Remove warning and add counter to track non utf8 values.

PiperOrigin-RevId: 331036516

Author: paulgc17

tensorflow_data_validation/statistics/generators/lift_stats_generator.py

@@ -39,7 +39,6 @@
 
 from __future__ import print_function
 
-import logging
 import operator
 import typing
 from typing import Any, Dict, Iterator, Iterable, Optional, Sequence, Text, Tuple, Union
@@ -108,6 +107,10 @@
                                 [('y', _YType), ('y_count', _CountType),
                                  ('lift_values', Iterable[_LiftValue])])
 
+# Beam counter to track the number of non-utf8 values.
+_NON_UTF8_VALUES_COUNTER = beam.metrics.Metrics.counter(
+    constants.METRICS_NAMESPACE, 'num_non_utf8_values_lift_generator')
+
 
 def _get_example_value_presence(
     record_batch: pa.RecordBatch, path: types.FeaturePath,
@@ -279,13 +282,12 @@ def _to_partial_x_counts(
       yield _SlicedXKey(slice_key, x_path, x), x_count
 
 
-def _get_unicode_value(value: Union[Text, bytes], path: types.FeaturePath
-                       ) -> Text:
+def _get_unicode_value(value: Union[Text, bytes]) -> Text:
+  """Get feature value decoded as utf-8."""
   decoded_value = stats_util.maybe_get_utf8(value)
   # Check if we have a valid utf-8 string. If not, assign a placeholder.
   if decoded_value is None:
-    logging.warning('Feature "%s" has bytes value "%s" which cannot be '
-                    'decoded as a UTF-8 string.', path, value)
+    _NON_UTF8_VALUES_COUNTER.inc()
     decoded_value = constants.NON_UTF8_PLACEHOLDER
   return decoded_value
 
@@ -341,7 +343,7 @@ def _make_dataset_feature_stats_proto(
       lift_series_proto.y_string = y
       y_display_val = y
     elif isinstance(y, six.binary_type):
-      y_string = _get_unicode_value(y, y_path)
+      y_string = _get_unicode_value(y)
       lift_series_proto.y_string = y_string
       y_display_val = y_string
     else:
@@ -370,7 +372,7 @@ def _make_dataset_feature_stats_proto(
         lift_value_proto.x_string = x
         x_display_val = x
       elif isinstance(x, six.binary_type):
-        x_string = _get_unicode_value(x, key.x_path)
+        x_string = _get_unicode_value(x)
         lift_value_proto.x_string = x_string
         x_display_val = x_string
       else:

tensorflow_data_validation/utils/top_k_uniques_stats_util.py

@@ -19,9 +19,9 @@
 from __future__ import print_function
 
 import collections
-import logging
 from typing import FrozenSet, List, Optional, Union
 
+import apache_beam as beam
 import six
 from tensorflow_data_validation import constants
 from tensorflow_data_validation import types
@@ -33,6 +33,10 @@
 FeatureValueCount = collections.namedtuple('FeatureValueCount',
                                            ['feature_value', 'count'])
 
+# Beam counter to track the number of non-utf8 values.
+_NON_UTF8_VALUES_COUNTER = beam.metrics.Metrics.counter(
+    constants.METRICS_NAMESPACE, 'num_non_utf8_values_topk_uniques_generator')
+
 
 def make_feature_stats_proto_topk_uniques(
     feature_path: types.FeaturePath, is_categorical: bool,
@@ -174,8 +178,7 @@ def _make_feature_stats_proto_topk(
     if isinstance(value, six.binary_type):
       decoded_value = stats_util.maybe_get_utf8(value)
       if decoded_value is None:
-        logging.warning('Feature "%s" has bytes value "%s" which cannot be '
-                        'decoded as a UTF-8 string.', feature_path, value)
+        _NON_UTF8_VALUES_COUNTER.inc()
         value = constants.NON_UTF8_PLACEHOLDER
       else:
         value = decoded_value