def execute(self, context: "Context"):
self.hook = AutoMLHook(
gcp_conn_id=self.gcp_conn_id,
delegate_to=self.delegate_to,
impersonation_chain=self.impersonation_chain,
)
model = self.hook.create_auto_ml_image_training_job(
project_id=self.project_id,
region=self.region,
display_name=self.display_name,
dataset=datasets.ImageDataset(dataset_name=self.dataset_id),
prediction_type=self.prediction_type,
multi_label=self.multi_label,
model_type=self.model_type,
base_model=self.base_model,
labels=self.labels,
training_encryption_spec_key_name=self.training_encryption_spec_key_name,
model_encryption_spec_key_name=self.model_encryption_spec_key_name,
training_fraction_split=self.training_fraction_split,
validation_fraction_split=self.validation_fraction_split,
test_fraction_split=self.test_fraction_split,
training_filter_split=self.training_filter_split,
validation_filter_split=self.validation_filter_split,
test_filter_split=self.test_filter_split,
budget_milli_node_hours=self.budget_milli_node_hours,
model_display_name=self.model_display_name,
model_labels=self.model_labels,
disable_early_stopping=self.disable_early_stopping,
sync=self.sync,
)
result = Model.to_dict(model)
model_id = self.hook.extract_model_id(result)
VertexAIModelLink.persist(context=context, task_instance=self, model_id=model_id)
return result
def execute(self, context: "Context"):
self.hook = AutoMLHook(
gcp_conn_id=self.gcp_conn_id,
delegate_to=self.delegate_to,
impersonation_chain=self.impersonation_chain,
)
model = self.hook.create_auto_ml_video_training_job(
project_id=self.project_id,
region=self.region,
display_name=self.display_name,
dataset=datasets.VideoDataset(dataset_name=self.dataset_id),
prediction_type=self.prediction_type,
model_type=self.model_type,
labels=self.labels,
training_encryption_spec_key_name=self.training_encryption_spec_key_name,
model_encryption_spec_key_name=self.model_encryption_spec_key_name,
training_fraction_split=self.training_fraction_split,
test_fraction_split=self.test_fraction_split,
training_filter_split=self.training_filter_split,
test_filter_split=self.test_filter_split,
model_display_name=self.model_display_name,
model_labels=self.model_labels,
sync=self.sync,
)
result = Model.to_dict(model)
model_id = self.hook.extract_model_id(result)
VertexAIModelLink.persist(context=context, task_instance=self, model_id=model_id)
return result
class CreateAutoMLVideoTrainingJobOperator(AutoMLTrainingJobBaseOperator):
"""Create Auto ML Video Training job"""
template_fields = [
'region',
'impersonation_chain',
]
operator_extra_links = (VertexAIModelLink(), )
def __init__(
self,
*,
dataset_id: str,
prediction_type: str = "classification",
model_type: str = "CLOUD",
training_filter_split: Optional[str] = None,
test_filter_split: Optional[str] = None,
**kwargs,
) -> None:
super().__init__(**kwargs)
self.dataset_id = dataset_id
self.prediction_type = prediction_type
self.model_type = model_type
self.training_filter_split = training_filter_split
self.test_filter_split = test_filter_split
def execute(self, context: "Context"):
self.hook = AutoMLHook(
gcp_conn_id=self.gcp_conn_id,
delegate_to=self.delegate_to,
impersonation_chain=self.impersonation_chain,
)
model = self.hook.create_auto_ml_video_training_job(
project_id=self.project_id,
region=self.region,
display_name=self.display_name,
dataset=datasets.VideoDataset(dataset_name=self.dataset_id),
prediction_type=self.prediction_type,
model_type=self.model_type,
labels=self.labels,
training_encryption_spec_key_name=self.
training_encryption_spec_key_name,
model_encryption_spec_key_name=self.model_encryption_spec_key_name,
training_fraction_split=self.training_fraction_split,
test_fraction_split=self.test_fraction_split,
training_filter_split=self.training_filter_split,
test_filter_split=self.test_filter_split,
model_display_name=self.model_display_name,
model_labels=self.model_labels,
sync=self.sync,
)
result = Model.to_dict(model)
model_id = self.hook.extract_model_id(result)
VertexAIModelLink.persist(context=context,
task_instance=self,
model_id=model_id)
return result
def execute(self, context: "Context"):
self.hook = AutoMLHook(
gcp_conn_id=self.gcp_conn_id,
delegate_to=self.delegate_to,
impersonation_chain=self.impersonation_chain,
)
model = self.hook.create_auto_ml_forecasting_training_job(
project_id=self.project_id,
region=self.region,
display_name=self.display_name,
dataset=datasets.TimeSeriesDataset(dataset_name=self.dataset_id),
target_column=self.target_column,
time_column=self.time_column,
time_series_identifier_column=self.time_series_identifier_column,
unavailable_at_forecast_columns=self.
unavailable_at_forecast_columns,
available_at_forecast_columns=self.available_at_forecast_columns,
forecast_horizon=self.forecast_horizon,
data_granularity_unit=self.data_granularity_unit,
data_granularity_count=self.data_granularity_count,
optimization_objective=self.optimization_objective,
column_specs=self.column_specs,
column_transformations=self.column_transformations,
labels=self.labels,
training_encryption_spec_key_name=self.
training_encryption_spec_key_name,
model_encryption_spec_key_name=self.model_encryption_spec_key_name,
training_fraction_split=self.training_fraction_split,
validation_fraction_split=self.validation_fraction_split,
test_fraction_split=self.test_fraction_split,
predefined_split_column_name=self.predefined_split_column_name,
weight_column=self.weight_column,
time_series_attribute_columns=self.time_series_attribute_columns,
context_window=self.context_window,
export_evaluated_data_items=self.export_evaluated_data_items,
export_evaluated_data_items_bigquery_destination_uri=(
self.export_evaluated_data_items_bigquery_destination_uri),
export_evaluated_data_items_override_destination=(
self.export_evaluated_data_items_override_destination),
quantiles=self.quantiles,
validation_options=self.validation_options,
budget_milli_node_hours=self.budget_milli_node_hours,
model_display_name=self.model_display_name,
model_labels=self.model_labels,
sync=self.sync,
)
result = Model.to_dict(model)
model_id = self.hook.extract_model_id(result)
VertexAIModelLink.persist(context=context,
task_instance=self,
model_id=model_id)
return result
def execute(self, context: "Context"):
self.hook = AutoMLHook(
gcp_conn_id=self.gcp_conn_id,
delegate_to=self.delegate_to,
impersonation_chain=self.impersonation_chain,
)
model = self.hook.create_auto_ml_tabular_training_job(
project_id=self.project_id,
region=self.region,
display_name=self.display_name,
dataset=datasets.TabularDataset(dataset_name=self.dataset_id),
target_column=self.target_column,
optimization_prediction_type=self.optimization_prediction_type,
optimization_objective=self.optimization_objective,
column_specs=self.column_specs,
column_transformations=self.column_transformations,
optimization_objective_recall_value=self.
optimization_objective_recall_value,
optimization_objective_precision_value=self.
optimization_objective_precision_value,
labels=self.labels,
training_encryption_spec_key_name=self.
training_encryption_spec_key_name,
model_encryption_spec_key_name=self.model_encryption_spec_key_name,
training_fraction_split=self.training_fraction_split,
validation_fraction_split=self.validation_fraction_split,
test_fraction_split=self.test_fraction_split,
predefined_split_column_name=self.predefined_split_column_name,
timestamp_split_column_name=self.timestamp_split_column_name,
weight_column=self.weight_column,
budget_milli_node_hours=self.budget_milli_node_hours,
model_display_name=self.model_display_name,
model_labels=self.model_labels,
disable_early_stopping=self.disable_early_stopping,
export_evaluated_data_items=self.export_evaluated_data_items,
export_evaluated_data_items_bigquery_destination_uri=(
self.export_evaluated_data_items_bigquery_destination_uri),
export_evaluated_data_items_override_destination=(
self.export_evaluated_data_items_override_destination),
sync=self.sync,
)
result = Model.to_dict(model)
model_id = self.hook.extract_model_id(result)
VertexAIModelLink.persist(context=context,
task_instance=self,
model_id=model_id)
return result
def execute(self, context: "Context"):
hook = AutoMLHook(
gcp_conn_id=self.gcp_conn_id,
delegate_to=self.delegate_to,
impersonation_chain=self.impersonation_chain,
)
results = hook.list_training_pipelines(
region=self.region,
project_id=self.project_id,
page_size=self.page_size,
page_token=self.page_token,
filter=self.filter,
read_mask=self.read_mask,
retry=self.retry,
timeout=self.timeout,
metadata=self.metadata,
)
VertexAITrainingPipelinesLink.persist(context=context, task_instance=self)
return [TrainingPipeline.to_dict(result) for result in results]
def execute(self, context: "Context"):
hook = AutoMLHook(
gcp_conn_id=self.gcp_conn_id,
delegate_to=self.delegate_to,
impersonation_chain=self.impersonation_chain,
)
try:
self.log.info("Deleting Auto ML training pipeline: %s", self.training_pipeline)
training_pipeline_operation = hook.delete_training_pipeline(
training_pipeline=self.training_pipeline,
region=self.region,
project_id=self.project_id,
retry=self.retry,
timeout=self.timeout,
metadata=self.metadata,
)
hook.wait_for_operation(timeout=self.timeout, operation=training_pipeline_operation)
self.log.info("Training pipeline was deleted.")
except NotFound:
self.log.info("The Training Pipeline ID %s does not exist.", self.training_pipeline)
class CreateAutoMLForecastingTrainingJobOperator(AutoMLTrainingJobBaseOperator):
"""Create AutoML Forecasting Training job"""
template_fields = [
'region',
'impersonation_chain',
]
operator_extra_links = (VertexAIModelLink(),)
def __init__(
self,
*,
dataset_id: str,
target_column: str,
time_column: str,
time_series_identifier_column: str,
unavailable_at_forecast_columns: List[str],
available_at_forecast_columns: List[str],
forecast_horizon: int,
data_granularity_unit: str,
data_granularity_count: int,
optimization_objective: Optional[str] = None,
column_specs: Optional[Dict[str, str]] = None,
column_transformations: Optional[List[Dict[str, Dict[str, str]]]] = None,
validation_fraction_split: Optional[float] = None,
predefined_split_column_name: Optional[str] = None,
weight_column: Optional[str] = None,
time_series_attribute_columns: Optional[List[str]] = None,
context_window: Optional[int] = None,
export_evaluated_data_items: bool = False,
export_evaluated_data_items_bigquery_destination_uri: Optional[str] = None,
export_evaluated_data_items_override_destination: bool = False,
quantiles: Optional[List[float]] = None,
validation_options: Optional[str] = None,
budget_milli_node_hours: int = 1000,
**kwargs,
) -> None:
super().__init__(**kwargs)
self.dataset_id = dataset_id
self.target_column = target_column
self.time_column = time_column
self.time_series_identifier_column = time_series_identifier_column
self.unavailable_at_forecast_columns = unavailable_at_forecast_columns
self.available_at_forecast_columns = available_at_forecast_columns
self.forecast_horizon = forecast_horizon
self.data_granularity_unit = data_granularity_unit
self.data_granularity_count = data_granularity_count
self.optimization_objective = optimization_objective
self.column_specs = column_specs
self.column_transformations = column_transformations
self.validation_fraction_split = validation_fraction_split
self.predefined_split_column_name = predefined_split_column_name
self.weight_column = weight_column
self.time_series_attribute_columns = time_series_attribute_columns
self.context_window = context_window
self.export_evaluated_data_items = export_evaluated_data_items
self.export_evaluated_data_items_bigquery_destination_uri = (
export_evaluated_data_items_bigquery_destination_uri
)
self.export_evaluated_data_items_override_destination = (
export_evaluated_data_items_override_destination
)
self.quantiles = quantiles
self.validation_options = validation_options
self.budget_milli_node_hours = budget_milli_node_hours
def execute(self, context: "Context"):
self.hook = AutoMLHook(
gcp_conn_id=self.gcp_conn_id,
delegate_to=self.delegate_to,
impersonation_chain=self.impersonation_chain,
)
model = self.hook.create_auto_ml_forecasting_training_job(
project_id=self.project_id,
region=self.region,
display_name=self.display_name,
dataset=datasets.TimeSeriesDataset(dataset_name=self.dataset_id),
target_column=self.target_column,
time_column=self.time_column,
time_series_identifier_column=self.time_series_identifier_column,
unavailable_at_forecast_columns=self.unavailable_at_forecast_columns,
available_at_forecast_columns=self.available_at_forecast_columns,
forecast_horizon=self.forecast_horizon,
data_granularity_unit=self.data_granularity_unit,
data_granularity_count=self.data_granularity_count,
optimization_objective=self.optimization_objective,
column_specs=self.column_specs,
column_transformations=self.column_transformations,
labels=self.labels,
training_encryption_spec_key_name=self.training_encryption_spec_key_name,
model_encryption_spec_key_name=self.model_encryption_spec_key_name,
training_fraction_split=self.training_fraction_split,
validation_fraction_split=self.validation_fraction_split,
test_fraction_split=self.test_fraction_split,
predefined_split_column_name=self.predefined_split_column_name,
weight_column=self.weight_column,
time_series_attribute_columns=self.time_series_attribute_columns,
context_window=self.context_window,
export_evaluated_data_items=self.export_evaluated_data_items,
export_evaluated_data_items_bigquery_destination_uri=(
self.export_evaluated_data_items_bigquery_destination_uri
),
export_evaluated_data_items_override_destination=(
self.export_evaluated_data_items_override_destination
),
quantiles=self.quantiles,
validation_options=self.validation_options,
budget_milli_node_hours=self.budget_milli_node_hours,
model_display_name=self.model_display_name,
model_labels=self.model_labels,
sync=self.sync,
)
result = Model.to_dict(model)
model_id = self.hook.extract_model_id(result)
VertexAIModelLink.persist(context=context, task_instance=self, model_id=model_id)
return result
class CreateAutoMLTabularTrainingJobOperator(AutoMLTrainingJobBaseOperator):
"""Create Auto ML Tabular Training job"""
template_fields = [
'region',
'impersonation_chain',
]
operator_extra_links = (VertexAIModelLink(),)
def __init__(
self,
*,
dataset_id: str,
target_column: str,
optimization_prediction_type: str,
optimization_objective: Optional[str] = None,
column_specs: Optional[Dict[str, str]] = None,
column_transformations: Optional[List[Dict[str, Dict[str, str]]]] = None,
optimization_objective_recall_value: Optional[float] = None,
optimization_objective_precision_value: Optional[float] = None,
validation_fraction_split: Optional[float] = None,
predefined_split_column_name: Optional[str] = None,
timestamp_split_column_name: Optional[str] = None,
weight_column: Optional[str] = None,
budget_milli_node_hours: int = 1000,
disable_early_stopping: bool = False,
export_evaluated_data_items: bool = False,
export_evaluated_data_items_bigquery_destination_uri: Optional[str] = None,
export_evaluated_data_items_override_destination: bool = False,
**kwargs,
) -> None:
super().__init__(**kwargs)
self.dataset_id = dataset_id
self.target_column = target_column
self.optimization_prediction_type = optimization_prediction_type
self.optimization_objective = optimization_objective
self.column_specs = column_specs
self.column_transformations = column_transformations
self.optimization_objective_recall_value = optimization_objective_recall_value
self.optimization_objective_precision_value = optimization_objective_precision_value
self.validation_fraction_split = validation_fraction_split
self.predefined_split_column_name = predefined_split_column_name
self.timestamp_split_column_name = timestamp_split_column_name
self.weight_column = weight_column
self.budget_milli_node_hours = budget_milli_node_hours
self.disable_early_stopping = disable_early_stopping
self.export_evaluated_data_items = export_evaluated_data_items
self.export_evaluated_data_items_bigquery_destination_uri = (
export_evaluated_data_items_bigquery_destination_uri
)
self.export_evaluated_data_items_override_destination = (
export_evaluated_data_items_override_destination
)
def execute(self, context: "Context"):
self.hook = AutoMLHook(
gcp_conn_id=self.gcp_conn_id,
delegate_to=self.delegate_to,
impersonation_chain=self.impersonation_chain,
)
model = self.hook.create_auto_ml_tabular_training_job(
project_id=self.project_id,
region=self.region,
display_name=self.display_name,
dataset=datasets.TabularDataset(dataset_name=self.dataset_id),
target_column=self.target_column,
optimization_prediction_type=self.optimization_prediction_type,
optimization_objective=self.optimization_objective,
column_specs=self.column_specs,
column_transformations=self.column_transformations,
optimization_objective_recall_value=self.optimization_objective_recall_value,
optimization_objective_precision_value=self.optimization_objective_precision_value,
labels=self.labels,
training_encryption_spec_key_name=self.training_encryption_spec_key_name,
model_encryption_spec_key_name=self.model_encryption_spec_key_name,
training_fraction_split=self.training_fraction_split,
validation_fraction_split=self.validation_fraction_split,
test_fraction_split=self.test_fraction_split,
predefined_split_column_name=self.predefined_split_column_name,
timestamp_split_column_name=self.timestamp_split_column_name,
weight_column=self.weight_column,
budget_milli_node_hours=self.budget_milli_node_hours,
model_display_name=self.model_display_name,
model_labels=self.model_labels,
disable_early_stopping=self.disable_early_stopping,
export_evaluated_data_items=self.export_evaluated_data_items,
export_evaluated_data_items_bigquery_destination_uri=(
self.export_evaluated_data_items_bigquery_destination_uri
),
export_evaluated_data_items_override_destination=(
self.export_evaluated_data_items_override_destination
),
sync=self.sync,
)
result = Model.to_dict(model)
model_id = self.hook.extract_model_id(result)
VertexAIModelLink.persist(context=context, task_instance=self, model_id=model_id)
return result
class CreateAutoMLImageTrainingJobOperator(AutoMLTrainingJobBaseOperator):
"""Create Auto ML Image Training job"""
template_fields = [
'region',
'impersonation_chain',
]
operator_extra_links = (VertexAIModelLink(),)
def __init__(
self,
*,
dataset_id: str,
prediction_type: str = "classification",
multi_label: bool = False,
model_type: str = "CLOUD",
base_model: Optional[Model] = None,
validation_fraction_split: Optional[float] = None,
training_filter_split: Optional[str] = None,
validation_filter_split: Optional[str] = None,
test_filter_split: Optional[str] = None,
budget_milli_node_hours: Optional[int] = None,
disable_early_stopping: bool = False,
**kwargs,
) -> None:
super().__init__(**kwargs)
self.dataset_id = dataset_id
self.prediction_type = prediction_type
self.multi_label = multi_label
self.model_type = model_type
self.base_model = base_model
self.validation_fraction_split = validation_fraction_split
self.training_filter_split = training_filter_split
self.validation_filter_split = validation_filter_split
self.test_filter_split = test_filter_split
self.budget_milli_node_hours = budget_milli_node_hours
self.disable_early_stopping = disable_early_stopping
def execute(self, context: "Context"):
self.hook = AutoMLHook(
gcp_conn_id=self.gcp_conn_id,
delegate_to=self.delegate_to,
impersonation_chain=self.impersonation_chain,
)
model = self.hook.create_auto_ml_image_training_job(
project_id=self.project_id,
region=self.region,
display_name=self.display_name,
dataset=datasets.ImageDataset(dataset_name=self.dataset_id),
prediction_type=self.prediction_type,
multi_label=self.multi_label,
model_type=self.model_type,
base_model=self.base_model,
labels=self.labels,
training_encryption_spec_key_name=self.training_encryption_spec_key_name,
model_encryption_spec_key_name=self.model_encryption_spec_key_name,
training_fraction_split=self.training_fraction_split,
validation_fraction_split=self.validation_fraction_split,
test_fraction_split=self.test_fraction_split,
training_filter_split=self.training_filter_split,
validation_filter_split=self.validation_filter_split,
test_filter_split=self.test_filter_split,
budget_milli_node_hours=self.budget_milli_node_hours,
model_display_name=self.model_display_name,
model_labels=self.model_labels,
disable_early_stopping=self.disable_early_stopping,
sync=self.sync,
)
result = Model.to_dict(model)
model_id = self.hook.extract_model_id(result)
VertexAIModelLink.persist(context=context, task_instance=self, model_id=model_id)
return result
