line number Diff line change
Expand Up @@ -3,11 +3,14 @@ from libcpp.memory cimport unique_ptr

from cudf._lib.pylibcudf.libcudf.column.column cimport column
from cudf._lib.pylibcudf.libcudf.column.column_view cimport column_view
from cudf._lib.pylibcudf.libcudf.scalar.scalar cimport string_scalar


cdef extern from "cudf/strings/capitalize.hpp" namespace "cudf::strings" nogil:
cdef unique_ptr[column] capitalize(
const column_view & strings) except +
const column_view & strings,
const string_scalar & delimiters
) except +

cdef unique_ptr[column] title(
const column_view & strings) except +
Expand Down
6 changes: 6 additions & 0 deletions python/cudf/cudf/_lib/pylibcudf/libcudf/strings/case.pxd
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,12 @@ from cudf._lib.pylibcudf.libcudf.column.column_view cimport column_view


cdef extern from "cudf/strings/case.hpp" namespace "cudf::strings" nogil:
cdef unique_ptr[column] capitalize(
const column_view & input) except +

cdef unique_ptr[column] is_title(
const column_view & input) except +

cdef unique_ptr[column] to_lower(
const column_view & strings) except +

Expand Down
23 changes: 12 additions & 11 deletions python/cudf/cudf/_lib/pylibcudf/libcudf/strings/char_types.pxd
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@
# Copyright (c) 2021-2024, NVIDIA CORPORATION.

from libc.stdint cimport uint32_t
from libcpp.memory cimport unique_ptr

from cudf._lib.pylibcudf.libcudf.column.column cimport column
Expand All @@ -10,17 +11,17 @@ from cudf._lib.pylibcudf.libcudf.scalar.scalar cimport string_scalar
cdef extern from "cudf/strings/char_types/char_types.hpp" \
namespace "cudf::strings" nogil:

ctypedef enum string_character_types:
DECIMAL 'cudf::strings::string_character_types::DECIMAL'
NUMERIC 'cudf::strings::string_character_types::NUMERIC'
DIGIT 'cudf::strings::string_character_types::DIGIT'
ALPHA 'cudf::strings::string_character_types::ALPHA'
SPACE 'cudf::strings::string_character_types::SPACE'
UPPER 'cudf::strings::string_character_types::UPPER'
LOWER 'cudf::strings::string_character_types::LOWER'
ALPHANUM 'cudf::strings::string_character_types::ALPHANUM'
CASE_TYPES 'cudf::strings::string_character_types::CASE_TYPES'
ALL_TYPES 'cudf::strings::string_character_types::ALL_TYPES'
cpdef enum class string_character_types(uint32_t):
DECIMAL
NUMERIC
DIGIT
ALPHA
SPACE
UPPER
LOWER
ALPHANUM
CASE_TYPES
ALL_TYPES

cdef extern from "cudf/strings/char_types/char_types.hpp" \
namespace "cudf::strings" nogil:
Expand Down
Empty file.
3 changes: 2 additions & 1 deletion python/cudf/cudf/_lib/pylibcudf/strings/CMakeLists.txt
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,8 @@
# the License.
# =============================================================================

set(cython_sources case.pyx find.pyx)
set(cython_sources capitalize.pyx case.pyx char_types.pyx find.pyx)

set(linked_libraries cudf::cudf)
rapids_cython_create_modules(
CXX
Expand Down
2 changes: 1 addition & 1 deletion python/cudf/cudf/_lib/pylibcudf/strings/__init__.pxd
Original file line number Diff line number Diff line change
@@ -1,3 +1,3 @@
# Copyright (c) 2024, NVIDIA CORPORATION.

from . cimport case, find
from . cimport capitalize, case, char_types, find
2 changes: 1 addition & 1 deletion python/cudf/cudf/_lib/pylibcudf/strings/__init__.py
Original file line number Diff line number Diff line change
@@ -1,3 +1,3 @@
# Copyright (c) 2024, NVIDIA CORPORATION.

from . import case, find
from . import capitalize, case, char_types, find
9 changes: 9 additions & 0 deletions python/cudf/cudf/_lib/pylibcudf/strings/capitalize.pxd
Original file line number Diff line number Diff line change
@@ -0,0 +1,9 @@
# Copyright (c) 2024, NVIDIA CORPORATION.

from cudf._lib.pylibcudf.column cimport Column
from cudf._lib.pylibcudf.scalar cimport Scalar


cpdef Column capitalize(Column input, Scalar delimiters=*)
cpdef Column title(Column input)
cpdef Column is_title(Column input)
62 changes: 62 additions & 0 deletions python/cudf/cudf/_lib/pylibcudf/strings/capitalize.pyx
Original file line number Diff line number Diff line change
@@ -0,0 +1,62 @@
# Copyright (c) 2024, NVIDIA CORPORATION.

from libcpp.memory cimport unique_ptr
from libcpp.utility cimport move

from cudf._lib.pylibcudf.column cimport Column
from cudf._lib.pylibcudf.libcudf.column.column cimport column
from cudf._lib.pylibcudf.libcudf.scalar.scalar cimport string_scalar
from cudf._lib.pylibcudf.libcudf.scalar.scalar_factories cimport (
make_string_scalar as cpp_make_string_scalar,
)
from cudf._lib.pylibcudf.libcudf.strings cimport capitalize as cpp_capitalize
from cudf._lib.pylibcudf.scalar cimport Scalar
from cudf._lib.pylibcudf.strings.char_types cimport string_character_types

from cython.operator import dereference


cpdef Column capitalize(
Column input,
Scalar delimiters=None
# TODO: default scalar values
# https://github.com/rapidsai/cudf/issues/15505
):

cdef unique_ptr[column] c_result

if delimiters is None:
delimiters = Scalar.from_libcudf(
cpp_make_string_scalar("".encode())
)

cdef const string_scalar* cpp_delimiters = <const string_scalar*>(
delimiters.c_obj.get()
)

with nogil:
c_result = cpp_capitalize.capitalize(
input.view(),
dereference(cpp_delimiters)
)

return Column.from_libcudf(move(c_result))


cpdef Column title(
Column input,
string_character_types sequence_type=string_character_types.ALPHA
):
cdef unique_ptr[column] c_result
with nogil:
c_result = cpp_capitalize.title(input.view())

return Column.from_libcudf(move(c_result))


cpdef Column is_title(Column input):
cdef unique_ptr[column] c_result
with nogil:
c_result = cpp_capitalize.is_title(input.view())

return Column.from_libcudf(move(c_result))
5 changes: 5 additions & 0 deletions python/cudf/cudf/_lib/pylibcudf/strings/char_types.pxd
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm fine leaving this as-is, but if the entire module is just the enum I sort of wonder if we shouldn't just lift it up directly into the __init__.py/pxd files.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

For me the main justification for leaving it is that if I'd usually #include a libcudf header to make use of some API or object, I think I should expect to import or cimport it from the corresponding pylibcudf "namespace".

Original file line number Diff line number Diff line change
@@ -0,0 +1,5 @@
# Copyright (c) 2024, NVIDIA CORPORATION.

from cudf._lib.pylibcudf.libcudf.strings.char_types cimport (
string_character_types,
)
4 changes: 4 additions & 0 deletions python/cudf/cudf/_lib/pylibcudf/strings/char_types.pyx
Original file line number Diff line number Diff line change
@@ -0,0 +1,4 @@
# Copyright (c) 2024, NVIDIA CORPORATION.

from cudf._lib.pylibcudf.libcudf.strings.char_types import \
string_character_types as StringCharacterTypes # no-cython-lint
48 changes: 17 additions & 31 deletions python/cudf/cudf/_lib/strings/capitalize.pyx
Original file line number Diff line number Diff line change
Expand Up @@ -2,47 +2,33 @@

from cudf.core.buffer import acquire_spill_lock

from libcpp.memory cimport unique_ptr
from libcpp.utility cimport move

from cudf._lib.column cimport Column
from cudf._lib.pylibcudf.libcudf.column.column cimport column
from cudf._lib.pylibcudf.libcudf.column.column_view cimport column_view
from cudf._lib.pylibcudf.libcudf.strings.capitalize cimport (
capitalize as cpp_capitalize,
is_title as cpp_is_title,
title as cpp_title,
)

import cudf._lib.pylibcudf as plc


@acquire_spill_lock()
def capitalize(Column source_strings):
cdef unique_ptr[column] c_result
cdef column_view source_view = source_strings.view()

with nogil:
c_result = move(cpp_capitalize(source_view))

return Column.from_unique_ptr(move(c_result))
return Column.from_pylibcudf(
plc.strings.capitalize.capitalize(
source_strings.to_pylibcudf(mode="read")
)
)


@acquire_spill_lock()
def title(Column source_strings):
cdef unique_ptr[column] c_result
cdef column_view source_view = source_strings.view()

with nogil:
c_result = move(cpp_title(source_view))

return Column.from_unique_ptr(move(c_result))
return Column.from_pylibcudf(
plc.strings.capitalize.title(
source_strings.to_pylibcudf(mode="read")
)
)


@acquire_spill_lock()
def is_title(Column source_strings):
cdef unique_ptr[column] c_result
cdef column_view source_view = source_strings.view()

with nogil:
c_result = move(cpp_is_title(source_view))

return Column.from_unique_ptr(move(c_result))
return Column.from_pylibcudf(
plc.strings.capitalize.is_title(
source_strings.to_pylibcudf(mode="read")
)
)
2 changes: 1 addition & 1 deletion python/cudf/cudf/pylibcudf_tests/common/utils.py
Original file line number Diff line number Diff line change
Expand Up @@ -35,7 +35,7 @@ def assert_column_eq(plc_column: plc.Column, pa_array: pa.Array) -> None:
plc_pa = plc_pa.combine_chunks()
if isinstance(pa_array, pa.ChunkedArray):
pa_array = pa_array.combine_chunks()

breakpoint()
assert plc_pa.equals(pa_array)


Expand Down
58 changes: 58 additions & 0 deletions python/cudf/cudf/pylibcudf_tests/test_string_capitalize.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,58 @@
# Copyright (c) 2024, NVIDIA CORPORATION.

import pyarrow as pa
import pytest
from utils import assert_column_eq

import cudf._lib.pylibcudf as plc


@pytest.fixture(scope="module")
def pa_data():
data = [
"leopard",
"Golden Eagle",
"SNAKE",
"",
"!A",
"hello World",
"A B C",
"#",
"AƻB",
"Ⓑⓖ",
"Art of War",
"The quick bRoWn fox juMps over the laze DOG",
'123nr98nv9rev!$#INF4390v03n1243<>?}{:-"',
"accénted",
None,
]
return pa.array(data)


@pytest.fixture(scope="module")
def plc_data(pa_data):
return plc.interop.from_arrow(pa_data)


def test_capitalize(plc_data, pa_data):
got = plc.strings.capitalize.capitalize(plc_data)
expected = pa.compute.utf8_capitalize(pa_data)
assert_column_eq(got, expected)


def test_title(plc_data, pa_data):
# A sequence shall be all characters that are not a space
# matches arrow for now
str_char_type = (
plc.strings.char_types.StringCharacterTypes.ALL_TYPES
& ~plc.strings.char_types.StringCharacterTypes.SPACE
)
brandon-b-miller marked this conversation as resolved.
Show resolved Hide resolved
got = plc.strings.capitalize.title(plc_data, str_char_type)
expected = pa.compute.utf8_title(pa_data)
assert_column_eq(got, expected)


def test_is_title(plc_data, pa_data):
got = plc.strings.capitalize.is_title(plc_data)
expected = pa.compute.utf8_is_title(pa_data)
assert_column_eq(got, expected)
Loading