Prepare code for breaking change in Protobuf C++ API.

Protobuf 6.30.0 will change the return types of Descriptor::name() and other methods to absl::string_view. This makes the code work both before and after such a change. PiperOrigin-RevId: 689135228

Author: tfx-copybara

.github/reusable-build/action.yml

@@ -1032,15 +1032,15 @@ void Schema::UpdateFeatureInternal(
  ->FindValueByNumber(view.type());
  string data_type_name = (descriptor == nullptr)
  ? absl::StrCat("unknown(", view.type(), ")")
- : descriptor->name();
+ : std::string(descriptor->name());
 
  const google::protobuf::EnumValueDescriptor* schema_descriptor =
  tensorflow::metadata::v0::FeatureType_descriptor()->FindValueByNumber(
  feature->type());
  string schema_type_name =
  (schema_descriptor == nullptr)
  ? absl::StrCat("unknown(", feature->type(), ")")
- : schema_descriptor->name();
+ : std::string(schema_descriptor->name());
  descriptions->push_back(
  {tensorflow::metadata::v0::AnomalyInfo::UNEXPECTED_DATA_TYPE,
  "Unexpected data type",

.github/workflows/build.yml

@@ -19,7 +19,6 @@
 from __future__ import print_function
 
 import os
-import pytest
 import tempfile
 from absl.testing import absltest
 import apache_beam as beam
@@ -44,7 +43,6 @@ class StatsAPITest(absltest.TestCase):
  def _get_temp_dir(self):
  return tempfile.mkdtemp()
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_stats_pipeline(self):
  record_batches = [
  pa.RecordBatch.from_arrays([
@@ -203,7 +201,6 @@ def test_stats_pipeline(self):
  }
  """, statistics_pb2.DatasetFeatureStatisticsList())
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_stats_pipeline_with_examples_with_no_values(self):
  record_batches = [
  pa.RecordBatch.from_arrays([
@@ -321,7 +318,6 @@ def test_stats_pipeline_with_examples_with_no_values(self):
  test_util.make_dataset_feature_stats_list_proto_equal_fn(
  self, expected_result, check_histograms=False))
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_stats_pipeline_with_zero_examples(self):
  expected_result = text_format.Parse(
  """
@@ -343,7 +339,6 @@ def test_stats_pipeline_with_zero_examples(self):
  test_util.make_dataset_feature_stats_list_proto_equal_fn(
  self, expected_result, check_histograms=False))
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_stats_pipeline_with_sample_rate(self):
  record_batches = [
  pa.RecordBatch.from_arrays(
@@ -493,7 +488,6 @@ def test_write_stats_to_tfrecord_and_binary(self):
 
 class MergeDatasetFeatureStatisticsListTest(absltest.TestCase):
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_merges_two_shards(self):
  stats1 = text_format.Parse(
  """

.github/workflows/test.yml

@@ -20,7 +20,6 @@
 from __future__ import print_function
 
 import os
-import pytest
 import tempfile
 
 from absl.testing import absltest
@@ -3173,14 +3172,6 @@ class IdentifyAnomalousExamplesTest(parameterized.TestCase):
  @parameterized.named_parameters(*IDENTIFY_ANOMALOUS_EXAMPLES_VALID_INPUTS)
  def test_identify_anomalous_examples(self, examples, schema_text,
  expected_result):
-
- if self._testMethodName in [
- "test_identify_anomalous_examples_same_anomaly_reason",
- "test_identify_anomalous_examples_no_anomalies",
- "test_identify_anomalous_examples_different_anomaly_reasons"
- ]:
- pytest.xfail(reason="PR 260 This test fails and needs to be fixed. ")
-
  schema = text_format.Parse(schema_text, schema_pb2.Schema())
  options = stats_options.StatsOptions(schema=schema)
 
@@ -3241,7 +3232,6 @@ def _assert_skew_pairs_equal(self, actual, expected) -> None:
  for each in actual:
  self.assertIn(each, expected)
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_detect_feature_skew(self):
  training_data = [
  text_format.Parse("""

tensorflow_data_validation/anomalies/schema.cc

@@ -21,7 +21,7 @@
 from __future__ import print_function
 
 import sys
-import pytest
+from absl.testing import absltest
 from absl.testing import parameterized
 import apache_beam as beam
 from apache_beam.testing import util
@@ -366,7 +366,6 @@
 ]
 
 
-@pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed. ")
 class CSVDecoderTest(parameterized.TestCase):
  """Tests for CSV decoder."""
 
@@ -406,3 +405,7 @@ def test_csv_decoder_invalid_row(self):
  | csv_decoder.DecodeCSV(column_names=column_names))
  util.assert_that(
  result, test_util.make_arrow_record_batches_equal_fn(self, None))
+
+
+if __name__ == '__main__':
+ absltest.main()

tensorflow_data_validation/api/stats_api_test.py

@@ -18,7 +18,6 @@
 from __future__ import print_function
 
 import copy
-import pytest
 import os
 
 from absl import flags
@@ -1738,7 +1737,6 @@
 ]
 
 
-@pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed. ")
 class SequenceExampleStatsTest(parameterized.TestCase):
 
  @classmethod
@@ -1789,6 +1787,7 @@ def _assert_features_equal(lhs, rhs):
  rhs_schema_copy.ClearField('feature')
  self.assertEqual(lhs_schema_copy, rhs_schema_copy)
  _assert_features_equal(lhs, rhs)
+
  @parameterized.named_parameters(*_TEST_CASES)
  def test_e2e(self, stats_options, expected_stats_pbtxt,
  expected_inferred_schema_pbtxt, schema_for_validation_pbtxt,

tensorflow_data_validation/api/validation_api_test.py

@@ -15,7 +15,6 @@
 
 import traceback
 
-import pytest
 from absl.testing import absltest
 from absl.testing import parameterized
 import apache_beam as beam
@@ -142,7 +141,6 @@ def _make_ex(identifier: str,
 
 class FeatureSkewDetectorTest(parameterized.TestCase):
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_detect_feature_skew(self):
  baseline_examples, test_examples, _ = get_test_input(
  include_skewed_features=True, include_close_floats=True)
@@ -194,7 +192,6 @@ def test_detect_feature_skew(self):
  skew_result,
  test_util.make_skew_result_equal_fn(self, expected_result))
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_detect_no_skew(self):
  baseline_examples, test_examples, _ = get_test_input(
  include_skewed_features=False, include_close_floats=False)
@@ -224,7 +221,6 @@ def test_detect_no_skew(self):
  util.assert_that(skew_sample, make_sample_equal_fn(self, 0, []),
  'CheckSkewSample')
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_obtain_skew_sample(self):
  baseline_examples, test_examples, skew_pairs = get_test_input(
  include_skewed_features=True, include_close_floats=False)
@@ -248,7 +244,6 @@ def test_obtain_skew_sample(self):
  skew_sample, make_sample_equal_fn(self, sample_size,
  potential_samples))
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_empty_inputs(self):
  baseline_examples, test_examples, _ = get_test_input(
  include_skewed_features=True, include_close_floats=True)
@@ -304,7 +299,6 @@ def test_empty_inputs(self):
  make_sample_equal_fn(self, 0, expected_result),
  'CheckSkewSample')
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_float_precision_configuration(self):
  baseline_examples, test_examples, _ = get_test_input(
  include_skewed_features=True, include_close_floats=True)
@@ -395,7 +389,6 @@ def test_no_identifier_features(self):
  _ = ((baseline_examples, test_examples)
  | feature_skew_detector.DetectFeatureSkewImpl([]))
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_duplicate_identifiers_allowed_with_duplicates(self):
  base_example_1 = text_format.Parse(
  """
@@ -469,7 +462,6 @@ def test_duplicate_identifiers_allowed_with_duplicates(self):
  skew_result,
  test_util.make_skew_result_equal_fn(self, expected_result))
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_duplicate_identifiers_not_allowed_with_duplicates(self):
  base_example_1 = text_format.Parse(
  """
@@ -535,7 +527,6 @@ def test_duplicate_identifiers_not_allowed_with_duplicates(self):
  self.assertLen(actual_counter, 1)
  self.assertEqual(actual_counter[0].committed, 1)
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_skips_missing_identifier_example(self):
  base_example_1 = text_format.Parse(
  """
@@ -576,7 +567,6 @@ def test_skips_missing_identifier_example(self):
  runner = p.run()
  runner.wait_until_finish()
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_empty_features_equivalent(self):
  base_example_1 = text_format.Parse(
  """
@@ -626,7 +616,6 @@ def test_empty_features_equivalent(self):
  runner = p.run()
  runner.wait_until_finish()
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_empty_features_not_equivalent_to_missing(self):
  base_example_1 = text_format.Parse(
  """
@@ -699,7 +688,6 @@ def test_telemetry(self):
  self.assertLen(actual_counter, 1)
  self.assertEqual(actual_counter[0].committed, 1)
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_confusion_analysis(self):
 
  baseline_examples = [
@@ -834,7 +822,6 @@ def test_confusion_analysis_errors(self, input_example, expected_error_regex):
  feature_skew_detector.ConfusionConfig(name='val'),
  ]))[feature_skew_detector.CONFUSION_KEY]
 
- @pytest.mark.xfail(run=False, reason="PR 260 This test fails and needs to be fixed.")
  def test_match_stats(self):
  baseline_examples = [
  _make_ex('id0'),