Skip to content

Commit fd597e0

Browse files
mtsokolcopybara-github
authored andcommitted
Revert the last cpp protobuf update and GCS reader support.
The required riegeli version depends on protobug that is incompatible with TF. Ported from #175 PiperOrigin-RevId: 797851410
1 parent 6376a81 commit fd597e0

File tree

5 files changed

+18
-40
lines changed

5 files changed

+18
-40
lines changed

MODULE.bazel

Lines changed: 7 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -13,25 +13,24 @@
1313
# limitations under the License.
1414

1515
# TODO(fchern): automate version string alignment with setup.py
16-
VERSION = "0.8.0"
16+
VERSION = "0.8.1"
1717

1818
module(
1919
name = "array_record",
2020
version = VERSION,
2121
repo_name = "com_google_array_record",
2222
)
2323

24-
bazel_dep(name = "rules_proto", version = "7.1.0")
25-
bazel_dep(name = "rules_python", version = "1.4.1")
26-
bazel_dep(name = "platforms", version = "0.0.11")
27-
bazel_dep(name = "protobuf", version = "31.1")
24+
bazel_dep(name = "rules_proto", version = "7.0.2")
25+
bazel_dep(name = "rules_python", version = "0.37.0")
26+
bazel_dep(name = "platforms", version = "0.0.10")
27+
bazel_dep(name = "protobuf", version = "28.3")
2828
bazel_dep(name = "googletest", version = "1.15.2")
29-
bazel_dep(name = "abseil-cpp", version = "20250127.1")
29+
bazel_dep(name = "abseil-cpp", version = "20240722.0")
3030
bazel_dep(name = "abseil-py", version = "2.1.0")
3131
bazel_dep(name = "eigen", version = "3.4.0.bcr.3")
32-
bazel_dep(name = "riegeli", version = "0.0.0-20250717-5b2e77e")
32+
bazel_dep(name = "riegeli", version = "0.0.0-20241218-3385e3c")
3333
bazel_dep(name = "pybind11_bazel", version = "2.12.0")
34-
bazel_dep(name = "google_cloud_cpp", version = "3.0.0-rc0")
3534

3635
SUPPORTED_PYTHON_VERSIONS = [
3736
"3.10",

oss/build_whl.sh

Lines changed: 4 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -90,12 +90,10 @@ function main() {
9090
$PYTHON_BIN -m pip install ${OUTPUT_DIR}/all_dist/array_record*.whl
9191
$PYTHON_BIN -c 'import array_record'
9292
$PYTHON_BIN -c 'from array_record.python import array_record_data_source'
93-
# TF is not available on Python 3.13 and above.
94-
if [ "$(uname)" != "Darwin" ] && (( "${PYTHON_MINOR_VERSION}" < 13 )); then
95-
$PYTHON_BIN -m pip install jax tensorflow>=2.20.0 grain
96-
$PYTHON_BIN oss/test_import_grain.py
97-
$PYTHON_BIN oss/test_import_tensorflow.py
98-
fi
93+
$PYTHON_BIN -m pip install jax tensorflow>=2.20.0 grain
94+
# Re-enable the grain import test once the new version is released.
95+
# $PYTHON_BIN oss/test_import_grain.py
96+
$PYTHON_BIN oss/test_import_tensorflow.py
9997
}
10098

10199
main

python/BUILD

Lines changed: 0 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -20,8 +20,6 @@ pybind_extension(
2020
"@riegeli//riegeli/base:initializer",
2121
"@riegeli//riegeli/bytes:fd_reader",
2222
"@riegeli//riegeli/bytes:fd_writer",
23-
"@riegeli//riegeli/gcs:gcs_object",
24-
"@riegeli//riegeli/gcs:gcs_reader",
2523
],
2624
)
2725

python/array_record_module.cc

Lines changed: 6 additions & 23 deletions
Original file line numberDiff line numberDiff line change
@@ -22,7 +22,6 @@ limitations under the License.
2222
#include <vector>
2323

2424
#include "absl/status/status.h"
25-
#include "absl/strings/match.h"
2625
#include "absl/strings/str_format.h"
2726
#include "absl/strings/string_view.h"
2827
#include "cpp/array_record_reader.h"
@@ -35,8 +34,6 @@ limitations under the License.
3534
#include "riegeli/base/maker.h"
3635
#include "riegeli/bytes/fd_reader.h"
3736
#include "riegeli/bytes/fd_writer.h"
38-
#include "riegeli/gcs/gcs_object.h"
39-
#include "riegeli/gcs/gcs_reader.h"
4037

4138
namespace py = pybind11;
4239

@@ -53,13 +50,10 @@ PYBIND11_MODULE(array_record_module, m) {
5350
throw py::value_error(
5451
std::string(status_or_option.status().message()));
5552
}
56-
riegeli::FdWriterBase::Options file_writer_options;
57-
file_writer_options.set_buffer_size(size_t{16} << 20);
5853
// Release the GIL because IO is time consuming.
5954
py::gil_scoped_release scoped_release;
6055
return new array_record::ArrayRecordWriter(
61-
riegeli::Maker<riegeli::FdWriter>(
62-
path, std::move(file_writer_options)),
56+
riegeli::Maker<riegeli::FdWriter>(path),
6357
status_or_option.value());
6458
}),
6559
py::arg("path"), py::arg("options") = "")
@@ -90,29 +84,18 @@ PYBIND11_MODULE(array_record_module, m) {
9084
std::string(status_or_option.status().message()));
9185
}
9286
riegeli::FdReaderBase::Options file_reader_options;
93-
riegeli::GcsReader::Options gcs_reader_options;
9487
if (kwargs.contains("file_reader_buffer_size")) {
9588
auto file_reader_buffer_size =
9689
kwargs["file_reader_buffer_size"].cast<int64_t>();
9790
file_reader_options.set_buffer_size(file_reader_buffer_size);
98-
gcs_reader_options.set_buffer_size(file_reader_buffer_size);
9991
}
10092
// Release the GIL because IO is time consuming.
10193
py::gil_scoped_release scoped_release;
102-
if (absl::StartsWith(path, "gs://")) {
103-
return new array_record::ArrayRecordReader(
104-
riegeli::Maker<riegeli::GcsReader>(
105-
google::cloud::storage::Client(),
106-
riegeli::GcsObject(path), std::move(gcs_reader_options)),
107-
status_or_option.value(),
108-
array_record::ArrayRecordGlobalPool());
109-
} else {
110-
return new array_record::ArrayRecordReader(
111-
riegeli::Maker<riegeli::FdReader>(
112-
path, std::move(file_reader_options)),
113-
status_or_option.value(),
114-
array_record::ArrayRecordGlobalPool());
115-
}
94+
return new array_record::ArrayRecordReader(
95+
riegeli::Maker<riegeli::FdReader>(
96+
path, std::move(file_reader_options)),
97+
status_or_option.value(),
98+
array_record::ArrayRecordGlobalPool());
11699
}),
117100
py::arg("path"), py::arg("options") = "", R"(
118101
ArrayRecordReader for fast sequential or random access.

setup.py

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -31,7 +31,7 @@ def has_ext_modules(self):
3131

3232
setup(
3333
name='array_record',
34-
version='0.8.0',
34+
version='0.8.1',
3535
description='A file format that achieves a new frontier of IO efficiency',
3636
author='ArrayRecord team',
3737
author_email='[email protected]',

0 commit comments

Comments
 (0)