def test_stream_train_component(self):
batch_input_example_meta = af.register_example(
name='stream_train_example',
support_type=ExampleSupportType.EXAMPLE_BOTH)
model_meta = af.register_model(model_name='mnist_model',
model_type=ModelType.SAVED_MODEL)
stream_input_example_meta = af.register_example(
name='stream_train_example',
support_type=ExampleSupportType.EXAMPLE_BOTH)
with af.config(
af.BaseJobConfig(platform='local',
engine='python',
job_name='stream_train')):
batch_input_example = af.read_example(
example_info=batch_input_example_meta,
executor=PythonObjectExecutor(
python_object=ReadBatchExample()))
batch_train = af.train(input_data_list=[batch_input_example],
executor=PythonObjectExecutor(
python_object=TrainBatchMnistModel()),
model_info=model_meta)
stream_input_example = af.read_example(
example_info=stream_input_example_meta,
executor=PythonObjectExecutor(
python_object=ReadStreamExample()))
stream_train = af.train(input_data_list=[stream_input_example],
executor=PythonObjectExecutor(
python_object=TrainStreamMnistModel()),
model_info=model_meta)
af.stop_before_control_dependency(stream_train, batch_train)
workflow_id = af.run(test_util.get_project_path())
res = af.wait_workflow_execution_finished(workflow_id)
self.assertEqual(0, res)
def test_stream_evaluate_component(self):
input_example_meta = af.register_example(
name='batch_train_example',
support_type=ExampleSupportType.EXAMPLE_BATCH)
model_meta = af.register_model(model_name='mnist_model',
model_type=ModelType.SAVED_MODEL)
stream_evaluate_example_meta = af.register_example(
name='stream_evaluate_example',
support_type=ExampleSupportType.EXAMPLE_STREAM)
stream_output_file = get_file_dir(__file__) + '/stream_evaluate'
evaluate_output = af.register_artifact(name='stream_evaluate',
stream_uri=stream_output_file)
stream_evaluate_result_example_meta = af.register_example(
name='stream_evaluate_result_example',
support_type=ExampleSupportType.EXAMPLE_STREAM,
stream_uri=stream_output_file)
if os.path.exists(stream_output_file):
os.remove(stream_output_file)
with af.config(
af.BaseJobConfig(platform='local',
engine='python',
job_name='stream_evaluate')):
input_example = af.read_example(
example_info=input_example_meta,
executor=PythonObjectExecutor(
python_object=ReadBatchExample()))
batch_train = af.train(input_data_list=[input_example],
executor=PythonObjectExecutor(
python_object=TrainBatchMnistModel()),
model_info=model_meta)
stream_evaluate_example = af.read_example(
example_info=stream_evaluate_example_meta,
executor=PythonObjectExecutor(
python_object=ReadStreamExample()))
stream_evaluate = af.evaluate(
input_data_list=[stream_evaluate_example],
model_info=model_meta,
executor=PythonObjectExecutor(
python_object=EvaluateStreamMnistModel()),
output_num=1)
af.write_example(input_data=stream_evaluate,
example_info=stream_evaluate_result_example_meta,
executor=PythonObjectExecutor(
python_object=WriteStreamExample()))
af.stop_before_control_dependency(stream_evaluate, batch_train)
workflow_id = af.run(test_util.get_project_path())
res = af.wait_workflow_execution_finished(workflow_id)
self.assertEqual(0, res)
def test_stream_transform_component(self):
file = get_file_dir(__file__) + '/test1.csv'
input_example_meta = af.register_example(
name='test_example',
support_type=ExampleSupportType.EXAMPLE_BOTH,
stream_uri=file)
output_file = get_file_dir(
__file__) + "/output_transform_stream_test1.csv"
output_example_meta = af.register_example(
name='test_example_output',
support_type=ExampleSupportType.EXAMPLE_BOTH,
stream_uri=output_file)
with af.config(
af.BaseJobConfig(platform='local',
engine='python',
job_name='stream_transform')):
input_example = af.read_example(
example_info=input_example_meta,
executor=PythonObjectExecutor(
python_object=ReadStreamExample()))
transform_example = af.transform(
input_data_list=[input_example],
executor=PythonObjectExecutor(
python_object=TransformStreamData()))
af.write_example(input_data=transform_example,
example_info=output_example_meta.name,
executor=PythonObjectExecutor(
python_object=WriteStreamExample()))
workflow_id = af.run(test_util.get_project_path())
res = af.wait_workflow_execution_finished(workflow_id)
self.assertEqual(0, res)
def run_task_function(client: NotificationClient):
with af.global_config_file(workflow_config_file()):
with af.config(task_config):
input_example = af.read_example(
example_info=example_1,
executor=faf.flink_executor.FlinkPythonExecutor(
python_object=Source()))
processed = af.transform(
input_data_list=[input_example],
executor=faf.flink_executor.FlinkPythonExecutor(
python_object=Transformer()))
af.write_example(
input_data=processed,
example_info=example_2,
executor=faf.flink_executor.FlinkPythonExecutor(
python_object=Sink()))
workflow_info = af.workflow_operation.submit_workflow(
workflow_name)
af.workflow_operation.start_new_workflow_execution(workflow_name)
while True:
with create_session() as session:
dag_run = session.query(DagRun).filter(
DagRun.dag_id == 'test_project.test_workflow').first()
if dag_run is not None and dag_run.state in State.finished:
break
else:
time.sleep(1)
def test_read_example_with_numpy_npy(self):
npy_name = 'test.npy'
np.save(file=npy_name, arr=np.arange(10))
input_example_meta = af.register_example(
name='input_numpy_example',
data_type='numpy',
data_format='npy',
support_type=ExampleSupportType.EXAMPLE_BATCH,
batch_uri=os.path.abspath(
os.path.dirname(__file__) + "/" + npy_name))
output_example_meta = af.register_example(
name='ouput_numpy_example',
data_type='numpy',
data_format='npy',
support_type=ExampleSupportType.EXAMPLE_BATCH,
batch_uri=os.path.abspath(
os.path.dirname(__file__) + '/numpy_output.npy'))
with af.config(
af.BaseJobConfig(platform='local',
engine='python',
job_name='test_npy')):
example_channel = af.read_example(example_info=input_example_meta)
af.write_example(input_data=example_channel,
example_info=output_example_meta)
workflow_id = af.run(test_util.get_project_path())
res = af.wait_workflow_execution_finished(workflow_id)
self.assertEqual(0, res)
def test_read_example_with_pandas(self):
input_example_meta = af.register_example(
name='input_pandas_example',
data_type='pandas',
data_format='csv',
support_type=ExampleSupportType.EXAMPLE_BATCH,
batch_uri=os.path.abspath(
os.path.dirname(__file__) + '/test1.csv'))
output_example_meta = af.register_example(
name='ouput_pandas_example',
data_type='pandas',
data_format='csv',
support_type=ExampleSupportType.EXAMPLE_BATCH,
batch_uri=os.path.abspath(
os.path.dirname(__file__) + '/pandas_output.csv'))
with af.config(
af.BaseJobConfig(platform='local',
engine='python',
job_name='test_csv')):
example_channel = af.read_example(example_info=input_example_meta)
af.write_example(input_data=example_channel,
example_info=output_example_meta)
workflow_id = af.run(test_util.get_project_path())
res = af.wait_workflow_execution_finished(workflow_id)
self.assertEqual(0, res)
def test_batch_train_component_with_an_output(self):
input_example_meta = af.register_example(
name='batch_train_example',
support_type=ExampleSupportType.EXAMPLE_BATCH)
model_meta = af.register_model(model_name='mnist_model',
model_type=ModelType.SAVED_MODEL)
example_meta = af.register_example(
name='output_example',
support_type=ExampleSupportType.EXAMPLE_BATCH,
data_type='numpy',
data_format='npz',
batch_uri=os.path.abspath(
os.path.dirname(__file__) + '/numpy_output.npz'))
with af.config(
af.BaseJobConfig(platform='local',
engine='python',
job_name='batch_train')):
input_example = af.read_example(
example_info=input_example_meta,
executor=PythonObjectExecutor(
python_object=ReadBatchExample()))
train_channel = af.train(
input_data_list=[input_example],
executor=PythonObjectExecutor(
python_object=TrainBatchMnistModelWithOutput()),
model_info=model_meta,
output_num=1)
af.write_example(input_data=train_channel,
example_info=example_meta)
workflow_id = af.run(test_util.get_project_path())
res = af.wait_workflow_execution_finished(workflow_id)
self.assertEqual(0, res)
def run_flink_job():
input_file = "/test1.csv"
output_file ="/output_test1.csv"
if os.path.exists(output_file):
os.remove(output_file)
example_1 = af.create_example(name="example_1",
support_type=af.ExampleSupportType.EXAMPLE_BOTH,
batch_uri=input_file,
stream_uri=input_file,
data_format="csv")
example_2 = af.create_example(name="example_2",
support_type=af.ExampleSupportType.EXAMPLE_BOTH,
batch_uri=output_file,
stream_uri=output_file,
data_format="csv")
flink_config = faf.LocalFlinkJobConfig()
flink_config.flink_home = "/Users/chenwuchao/soft/apache/flink-1.10.0"
with af.config(flink_config):
batch_args_1: Properties = {}
ddl = """CREATE TABLE input_table (a STRING, b STRING, c STRING) WITH ('connector' = 'filesystem',
'path' = 'INPUT',
'format' = 'csv'
)"""
table_name = "input_table"
batch_args_1['ddl'] = ddl
batch_args_1['table_name'] = table_name
stream_args_1 = batch_args_1
batch_args_2: Properties = {}
ddl = """CREATE TABLE output_table (aa STRING, bb STRING) WITH ('connector' = 'filesystem',
'path' = 'OUTPUT',
'format' = 'csv'
)"""
table_name = "output_table"
batch_args_2['ddl'] = ddl
batch_args_2['table_name'] = table_name
stream_args_2 = batch_args_2
input_example = af.read_example(example_info=example_1,
exec_args=ExecuteProperties(
batch_properties=batch_args_1,
stream_properties=stream_args_1)
)
processed = af.transform(input_data_list=[input_example],
executor=faf.FlinkJavaExecutor(
java_class="com.apache.flink.ai.flow.TestTransformer"))
af.write_example(input_data=processed,
example_info=example_2,
exec_args=ExecuteProperties(
batch_properties=batch_args_2,
stream_properties=stream_args_2)
)
workflow = af.compile_workflow(test_util.get_project_path())
print(json_utils.dumps(list(workflow.jobs.values())[0]))
def run_project(project_root_path):
af.set_project_config_file(project_root_path + '/project.yaml')
# Config command line job, we set platform to local and engine to cmd_line here
cmd_job_config = af.BaseJobConfig(platform=LocalPlatform.platform(), engine=CMDEngine().engine())
with af.config(cmd_job_config):
# Command line job executor
cmd_job = af.user_define_operation(executor=CmdExecutor(cmd_line="echo Start AI flow"))
# Config python job, we set platform to local and engine to python here
python_job_config = af.BaseJobConfig(platform=LocalPlatform.platform(), engine=PythonEngine.engine())
# Set execution mode of this python job to BATCH,
# which indicates jobs with this config is running in the form of batch.
python_job_config.exec_mode = af.ExecutionMode.BATCH
with af.config(python_job_config):
# Path of Source data(under '..../simple_transform_airflow' dir)
source_path = os.path.dirname(os.path.abspath(__file__)) + '/source_data.csv'
# Path of Sink data
sink_path = os.path.dirname(os.path.abspath(__file__)) + '/sink_data.csv'
# To make the project replaceable, we register the example in metadata service
read_example_meta = af.register_example(name='read_example', support_type=ExampleSupportType.EXAMPLE_BATCH,
data_format='csv', data_type='pandas', batch_uri=source_path)
# Read training example using af.read_example()
# example_info is the meta information of the example
read_example_channel = af.read_example(example_info=read_example_meta, exec_args=ExecuteArgs(
batch_properties=Args(header=None, names=["a", "b", "c"])))
# Transform examples using af.transform()
transform_channel = af.transform(input_data_list=[read_example_channel],
executor=PythonObjectExecutor(python_object=SimpleTransform()))
write_example_meta = af.register_example(name='write_example', support_type=ExampleSupportType.EXAMPLE_BATCH,
data_format='csv', data_type='pandas', batch_uri=sink_path)
# Write example to specific path
write = af.write_example(input_data=transform_channel, example_info=write_example_meta,
exec_args=ExecuteArgs(batch_properties=Args(sep=',', header=False, index=False)))
# Add control dependency, which means read_example job will start right after command line job finishes.
af.stop_before_control_dependency(read_example_channel, cmd_job)
transform_dag = 'simple_transform'
af.deploy_to_airflow(project_root_path, dag_id=transform_dag)
context = af.run(project_path=project_root_path,
dag_id=transform_dag,
scheduler_type=SchedulerType.AIRFLOW)
def test_run_pyflink_job(self):
project_path = os.path.dirname(__file__) + '/../'
af.set_project_config_file(project_path + "project.yaml")
input_file = get_parent_dir(
get_file_dir(__file__)) + '/resources/word_count.txt'
output_file = get_file_dir(__file__) + "/word_count_output.csv"
if os.path.exists(output_file):
os.remove(output_file)
example_1 = af.create_example(
name="example_1",
support_type=af.ExampleSupportType.EXAMPLE_BOTH,
batch_uri=input_file,
stream_uri=input_file,
data_format="csv")
example_2 = af.create_example(
name="example_2",
support_type=af.ExampleSupportType.EXAMPLE_BOTH,
batch_uri=output_file,
stream_uri=output_file,
data_format="csv")
flink_config = faf.LocalFlinkJobConfig()
flink_config.local_mode = 'cluster'
flink_config.flink_home = '/Users/chenwuchao/soft/apache/flink-1.11.0/'
flink_config.set_table_env_create_func(TableEnvCreator())
with af.config(flink_config):
input_example = af.read_example(
example_info=example_1,
executor=faf.flink_executor.FlinkPythonExecutor(
python_object=Source()))
processed = af.transform(
input_data_list=[input_example],
executor=faf.flink_executor.FlinkPythonExecutor(
python_object=Transformer()))
af.write_example(input_data=processed,
example_info=example_2,
executor=faf.flink_executor.FlinkPythonExecutor(
python_object=Sink()))
workflow_id = af.run(project_path)
res = af.wait_workflow_execution_finished(workflow_id)
def test_batch_predict_component(self):
input_example_meta = af.register_example(
name='input_train_example',
support_type=ExampleSupportType.EXAMPLE_BOTH)
model_meta = af.register_model(model_name='mnist_model',
model_type=ModelType.SAVED_MODEL)
batch_output_file = get_file_dir(__file__) + '/batch_predict'
evaluate_output = af.register_artifact(name='batch_evaluate',
batch_uri=batch_output_file)
output_example_meta = af.register_example(
name='output_result_example',
support_type=ExampleSupportType.EXAMPLE_BATCH,
data_type='numpy',
data_format='txt',
batch_uri=batch_output_file)
if os.path.exists(batch_output_file):
os.remove(batch_output_file)
with af.config(
af.BaseJobConfig(platform='local',
engine='python',
job_name='batch_predict')):
batch_example = af.read_example(
example_info=input_example_meta,
executor=PythonObjectExecutor(
python_object=ReadBatchExample()))
batch_train = af.train(input_data_list=[batch_example],
executor=PythonObjectExecutor(
python_object=TrainBatchMnistModel()),
model_info=model_meta)
batch_predict = af.predict(
input_data_list=[batch_example],
model_info=model_meta,
executor=PythonObjectExecutor(
python_object=PredictBatchMnistModel()),
output_num=1)
af.write_example(input_data=batch_predict,
example_info=output_example_meta)
af.stop_before_control_dependency(batch_predict, batch_train)
workflow_id = af.run(test_util.get_project_path())
res = af.wait_workflow_execution_finished(workflow_id)
self.assertEqual(0, res)
def test_read_example_with_numpy_npz(self):
npy_name = 'test.npz'
np.savez(npy_name, np.arange(10), np.sin(np.arange(10)))
input_example_meta = af.register_example(
name='input_numpy_example',
data_type='numpy',
data_format='npz',
support_type=ExampleSupportType.EXAMPLE_BATCH,
batch_uri=os.path.abspath(
os.path.dirname(__file__) + "/" + npy_name))
output_example_meta_first = af.register_example(
name='ouput_numpy_example_1',
data_type='numpy',
data_format='npz',
support_type=ExampleSupportType.EXAMPLE_BATCH,
batch_uri=os.path.abspath(
os.path.dirname(__file__) + '/numpy_output_1.npz'))
output_example_meta_second = af.register_example(
name='ouput_numpy_example_2',
data_type='numpy',
data_format='npz',
support_type=ExampleSupportType.EXAMPLE_BATCH,
batch_uri=os.path.abspath(
os.path.dirname(__file__) + '/numpy_output_2.npz'))
with af.config(
af.BaseJobConfig(platform='local',
engine='python',
job_name='test_npz')):
example_channel = af.read_example(example_info=input_example_meta)
transform_channel = af.transform(
input_data_list=[example_channel],
executor=PythonObjectExecutor(
python_object=TransformTrainData()),
output_num=2)
af.write_example(input_data=transform_channel[0],
example_info=output_example_meta_first)
af.write_example(input_data=transform_channel[1],
example_info=output_example_meta_second)
workflow_id = af.run(test_util.get_project_path())
res = af.wait_workflow_execution_finished(workflow_id)
self.assertEqual(0, res)
def test_batch_train_component(self):
input_example_meta = af.register_example(
name='batch_train_example',
support_type=ExampleSupportType.EXAMPLE_BATCH)
model_meta = af.register_model(model_name='mnist_model',
model_type=ModelType.SAVED_MODEL)
with af.config(
af.BaseJobConfig(platform='local',
engine='python',
job_name='batch_train')):
input_example = af.read_example(
example_info=input_example_meta,
executor=PythonObjectExecutor(
python_object=ReadBatchExample()))
af.train(input_data_list=[input_example],
executor=PythonObjectExecutor(
python_object=TrainBatchMnistModel()),
model_info=model_meta)
workflow_id = af.run(test_util.get_project_path())
res = af.wait_workflow_execution_finished(workflow_id)
self.assertEqual(0, res)
def test_batch_model_validate(self):
input_example_meta = af.register_example(name='batch_train_example',
support_type=ExampleSupportType.EXAMPLE_BOTH)
model_meta = af.register_model(model_name='mnist_model',
model_type=ModelType.SAVED_MODEL)
with af.config(af.BaseJobConfig(platform='local', engine='python', job_name='evaluate')):
input_example = af.read_example(example_info=input_example_meta,
executor=PythonObjectExecutor(python_object=ReadBatchExample()))
batch_train = af.train(input_data_list=[input_example],
executor=PythonObjectExecutor(python_object=TrainBatchMnistModel()),
model_info=model_meta)
model_validate = af.model_validate(input_data_list=[input_example],
model_info=model_meta,
executor=PythonObjectExecutor(python_object=BatchModelValidate()),
output_num=0)
af.stop_before_control_dependency(model_validate, batch_train)
workflow_id = af.run(test_util.get_project_path())
res = af.wait_workflow_execution_finished(workflow_id)
self.assertEqual(0, res)
def test_run_pyflink_job(self):
input_file = get_parent_dir(
get_file_dir(__file__)) + '/resources/word_count.txt'
output_file = get_file_dir(__file__) + "/word_count_output.csv"
if os.path.exists(output_file):
os.remove(output_file)
example_1 = af.create_example(
name="example_1",
support_type=af.ExampleSupportType.EXAMPLE_BOTH,
batch_uri=input_file,
stream_uri=input_file,
data_format="csv")
example_2 = af.create_example(
name="example_2",
support_type=af.ExampleSupportType.EXAMPLE_BOTH,
batch_uri=output_file,
stream_uri=output_file,
data_format="csv")
flink_config = faf.LocalFlinkJobConfig()
flink_config.local_mode = 'python'
flink_config.set_table_env_create_func(TableEnvCreator())
with af.config(flink_config):
input_example = af.read_example(
example_info=example_1,
executor=faf.flink_executor.FlinkPythonExecutor(
python_object=Source()))
processed = af.transform(
input_data_list=[input_example],
executor=faf.flink_executor.FlinkPythonExecutor(
python_object=Transformer()))
af.write_example(input_data=processed,
example_info=example_2,
executor=faf.flink_executor.FlinkPythonExecutor(
python_object=Sink()))
workflow_id = af.run(test_util.get_project_path())
res = af.wait_workflow_execution_finished(workflow_id)
self.assertEqual(0, res)
def run_project(project_root_path):
af.set_project_config_file(project_root_path + "/project.yaml")
project_name = af.project_config().get_project_name()
artifact_prefix = project_name + "."
validate_trigger = af.external_trigger(name='validate')
push_trigger = af.external_trigger(name='push')
with af.global_config_file(project_root_path + '/resources/workflow_config.yaml'):
with af.config('train_job'):
train_example = af.register_example(name=artifact_prefix + 'train_example',
support_type=ExampleSupportType.EXAMPLE_STREAM,
stream_uri=EXAMPLE_URI.format('train'))
train_read_example = af.read_example(example_info=train_example,
executor=PythonObjectExecutor(python_object=TrainExampleReader()))
train_transform = af.transform(input_data_list=[train_read_example],
executor=PythonObjectExecutor(python_object=TrainExampleTransformer()))
train_model = af.register_model(model_name=artifact_prefix + 'logistic-regression',
model_type=ModelType.SAVED_MODEL,
model_desc='logistic regression model')
train_channel = af.train(input_data_list=[train_transform],
executor=PythonObjectExecutor(python_object=ModelTrainer()),
model_info=train_model)
with af.config('validate_job'):
validate_example = af.register_example(name=artifact_prefix + 'validate_example',
support_type=ExampleSupportType.EXAMPLE_STREAM,
stream_uri=EXAMPLE_URI.format('evaluate'),
data_format='npz')
validate_read_example = af.read_example(example_info=validate_example,
executor=PythonObjectExecutor(
python_object=ValidateExampleReader()))
validate_transform = af.transform(input_data_list=[validate_read_example],
executor=PythonObjectExecutor(python_object=ValidateTransformer()))
validate_artifact_name = artifact_prefix + 'validate_artifact'
validate_artifact = af.register_artifact(name=validate_artifact_name,
stream_uri=get_file_dir(__file__) + '/validate_result')
validate_channel = af.model_validate(input_data_list=[validate_transform],
model_info=train_model,
executor=PythonObjectExecutor(
python_object=ModelValidator(validate_artifact_name)),
)
with af.config('push_job'):
# Push model to serving
# Register metadata of pushed model
push_model_artifact_name = artifact_prefix + 'push_model_artifact'
push_model_artifact = af.register_artifact(name=push_model_artifact_name,
stream_uri=get_file_dir(__file__) + '/pushed_model')
push_channel = af.push_model(model_info=train_model,
executor=PythonObjectExecutor(
python_object=ModelPusher(push_model_artifact_name)))
with af.config('predict_job'):
predict_example = af.register_example(name=artifact_prefix + 'predict_example',
support_type=ExampleSupportType.EXAMPLE_STREAM,
stream_uri=EXAMPLE_URI.format('predict'))
predict_read_example = af.read_example(example_info=predict_example,
executor=PythonObjectExecutor(python_object=PredictExampleReader()))
predict_transform = af.transform(input_data_list=[predict_read_example],
executor=PythonObjectExecutor(python_object=PredictTransformer()))
predict_channel = af.predict(input_data_list=[predict_transform],
model_info=train_model,
executor=PythonObjectExecutor(python_object=ModelPredictor()))
write_example = af.register_example(name=artifact_prefix + 'export_example',
support_type=ExampleSupportType.EXAMPLE_STREAM,
stream_uri=get_file_dir(__file__) + '/predict_result')
af.write_example(input_data=predict_channel,
example_info=write_example,
executor=PythonObjectExecutor(python_object=ExampleWriter()))
af.model_version_control_dependency(src=validate_channel,
model_version_event_type=ModelVersionEventType.MODEL_GENERATED,
dependency=validate_trigger, model_name=train_model.name)
af.model_version_control_dependency(src=push_channel,
model_version_event_type=ModelVersionEventType.MODEL_VALIDATED,
dependency=push_trigger, model_name=train_model.name)
# Run workflow
transform_dag = project_name
af.deploy_to_airflow(project_root_path, dag_id=transform_dag)
af.run(project_path=project_root_path,
dag_id=transform_dag,
scheduler_type=SchedulerType.AIRFLOW)
def run_workflow():
"""
Run the user-defined workflow definition.
"""
train_example_meta, label_example_meta, test_example_meta, test_output_example_meta, train_model_meta = prepare_workflow(
)
python_job_config_0 = BaseJobConfig(job_name='read_train',
platform='local',
engine='python')
python_job_config_1 = BaseJobConfig(job_name='train',
platform='local',
engine='python')
flink_job_config_2 = LocalFlinkJobConfig()
flink_job_config_2.job_name = 'test'
flink_job_config_2.local_mode = 'python'
flink_job_config_2.flink_home = os.environ['FLINK_HOME']
flink_job_config_2.set_table_env_create_func(MyStreamTableEnvCreator())
with af.config(python_job_config_0):
python_job_0_read_train_data = af.read_example(
example_info=train_example_meta,
executor=PythonObjectExecutor(python_object=ReadTrainCsvExample()))
python_job_0_read_label_data = af.read_example(
example_info=label_example_meta,
executor=PythonObjectExecutor(python_object=ReadLabelCsvExample()))
write_train_data_example = af.register_example(
name='write_train_data',
support_type=ExampleSupportType.EXAMPLE_BATCH,
data_type='pandas',
data_format='csv',
batch_uri='/tmp/write_train_data.csv')
python_job_0_write_train_result = af.write_example(
input_data=python_job_0_read_train_data,
example_info=write_train_data_example,
executor=PythonObjectExecutor(
python_object=WriteTrainCsvExample()))
with af.config(python_job_config_1):
python_job_1_train_model = af.train(
name='trainer_0',
input_data_list=[
python_job_0_read_train_data, python_job_0_read_label_data
],
executor=PythonObjectExecutor(python_object=TrainModel()),
model_info=train_model_meta)
with af.config(flink_job_config_2):
flink_job_2_read_test_data = af.read_example(
example_info=test_example_meta,
executor=FlinkPythonExecutor(python_object=ReadTestCsvExample()))
flink_job_2_predict_test_data = af.transform(
input_data_list=[flink_job_2_read_test_data],
executor=FlinkPythonExecutor(
python_object=PredictTestLabelExecutor()))
write_result = af.write_example(
input_data=flink_job_2_predict_test_data,
example_info=test_output_example_meta,
executor=FlinkPythonExecutor(
python_object=WritePredictTestExample()))
af.stop_before_control_dependency(python_job_1_train_model,
python_job_0_write_train_result)
af.stop_before_control_dependency(write_result, python_job_1_train_model)
workflow_id = af.run(get_project_path() + '/')
res = af.wait_workflow_execution_finished(workflow_id)
sys.exit(res)
def run_project(project_root_path):
af.set_project_config_file(project_root_path + "/project.yaml")
project_name = af.project_config().get_project_name()
artifact_prefix = project_name + "."
validate_trigger = af.external_trigger(name='validate')
push_trigger = af.external_trigger(name='push')
with af.global_config_file(project_root_path + '/resources/workflow_config.yaml'):
# the config of train job is a periodic job which means it will
# run every `interval`(defined in workflow_config.yaml) seconds
with af.config('train_job'):
# Register metadata raw training data(example) and read example(i.e. training dataset)
train_example = af.register_example(name=artifact_prefix + 'train_example',
support_type=ExampleSupportType.EXAMPLE_BATCH,
batch_uri=EXAMPLE_URI.format('train'))
train_read_example = af.read_example(example_info=train_example,
executor=PythonObjectExecutor(python_object=ExampleReader()))
# Transform(preprocessing) example
train_transform = af.transform(input_data_list=[train_read_example],
executor=PythonObjectExecutor(python_object=ExampleTransformer()))
# Register model metadata and train model
train_model = af.register_model(model_name=artifact_prefix + 'logistic-regression',
model_type=ModelType.SAVED_MODEL,
model_desc='logistic regression model')
train_channel = af.train(input_data_list=[train_transform],
executor=PythonObjectExecutor(python_object=ModelTrainer()),
model_info=train_model)
with af.config('validate_job'):
# Validation of model
# Read validation dataset and validate model before it is used to predict
validate_example = af.register_example(name=artifact_prefix + 'validate_example',
support_type=ExampleSupportType.EXAMPLE_STREAM,
batch_uri=EXAMPLE_URI.format('evaluate'))
validate_read_example = af.read_example(example_info=validate_example,
executor=PythonObjectExecutor(
python_object=ValidateExampleReader()))
validate_transform = af.transform(input_data_list=[validate_read_example],
executor=PythonObjectExecutor(python_object=ValidateTransformer()))
validate_artifact_name = artifact_prefix + 'validate_artifact'
validate_artifact = af.register_artifact(name=validate_artifact_name,
batch_uri=get_file_dir(__file__) + '/validate_result')
validate_channel = af.model_validate(input_data_list=[validate_transform],
model_info=train_model,
executor=PythonObjectExecutor(
python_object=ModelValidator(validate_artifact_name)))
with af.config('push_job'):
# Push model to serving
# Register metadata of pushed model
push_model_artifact_name = artifact_prefix + 'push_model_artifact'
push_model_artifact = af.register_artifact(name=push_model_artifact_name,
batch_uri=get_file_dir(__file__) + '/pushed_model')
push_channel = af.push_model(model_info=train_model,
executor=PythonObjectExecutor(
python_object=ModelPusher(push_model_artifact_name)))
with af.config('predict_job'):
# Prediction(Inference)
predict_example = af.register_example(name=artifact_prefix + 'predict_example',
support_type=ExampleSupportType.EXAMPLE_STREAM,
stream_uri=EXAMPLE_URI.format('predict'))
predict_read_example = af.read_example(example_info=predict_example,
executor=PythonObjectExecutor(python_object=PredictExampleReader()))
predict_transform = af.transform(input_data_list=[predict_read_example],
executor=PythonObjectExecutor(python_object=PredictTransformer()))
predict_channel = af.predict(input_data_list=[predict_transform],
model_info=train_model,
executor=PythonObjectExecutor(python_object=ModelPredictor()))
# Save prediction result
write_example = af.register_example(name=artifact_prefix + 'write_example',
support_type=ExampleSupportType.EXAMPLE_STREAM,
stream_uri=get_file_dir(__file__) + '/predict_result')
af.write_example(input_data=predict_channel,
example_info=write_example,
executor=PythonObjectExecutor(python_object=ExampleWriter()))
# Define relation graph connected by control edge:
# Once a round of training is done, validator will be launched and
# pusher will be launched if the new model is better.
# Prediction will start once the first round of training is done and
# when pusher pushes(deploys) a new model, the predictor will use the latest deployed model as well.
af.model_version_control_dependency(src=validate_channel,
model_version_event_type=ModelVersionEventType.MODEL_GENERATED,
dependency=validate_trigger, model_name=train_model.name)
af.model_version_control_dependency(src=push_channel,
model_version_event_type=ModelVersionEventType.MODEL_VALIDATED,
dependency=push_trigger, model_name=train_model.name)
# Run workflow
transform_dag = project_name
af.deploy_to_airflow(project_root_path, dag_id=transform_dag)
af.run(project_path=project_root_path,
dag_id=transform_dag,
scheduler_type=SchedulerType.AIRFLOW)
def run_flink_predict_job():
input_file = "/test1.csv"
output_file = "/output_test2.csv"
example_1 = af.create_example(
name="example_1",
support_type=af.ExampleSupportType.EXAMPLE_BOTH,
batch_uri=input_file,
stream_uri=input_file,
data_format="csv")
example_2 = af.create_example(
name="example_2",
support_type=af.ExampleSupportType.EXAMPLE_BOTH,
batch_uri=output_file,
stream_uri=output_file,
data_format="csv")
flink_config = faf.LocalFlinkJobConfig()
flink_config.flink_home = ''
with af.config(flink_config):
batch_args_1: Properties = {}
ddl = """CREATE TABLE input_table (a INT, b INT, c INT) WITH ('connector' = 'filesystem',
'path' = 'INPUT',
'format' = 'csv'
)"""
table_name = "input_table"
batch_args_1['ddl'] = ddl
batch_args_1['table_name'] = table_name
stream_args_1 = batch_args_1
batch_args_2: Properties = {}
ddl = """CREATE TABLE output_table (aa INT, cc INT) WITH ('connector' = 'filesystem',
'path' = 'OUTPUT',
'format' = 'csv'
)"""
table_name = "output_table"
batch_args_2['ddl'] = ddl
batch_args_2['table_name'] = table_name
stream_args_2 = batch_args_2
input_example = af.read_example(example_info=example_1,
exec_args=ExecuteArgs(
batch_properties=batch_args_1,
stream_properties=stream_args_1))
model_meta = af.ModelMeta(name="test", model_type="saved_model")
model_version = af.ModelVersionMeta(version="11111",
model_path="./tmp/saved_model/",
model_metric="./tmp/saved_model/",
model_id=0)
processed = af.predict(
input_data_list=[input_example],
model_info=model_meta,
model_version_info=model_version,
executor=faf.flink_executor.FlinkJavaExecutor(
java_class="com.apache.flink.ai.flow.TestPredict"))
af.write_example(input_data=processed,
example_info=example_2,
exec_args=ExecuteArgs(
batch_properties=batch_args_2,
stream_properties=stream_args_2))
g = af.default_graph()
workflow = af.compile_workflow(project_path=test_util.get_project_path())
print(dumps(list(workflow.jobs.values())[0]))
def run_workflow():
"""
Run the user-defined workflow definition.
"""
train_data_file, predict_result_directory, merge_predict_result_path, \
first_test_data_file, first_result_data_file = collect_data_file()
# Prepare workflow: Example & Model Metadata registration.
train_example_meta, predict_result_meta, merge_data_meta, first_test_example_meta, second_test_example_meta, \
first_result_example_meta, second_result_example_meta, train_model_meta = \
prepare_workflow(train_data_file=train_data_file,
predict_result_directory=predict_result_directory,
merge_predict_result_path=merge_predict_result_path,
first_test_data_file=first_test_data_file,
first_result_data_file=first_result_data_file)
# Save proxima indexes under the following index path.
index_path = '{}/codes/{}/'.format(os.environ['ENV_HOME'], os.environ['TASK_ID']) + 'test.index'
# Set Python job config to train model.
python_job_config_0 = BaseJobConfig(platform='local', engine='python', job_name='train')
python_job_config_1 = BaseJobConfig(platform='local', engine='python', job_name='start_cluster_serving')
python_job_config_2 = BaseJobConfig(platform='local', engine='python', job_name='merge_predict_result')
# Set Flink job config to predict with cluster serving
global_job_config_1 = LocalFlinkJobConfig()
global_job_config_1.local_mode = 'cluster'
global_job_config_1.flink_home = os.environ['FLINK_HOME']
global_job_config_1.job_name = 'cluster_serving'
global_job_config_1.set_table_env_create_func(StreamTableEnvCreatorBuildIndex())
# Set Flink job config to build index.
global_job_config_2 = LocalFlinkJobConfig()
global_job_config_2.local_mode = 'cluster'
global_job_config_2.flink_home = os.environ['FLINK_HOME']
global_job_config_2.job_name = 'build_index'
global_job_config_2.set_table_env_create_func(StreamTableEnvCreator())
# Set Flink job config to fink sick.
global_job_config_3 = LocalFlinkJobConfig()
global_job_config_3.local_mode = 'cluster'
global_job_config_3.flink_home = os.environ['FLINK_HOME']
global_job_config_3.job_name = 'find_sick'
global_job_config_3.set_table_env_create_func(StreamTableEnvCreator())
# Set Flink job config to online cluster.
global_job_config_4 = LocalFlinkJobConfig()
global_job_config_4.local_mode = 'cluster'
global_job_config_4.flink_home = os.environ['FLINK_HOME']
global_job_config_4.job_name = 'online_cluster'
global_job_config_4.set_table_env_create_func(StreamTableEnvCreator())
with af.config(python_job_config_0):
# Under first job config, we construct the first job, the job is going to train an auto_encoder model.
python_job_0_read_train_example = af.read_example(example_info=train_example_meta,
executor=PythonObjectExecutor(python_object=ReadCsvExample()))
python_job_0_train_model = af.train(input_data_list=[python_job_0_read_train_example],
executor=PythonObjectExecutor(python_object=TrainAutoEncoder()),
model_info=train_model_meta,
name='trainer_0')
with af.config(python_job_config_1):
python_job_1_cluster_serving_channel = af.cluster_serving(model_info=train_model_meta, parallelism=2)
# python_job_1_cluster_serving_channel = af.cluster_serving(model_info=train_model_meta, parallelism=16)
with af.config(global_job_config_1):
flink_job_0_read_train_example = af.read_example(example_info=train_example_meta,
executor=FlinkPythonExecutor(python_object=ReadTrainExample()))
flink_job_0_predict_model = af.predict(input_data_list=[flink_job_0_read_train_example],
model_info=train_model_meta,
executor=FlinkPythonExecutor(
python_object=PredictAutoEncoderWithTrain()))
flink_job_0_write_predict_data = af.write_example(input_data=flink_job_0_predict_model,
example_info=predict_result_meta,
executor=FlinkPythonExecutor(
python_object=WritePredictResult()))
with af.config(python_job_config_2):
python_job_2_merge_train_data_file = af.user_define_operation(executor=PythonObjectExecutor(
python_object=MergePredictResult()))
with af.config(global_job_config_2):
flink_job_1_read_train_example = af.read_example(example_info=merge_data_meta,
executor=FlinkPythonExecutor(python_object=ReadMergeExample()))
flink_job_1_build_index_channel = af.transform([flink_job_1_read_train_example],
executor=FlinkPythonExecutor(
python_object=BuildIndexExecutor(index_path, FloatDataType(),
128)))
with af.config(global_job_config_3):
flink_job_2_read_history_example = af.read_example(example_info=first_test_example_meta,
executor=FlinkPythonExecutor(
python_object=ReadPredictExample()))
flink_job_2_predict_model = af.predict(input_data_list=[flink_job_2_read_history_example],
model_info=train_model_meta,
executor=FlinkPythonExecutor(python_object=PredictAutoEncoder()))
flink_job_2_transformed_data = af.transform([flink_job_2_predict_model],
executor=FlinkPythonExecutor(
python_object=SearchExecutor(index_path, FloatDataType(), 2)))
flink_job_2_read_train_example = af.read_example(example_info=train_example_meta,
executor=FlinkPythonExecutor(python_object=ReadTrainExample()))
flink_job_2_join_channel = af.transform(
input_data_list=[flink_job_2_transformed_data, flink_job_2_read_train_example],
executor=FlinkPythonExecutor(python_object=FindHistory()))
flink_job_2_write_result = af.write_example(input_data=flink_job_2_join_channel,
example_info=first_result_example_meta,
executor=FlinkPythonExecutor(python_object=SearchSink()))
with af.config(global_job_config_4):
flink_job_3_read_online_example = af.read_example(example_info=second_test_example_meta,
executor=FlinkPythonExecutor(
python_object=ReadOnlinePredictExample()))
flink_job_3_predict_model = af.predict(input_data_list=[flink_job_3_read_online_example],
model_info=train_model_meta,
executor=FlinkPythonExecutor(python_object=OnlinePredictAutoEncoder()))
flink_job_3_transformed_data = af.transform([flink_job_3_predict_model],
executor=FlinkPythonExecutor(
python_object=SearchExecutor3(index_path, FloatDataType(), 2)))
af.write_example(input_data=flink_job_3_transformed_data,
example_info=second_result_example_meta,
executor=FlinkPythonExecutor(python_object=WriteSecondResult()))
af.stop_before_control_dependency(python_job_1_cluster_serving_channel, python_job_0_train_model)
af.stop_before_control_dependency(flink_job_0_read_train_example, python_job_1_cluster_serving_channel)
af.stop_before_control_dependency(python_job_2_merge_train_data_file, flink_job_0_read_train_example)
af.stop_before_control_dependency(flink_job_1_build_index_channel, python_job_2_merge_train_data_file)
af.stop_before_control_dependency(flink_job_2_read_history_example, flink_job_1_build_index_channel)
af.stop_before_control_dependency(flink_job_3_read_online_example, flink_job_2_write_result)
workflow_id = af.run(get_project_path()+'/')
res = af.wait_workflow_execution_finished(workflow_id)
sys.exit(res)
