chore: merge recent changes from master

.github/sync-repo-settings.yaml

@@ -3,7 +3,7 @@
 branchProtectionRules:
 # Identifies the protection rule pattern. Name of the branch to be protected.
 # Defaults to `master`
-- pattern: master
+- pattern: '{master,v3}'
   requiredStatusCheckContexts:
     - 'Kokoro'
     - 'Kokoro snippets-3.8'

CHANGELOG.md

@@ -5,6 +5,24 @@
 [1]: https://pypi.org/project/google-cloud-bigquery/#history
 
 
+## [2.23.0](https://www.github.com/googleapis/python-bigquery/compare/v2.22.1...v2.23.0) (2021-07-27)
+
+
+### Features
+
+* Update proto definitions for bigquery/v2 to support new proto fields for BQML. ([#817](https://www.github.com/googleapis/python-bigquery/issues/817)) ([fe7a902](https://www.github.com/googleapis/python-bigquery/commit/fe7a902e8b3e723ace335c9b499aea6d180a025b))
+
+
+### Bug Fixes
+
+* no longer raise a warning in `to_dataframe` if `max_results` set ([#815](https://www.github.com/googleapis/python-bigquery/issues/815)) ([3c1be14](https://www.github.com/googleapis/python-bigquery/commit/3c1be149e76b1d1d8879fdcf0924ddb1c1839e94))
+* retry ChunkedEncodingError by default ([#802](https://www.github.com/googleapis/python-bigquery/issues/802)) ([419d36d](https://www.github.com/googleapis/python-bigquery/commit/419d36d6b1887041e5795dbc8fc808890e91ab11))
+
+
+### Documentation
+
+* correct docs for `LoadJobConfig.destination_table_description` ([#810](https://www.github.com/googleapis/python-bigquery/issues/810)) ([da87fd9](https://www.github.com/googleapis/python-bigquery/commit/da87fd921cc8067b187d7985c978aac8eb58d107))
+
 ### [2.22.1](https://www.github.com/googleapis/python-bigquery/compare/v2.22.0...v2.22.1) (2021-07-22)
 
 

google/cloud/bigquery/table.py

@@ -1544,11 +1544,6 @@ def _validate_bqstorage(self, bqstorage_client, create_bqstorage_client):
             return False
 
         if self.max_results is not None:
-            warnings.warn(
-                "Cannot use bqstorage_client if max_results is set, "
-                "reverting to fetching data with the REST endpoint.",
-                stacklevel=2,
-            )
             return False
 
         return True
@@ -1585,6 +1580,25 @@ def total_rows(self):
         """int: The total number of rows in the table."""
         return self._total_rows
 
+    def _maybe_warn_max_results(
+        self, bqstorage_client: Optional["bigquery_storage.BigQueryReadClient"],
+    ):
+        """Issue a warning if BQ Storage client is not ``None`` with ``max_results`` set.
+
+        This helper method should be used directly in the relevant top-level public
+        methods, so that the warning is issued for the correct line in user code.
+
+        Args:
+            bqstorage_client:
+                The BigQuery Storage client intended to use for downloading result rows.
+        """
+        if bqstorage_client is not None and self.max_results is not None:
+            warnings.warn(
+                "Cannot use bqstorage_client if max_results is set, "
+                "reverting to fetching data with the REST endpoint.",
+                stacklevel=3,
+            )
+
     def _to_page_iterable(
         self, bqstorage_download, tabledata_list_download, bqstorage_client=None
     ):
@@ -1674,6 +1688,8 @@ def to_arrow(
 
         .. versionadded:: 1.17.0
         """
+        self._maybe_warn_max_results(bqstorage_client)
+
         if not self._validate_bqstorage(bqstorage_client, create_bqstorage_client):
             create_bqstorage_client = False
             bqstorage_client = None
@@ -1763,6 +1779,8 @@ def to_dataframe_iterable(
         if dtypes is None:
             dtypes = {}
 
+        self._maybe_warn_max_results(bqstorage_client)
+
         column_names = [field.name for field in self._schema]
         bqstorage_download = functools.partial(
             _pandas_helpers.download_dataframe_bqstorage,
@@ -1866,6 +1884,8 @@ def to_dataframe(
         if dtypes is None:
             dtypes = {}
 
+        self._maybe_warn_max_results(bqstorage_client)
+
         if not self._validate_bqstorage(bqstorage_client, create_bqstorage_client):
             create_bqstorage_client = False
             bqstorage_client = None

google/cloud/bigquery/version.py

@@ -12,4 +12,4 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__version__ = "2.22.1"
+__version__ = "2.23.0"

google/cloud/bigquery_v2/types/model.py

@@ -96,6 +96,8 @@ class Model(proto.Message):
             Output only. Label columns that were used to train this
             model. The output of the model will have a `predicted_`
             prefix to these columns.
+        best_trial_id (int):
+            The best trial_id across all training runs.
     """
 
     class ModelType(proto.Enum):
@@ -113,6 +115,7 @@ class ModelType(proto.Enum):
         ARIMA = 11
         AUTOML_REGRESSOR = 12
         AUTOML_CLASSIFIER = 13
+        ARIMA_PLUS = 19
 
     class LossType(proto.Enum):
         r"""Loss metric to evaluate model training performance."""
@@ -151,6 +154,7 @@ class DataFrequency(proto.Enum):
         WEEKLY = 5
         DAILY = 6
         HOURLY = 7
+        PER_MINUTE = 8
 
     class HolidayRegion(proto.Enum):
         r"""Type of supported holiday regions for time series forecasting
@@ -285,7 +289,7 @@ class RegressionMetrics(proto.Message):
             median_absolute_error (google.protobuf.wrappers_pb2.DoubleValue):
                 Median absolute error.
             r_squared (google.protobuf.wrappers_pb2.DoubleValue):
-                R^2 score.
+                R^2 score. This corresponds to r2_score in ML.EVALUATE.
         """
 
         mean_absolute_error = proto.Field(
@@ -528,7 +532,7 @@ class ClusteringMetrics(proto.Message):
                 Mean of squared distances between each sample
                 to its cluster centroid.
             clusters (Sequence[google.cloud.bigquery_v2.types.Model.ClusteringMetrics.Cluster]):
-                [Beta] Information for all clusters.
+                Information for all clusters.
         """
 
         class Cluster(proto.Message):
@@ -697,10 +701,29 @@ class ArimaSingleModelForecastingMetrics(proto.Message):
                     Is arima model fitted with drift or not. It
                     is always false when d is not 1.
                 time_series_id (str):
-                    The id to indicate different time series.
+                    The time_series_id value for this time series. It will be
+                    one of the unique values from the time_series_id_column
+                    specified during ARIMA model training. Only present when
+                    time_series_id_column training option was used.
+                time_series_ids (Sequence[str]):
+                    The tuple of time_series_ids identifying this time series.
+                    It will be one of the unique tuples of values present in the
+                    time_series_id_columns specified during ARIMA model
+                    training. Only present when time_series_id_columns training
+                    option was used and the order of values here are same as the
+                    order of time_series_id_columns.
                 seasonal_periods (Sequence[google.cloud.bigquery_v2.types.Model.SeasonalPeriod.SeasonalPeriodType]):
                     Seasonal periods. Repeated because multiple
                     periods are supported for one time series.
+                has_holiday_effect (google.protobuf.wrappers_pb2.BoolValue):
+                    If true, holiday_effect is a part of time series
+                    decomposition result.
+                has_spikes_and_dips (google.protobuf.wrappers_pb2.BoolValue):
+                    If true, spikes_and_dips is a part of time series
+                    decomposition result.
+                has_step_changes (google.protobuf.wrappers_pb2.BoolValue):
+                    If true, step_changes is a part of time series decomposition
+                    result.
             """
 
             non_seasonal_order = proto.Field(
@@ -711,9 +734,19 @@ class ArimaSingleModelForecastingMetrics(proto.Message):
             )
             has_drift = proto.Field(proto.BOOL, number=3,)
             time_series_id = proto.Field(proto.STRING, number=4,)
+            time_series_ids = proto.RepeatedField(proto.STRING, number=9,)
             seasonal_periods = proto.RepeatedField(
                 proto.ENUM, number=5, enum="Model.SeasonalPeriod.SeasonalPeriodType",
             )
+            has_holiday_effect = proto.Field(
+                proto.MESSAGE, number=6, message=wrappers_pb2.BoolValue,
+            )
+            has_spikes_and_dips = proto.Field(
+                proto.MESSAGE, number=7, message=wrappers_pb2.BoolValue,
+            )
+            has_step_changes = proto.Field(
+                proto.MESSAGE, number=8, message=wrappers_pb2.BoolValue,
+            )
 
         non_seasonal_order = proto.RepeatedField(
             proto.MESSAGE, number=1, message="Model.ArimaOrder",
@@ -901,7 +934,7 @@ class TrainingRun(proto.Message):
         """
 
         class TrainingOptions(proto.Message):
-            r"""
+            r"""Options used in model training.
             Attributes:
                 max_iterations (int):
                     The maximum number of iterations in training.
@@ -972,8 +1005,9 @@ class TrainingOptions(proto.Message):
                 num_clusters (int):
                     Number of clusters for clustering models.
                 model_uri (str):
-                    [Beta] Google Cloud Storage URI from which the model was
-                    imported. Only applicable for imported models.
+                    Google Cloud Storage URI from which the model
+                    was imported. Only applicable for imported
+                    models.
                 optimization_strategy (google.cloud.bigquery_v2.types.Model.OptimizationStrategy):
                     Optimization strategy for training linear
                     regression models.
@@ -1030,8 +1064,11 @@ class TrainingOptions(proto.Message):
                     If a valid value is specified, then holiday
                     effects modeling is enabled.
                 time_series_id_column (str):
-                    The id column that will be used to indicate
-                    different time series to forecast in parallel.
+                    The time series id column that was used
+                    during ARIMA model training.
+                time_series_id_columns (Sequence[str]):
+                    The time series id columns that were used
+                    during ARIMA model training.
                 horizon (int):
                     The number of periods ahead that need to be
                     forecasted.
@@ -1042,6 +1079,15 @@ class TrainingOptions(proto.Message):
                     output feature name is A.b.
                 auto_arima_max_order (int):
                     The max value of non-seasonal p and q.
+                decompose_time_series (google.protobuf.wrappers_pb2.BoolValue):
+                    If true, perform decompose time series and
+                    save the results.
+                clean_spikes_and_dips (google.protobuf.wrappers_pb2.BoolValue):
+                    If true, clean spikes and dips in the input
+                    time series.
+                adjust_step_changes (google.protobuf.wrappers_pb2.BoolValue):
+                    If true, detect step changes and make data
+                    adjustment in the input time series.
             """
 
             max_iterations = proto.Field(proto.INT64, number=1,)
@@ -1120,9 +1166,19 @@ class TrainingOptions(proto.Message):
                 proto.ENUM, number=42, enum="Model.HolidayRegion",
             )
             time_series_id_column = proto.Field(proto.STRING, number=43,)
+            time_series_id_columns = proto.RepeatedField(proto.STRING, number=51,)
             horizon = proto.Field(proto.INT64, number=44,)
             preserve_input_structs = proto.Field(proto.BOOL, number=45,)
             auto_arima_max_order = proto.Field(proto.INT64, number=46,)
+            decompose_time_series = proto.Field(
+                proto.MESSAGE, number=50, message=wrappers_pb2.BoolValue,
+            )
+            clean_spikes_and_dips = proto.Field(
+                proto.MESSAGE, number=52, message=wrappers_pb2.BoolValue,
+            )
+            adjust_step_changes = proto.Field(
+                proto.MESSAGE, number=53, message=wrappers_pb2.BoolValue,
+            )
 
         class IterationResult(proto.Message):
             r"""Information about a single iteration of the training run.
@@ -1218,10 +1274,29 @@ class ArimaModelInfo(proto.Message):
                             Whether Arima model fitted with drift or not.
                             It is always false when d is not 1.
                         time_series_id (str):
-                            The id to indicate different time series.
+                            The time_series_id value for this time series. It will be
+                            one of the unique values from the time_series_id_column
+                            specified during ARIMA model training. Only present when
+                            time_series_id_column training option was used.
+                        time_series_ids (Sequence[str]):
+                            The tuple of time_series_ids identifying this time series.
+                            It will be one of the unique tuples of values present in the
+                            time_series_id_columns specified during ARIMA model
+                            training. Only present when time_series_id_columns training
+                            option was used and the order of values here are same as the
+                            order of time_series_id_columns.
                         seasonal_periods (Sequence[google.cloud.bigquery_v2.types.Model.SeasonalPeriod.SeasonalPeriodType]):
                             Seasonal periods. Repeated because multiple
                             periods are supported for one time series.
+                        has_holiday_effect (google.protobuf.wrappers_pb2.BoolValue):
+                            If true, holiday_effect is a part of time series
+                            decomposition result.
+                        has_spikes_and_dips (google.protobuf.wrappers_pb2.BoolValue):
+                            If true, spikes_and_dips is a part of time series
+                            decomposition result.
+                        has_step_changes (google.protobuf.wrappers_pb2.BoolValue):
+                            If true, step_changes is a part of time series decomposition
+                            result.
                     """
 
                     non_seasonal_order = proto.Field(
@@ -1237,11 +1312,21 @@ class ArimaModelInfo(proto.Message):
                     )
                     has_drift = proto.Field(proto.BOOL, number=4,)
                     time_series_id = proto.Field(proto.STRING, number=5,)
+                    time_series_ids = proto.RepeatedField(proto.STRING, number=10,)
                     seasonal_periods = proto.RepeatedField(
                         proto.ENUM,
                         number=6,
                         enum="Model.SeasonalPeriod.SeasonalPeriodType",
                     )
+                    has_holiday_effect = proto.Field(
+                        proto.MESSAGE, number=7, message=wrappers_pb2.BoolValue,
+                    )
+                    has_spikes_and_dips = proto.Field(
+                        proto.MESSAGE, number=8, message=wrappers_pb2.BoolValue,
+                    )
+                    has_step_changes = proto.Field(
+                        proto.MESSAGE, number=9, message=wrappers_pb2.BoolValue,
+                    )
 
                 arima_model_info = proto.RepeatedField(
                     proto.MESSAGE,
@@ -1319,6 +1404,7 @@ class ArimaModelInfo(proto.Message):
     label_columns = proto.RepeatedField(
         proto.MESSAGE, number=11, message=standard_sql.StandardSqlField,
     )
+    best_trial_id = proto.Field(proto.INT64, number=19,)
 
 
 class GetModelRequest(proto.Message):

google/cloud/bigquery_v2/types/table_reference.py

@@ -36,11 +36,23 @@ class TableReference(proto.Message):
             maximum length is 1,024 characters. Certain operations allow
             suffixing of the table ID with a partition decorator, such
             as ``sample_table$20190123``.
+        project_id_alternative (Sequence[str]):
+            The alternative field that will be used when ESF is not able
+            to translate the received data to the project_id field.
+        dataset_id_alternative (Sequence[str]):
+            The alternative field that will be used when ESF is not able
+            to translate the received data to the project_id field.
+        table_id_alternative (Sequence[str]):
+            The alternative field that will be used when ESF is not able
+            to translate the received data to the project_id field.
     """
 
     project_id = proto.Field(proto.STRING, number=1,)
     dataset_id = proto.Field(proto.STRING, number=2,)
     table_id = proto.Field(proto.STRING, number=3,)
+    project_id_alternative = proto.RepeatedField(proto.STRING, number=4,)
+    dataset_id_alternative = proto.RepeatedField(proto.STRING, number=5,)
+    table_id_alternative = proto.RepeatedField(proto.STRING, number=6,)
 
 
 __all__ = tuple(sorted(__protobuf__.manifest))