def run_pipeline(client: kfp.Client, pipeline_name: str, pipeline_id: str,
pipeline_paramters_path: dict):
experiment_id = find_experiment_id(
experiment_name=os.environ["INPUT_EXPERIMENT_NAME"], client=client)
if not experiment_id:
raise ValueError("Failed to find experiment with the name: {}".format(
os.environ["INPUT_EXPERIMENT_NAME"]))
logging.info(f"The expriment id is: {experiment_id}")
namespace = None
if (os.getenv("INPUT_PIPELINE_NAMESPACE") !=
None) and (str.isspace(os.getenv("INPUT_PIPELINE_NAMESPACE"))
== False) and os.getenv("INPUT_PIPELINE_NAMESPACE"):
namespace = os.environ["INPUT_PIPELINE_NAMESPACE"]
logging.info(f"The namespace that will be used is: {namespace}")
# [TODO] What would be a good way to name the jobs
job_name = pipeline_name + datetime.now().strftime("%Y_%m_%d_%H_%M_%S")
logging.info(f"The job name is: {job_name}")
pipeline_params = read_pipeline_params(
pipeline_paramters_path=pipeline_paramters_path)
pipeline_params = pipeline_params if pipeline_params != None else {}
logging.info(
f"experiment_id: {experiment_id}, job_name:{job_name}, pipeline_params:{pipeline_params}, pipeline_id:{pipeline_id}, namespace:{namespace}"
)
client.run_pipeline(
experiment_id=experiment_id,
job_name=job_name,
# Read this as a yaml, people seam to prefer that to json.
params=pipeline_params,
pipeline_id=pipeline_id,
namespace=namespace)
logging.info(
"Successfully started the pipeline, head over to kubeflow to check it out"
)
def run_pipeline_func(client: kfp.Client,
pipeline_name: str,
pipeline_id: str,
pipeline_paramters_path: dict,
recurring_flag: bool = False,
cron_exp: str = ''):
pipeline_params = read_pipeline_params(
pipeline_paramters_path=pipeline_paramters_path)
pipeline_params = pipeline_params if pipeline_params is not None else {}
experiment_id = None
experiment_name = "{}-{}".format(pipeline_name,
os.environ["INPUT_EXPERIMENT_NAME"])
try:
experiment_id = client.get_experiment(
experiment_name=experiment_name).to_dict()["id"]
except ValueError:
experiment_id = client.create_experiment(
name=experiment_name).to_dict()["id"]
namespace = os.getenv("INPUT_PIPELINE_NAMESPACE") if not str.isspace(
os.getenv("INPUT_PIPELINE_NAMESPACE")) else None
job_name = 'Run {} on {}'.format(
pipeline_name,
datetime.now().strftime("%Y-%m-%d %H:%M:%S"))
logging.info(f"experiment_id: {experiment_id}, \
job_name: {job_name}, \
pipeline_params: {pipeline_params}, \
pipeline_id: {pipeline_id}, \
namespace: {namespace}, \
cron_exp: {cron_exp}")
if recurring_flag == "true":
client.create_recurring_run(experiment_id=experiment_id,
job_name=job_name,
params=pipeline_params,
pipeline_id=pipeline_id,
cron_expression=cron_exp)
logging.info(
"Successfully started the recurring pipeline, head over to kubeflow to check it out"
)
client.run_pipeline(experiment_id=experiment_id,
job_name=job_name,
params=pipeline_params,
pipeline_id=pipeline_id)
logging.info(
"Successfully started the pipeline, head over to kubeflow to check it out"
)
def run_pipeline_in_experiment(api_pipeline: ApiPipeline, parameters: dict = None, run_name: str = None,
wait_for_status: bool = False):
try:
client = KfpClient(_pipeline_service_url)
experiment = client.create_experiment('PIPELINE_RUNS')
run_result = client.run_pipeline(experiment_id=experiment.id,
job_name=run_name or api_pipeline.name,
params=parameters,
pipeline_id=api_pipeline.id)
run_id = run_result.id
if wait_for_status:
run_details = wait_for_run_status(client, run_id, 10)
run_status = json.loads(run_details.pipeline_runtime.workflow_manifest)["status"]
if run_status \
and run_status.get("phase", "").lower() in ["failed", "error"] \
and run_status.get("message"):
raise RuntimeError(f"Run {run_id} failed with error: {run_status['message']}")
return run_id
except Exception as e:
print(f"Exception trying to run pipeline {api_pipeline.id} '{api_pipeline.name}'"
f" with parameters {parameters}:"
f" %s\n" % e)
raise ApiError(message=f"{e.body}\nKFP URL: {_pipeline_service_url}", http_status_code=e.status or 422)
return None
def main():
args = parse_arguments()
test_cases = []
test_name = args.testname + ' Sample Test'
###### Initialization ######
client = Client(namespace=args.namespace)
###### Check Input File ######
utils.add_junit_test(test_cases, 'input generated yaml file',
os.path.exists(args.input),
'yaml file is not generated')
if not os.path.exists(args.input):
utils.write_junit_xml(test_name, args.result, test_cases)
print('Error: job not found.')
exit(1)
###### Create Experiment ######
experiment_name = args.testname + ' sample experiment'
response = client.create_experiment(experiment_name)
experiment_id = response.id
utils.add_junit_test(test_cases, 'create experiment', True)
###### Create Job ######
job_name = args.testname + '_sample'
params = {}
response = client.run_pipeline(experiment_id, job_name, args.input, params)
run_id = response.id
utils.add_junit_test(test_cases, 'create pipeline run', True)
###### Monitor Job ######
start_time = datetime.now()
response = client.wait_for_run_completion(run_id, 1200)
succ = (response.run.status.lower() == 'succeeded')
end_time = datetime.now()
elapsed_time = (end_time - start_time).seconds
utils.add_junit_test(test_cases, 'job completion', succ,
'waiting for job completion failure', elapsed_time)
###### Output Argo Log for Debugging ######
workflow_json = client._get_workflow_json(run_id)
workflow_id = workflow_json['metadata']['name']
argo_log, _ = utils.run_bash_command('argo logs -n {} -w {}'.format(
args.namespace, workflow_id))
print("=========Argo Workflow Log=========")
print(argo_log)
if not succ:
utils.write_junit_xml(test_name, args.result, test_cases)
exit(1)
###### Delete Job ######
#TODO: add deletion when the backend API offers the interface.
###### Write out the test result in junit xml ######
utils.write_junit_xml(test_name, args.result, test_cases)
def _mock_pipelines_creation(kfp_client_mock: kfp.Client):
def _mock_create_experiment(name, description=None, namespace=None):
return kfp_server_api.models.ApiExperiment(
id="some-exp-id",
name=name,
description=description,
)
def _mock_run_pipeline(
experiment_id,
job_name,
pipeline_package_path=None,
params={},
pipeline_id=None,
version_id=None,
):
return kfp_server_api.models.ApiRun(id="some-run-id", name=job_name)
kfp_client_mock.create_experiment = _mock_create_experiment
kfp_client_mock.run_pipeline = _mock_run_pipeline
def run_pipeline(
pipeline,
arguments=None,
project=None,
experiment=None,
run=None,
namespace=None,
artifact_path=None,
ops=None,
url=None,
ttl=None,
):
"""remote KubeFlow pipeline execution
Submit a workflow task to KFP via mlrun API service
:param pipeline: KFP pipeline function or path to .yaml/.zip pipeline file
:param arguments: pipeline arguments
:param experiment: experiment name
:param run: optional, run name
:param namespace: Kubernetes namespace (if not using default)
:param url: optional, url to mlrun API service
:param artifact_path: target location/url for mlrun artifacts
:param ops: additional operators (.apply() to all pipeline functions)
:param ttl: pipeline ttl in secs (after that the pods will be removed)
:returns: kubeflow pipeline id
"""
remote = not get_k8s_helper(
silent=True).is_running_inside_kubernetes_cluster()
artifact_path = artifact_path or mlconf.artifact_path
if artifact_path and "{{run.uid}}" in artifact_path:
artifact_path.replace("{{run.uid}}", "{{workflow.uid}}")
if artifact_path and "{{run.project}}" in artifact_path:
if not project:
raise ValueError("project name must be specified with this" +
f" artifact_path template {artifact_path}")
artifact_path.replace("{{run.project}}", project)
if not artifact_path:
raise ValueError("artifact path was not specified")
namespace = namespace or mlconf.namespace
arguments = arguments or {}
if remote or url:
mldb = get_run_db(url)
if mldb.kind != "http":
raise ValueError(
"run pipeline require access to remote api-service"
", please set the dbpath url")
id = mldb.submit_pipeline(
pipeline,
arguments,
experiment=experiment,
run=run,
namespace=namespace,
ops=ops,
artifact_path=artifact_path,
)
else:
client = Client(namespace=namespace)
if isinstance(pipeline, str):
experiment = client.create_experiment(name=experiment)
run_result = client.run_pipeline(experiment.id,
run,
pipeline,
params=arguments)
else:
conf = new_pipe_meta(artifact_path, ttl, ops)
run_result = client.create_run_from_pipeline_func(
pipeline,
arguments,
run_name=run,
experiment_name=experiment,
pipeline_conf=conf,
)
id = run_result.run_id
logger.info("Pipeline run id={}, check UI or DB for progress".format(id))
return id
class PySampleChecker(object):
def __init__(self, testname, input, output, result, namespace='kubeflow'):
"""Util class for checking python sample test running results.
:param testname: test name.
:param input: The path of a pipeline file that will be submitted.
:param output: The path of the test output.
:param result: The path of the test result that will be exported.
:param namespace: namespace of the deployed pipeline system. Default: kubeflow
"""
self._testname = testname
self._input = input
self._output = output
self._result = result
self._namespace = namespace
self._run_pipeline = None
self._test_timeout = None
self._test_cases = []
self._test_name = self._testname + ' Sample Test'
self._client = None
self._experiment_id = None
self._job_name = None
self._test_args = None
self._run_id = None
def run(self):
"""Run compiled KFP pipeline."""
###### Initialization ######
host = 'ml-pipeline.%s.svc.cluster.local:8888' % self._namespace
self._client = Client(host=host)
###### Check Input File ######
utils.add_junit_test(self._test_cases, 'input generated yaml file',
os.path.exists(self._input),
'yaml file is not generated')
if not os.path.exists(self._input):
utils.write_junit_xml(self._test_name, self._result,
self._test_cases)
print('Error: job not found.')
exit(1)
###### Create Experiment ######
experiment_name = self._testname + ' sample experiment'
response = self._client.create_experiment(experiment_name)
self._experiment_id = response.id
utils.add_junit_test(self._test_cases, 'create experiment', True)
###### Create Job ######
self._job_name = self._testname + '_sample'
###### Figure out arguments from associated config files. #######
self._test_args = {}
config_schema = yamale.make_schema(SCHEMA_CONFIG)
try:
with open(DEFAULT_CONFIG, 'r') as f:
raw_args = yaml.safe_load(f)
default_config = yamale.make_data(DEFAULT_CONFIG)
yamale.validate(
config_schema,
default_config) # If fails, a ValueError will be raised.
except yaml.YAMLError as yamlerr:
raise RuntimeError('Illegal default config:{}'.format(yamlerr))
except OSError as ose:
raise FileExistsError('Default config not found:{}'.format(ose))
else:
self._test_timeout = raw_args['test_timeout']
self._run_pipeline = raw_args['run_pipeline']
try:
with open(
os.path.join(CONFIG_DIR,
'%s.config.yaml' % self._testname), 'r') as f:
raw_args = yaml.safe_load(f)
test_config = yamale.make_data(
os.path.join(CONFIG_DIR, '%s.config.yaml' % self._testname))
yamale.validate(
config_schema,
test_config) # If fails, a ValueError will be raised.
except yaml.YAMLError as yamlerr:
print(
'No legit yaml config file found, use default args:{}'.format(
yamlerr))
except OSError as ose:
print(
'Config file with the same name not found, use default args:{}'
.format(ose))
else:
self._test_args.update(raw_args['arguments'])
if 'output' in self._test_args.keys(
): # output is a special param that has to be specified dynamically.
self._test_args['output'] = self._output
if 'test_timeout' in raw_args.keys():
self._test_timeout = raw_args['test_timeout']
if 'run_pipeline' in raw_args.keys():
self._run_pipeline = raw_args['run_pipeline']
# Submit for pipeline running.
if self._run_pipeline:
response = self._client.run_pipeline(self._experiment_id,
self._job_name, self._input,
self._test_args)
self._run_id = response.id
utils.add_junit_test(self._test_cases, 'create pipeline run', True)
def check(self):
"""Check pipeline run results."""
if self._run_pipeline:
###### Monitor Job ######
try:
start_time = datetime.now()
response = self._client.wait_for_run_completion(
self._run_id, self._test_timeout)
succ = (response.run.status.lower() == 'succeeded')
end_time = datetime.now()
elapsed_time = (end_time - start_time).seconds
utils.add_junit_test(self._test_cases, 'job completion', succ,
'waiting for job completion failure',
elapsed_time)
finally:
###### Output Argo Log for Debugging ######
workflow_json = self._client._get_workflow_json(self._run_id)
workflow_id = workflow_json['metadata']['name']
argo_log, _ = utils.run_bash_command(
'argo logs -n {} -w {}'.format(self._namespace,
workflow_id))
print('=========Argo Workflow Log=========')
print(argo_log)
if not succ:
utils.write_junit_xml(self._test_name, self._result,
self._test_cases)
exit(1)
###### Validate the results for specific test cases ######
#TODO: Add result check for tfx-cab-classification after launch.
if self._testname == 'xgboost_training_cm':
# For xgboost sample, check its confusion matrix.
cm_tar_path = './confusion_matrix.tar.gz'
utils.get_artifact_in_minio(workflow_json, 'confusion-matrix',
cm_tar_path,
'mlpipeline-ui-metadata')
with tarfile.open(cm_tar_path) as tar_handle:
file_handles = tar_handle.getmembers()
assert len(file_handles) == 1
with tar_handle.extractfile(file_handles[0]) as f:
cm_data = f.read()
utils.add_junit_test(
self._test_cases, 'confusion matrix format',
(len(cm_data) > 0),
'the confusion matrix file is empty')
###### Delete Job ######
#TODO: add deletion when the backend API offers the interface.
###### Write out the test result in junit xml ######
utils.write_junit_xml(self._test_name, self._result, self._test_cases)
def main():
args = parse_arguments()
test_cases = []
test_name = 'Kubeflow Sample Test'
###### Initialization ######
client = Client()
###### Check Input File ######
utils.add_junit_test(test_cases, 'input generated yaml file',
os.path.exists(args.input),
'yaml file is not generated')
if not os.path.exists(args.input):
utils.write_junit_xml(test_name, args.result, test_cases)
exit()
###### Create Experiment ######
experiment_name = 'kubeflow sample experiment'
response = client.create_experiment(experiment_name)
experiment_id = response.id
utils.add_junit_test(test_cases, 'create experiment', True)
###### Create Job ######
job_name = 'kubeflow_sample'
params = {
'output': args.output,
'project': 'ml-pipeline-test',
'evaluation': 'gs://ml-pipeline-dataset/sample-test/flower/eval15.csv',
'train': 'gs://ml-pipeline-dataset/sample-test/flower/train30.csv',
'hidden-layer-size': '10,5',
'steps': '5'
}
response = client.run_pipeline(experiment_id, job_name, args.input, params)
run_id = response.id
utils.add_junit_test(test_cases, 'create pipeline run', True)
###### Monitor Job ######
start_time = datetime.now()
response = client.wait_for_run_completion(run_id, 1200)
succ = (response.run.status.lower() == 'succeeded')
end_time = datetime.now()
elapsed_time = (end_time - start_time).seconds
utils.add_junit_test(test_cases, 'job completion', succ,
'waiting for job completion failure', elapsed_time)
if not succ:
utils.write_junit_xml(test_name, args.result, test_cases)
exit()
###### Output Argo Log for Debugging ######
workflow_json = client._get_workflow_json(run_id)
workflow_id = workflow_json['metadata']['name']
#TODO: remove the namespace dependency or make is configurable.
argo_log, _ = utils.run_bash_command(
'argo logs -n kubeflow -w {}'.format(workflow_id))
print("=========Argo Workflow Log=========")
print(argo_log)
###### Validate the results ######
# confusion matrix should show three columns for the flower data
# target, predicted, count
cm_tar_path = './confusion_matrix.tar.gz'
cm_filename = 'mlpipeline-ui-metadata.json'
utils.get_artifact_in_minio(workflow_json, 'confusionmatrix', cm_tar_path)
tar_handler = tarfile.open(cm_tar_path)
tar_handler.extractall()
with open(cm_filename, 'r') as f:
cm_data = json.load(f)
utils.add_junit_test(
test_cases, 'confusion matrix format',
(len(cm_data['outputs'][0]['schema']) == 3),
'the column number of the confusion matrix output is not equal to three'
)
###### Delete Job ######
#TODO: add deletion when the backend API offers the interface.
###### Write out the test result in junit xml ######
utils.write_junit_xml(test_name, args.result, test_cases)
def run_pipeline(pipeline,
arguments=None,
experiment=None,
run=None,
namespace=None,
artifact_path=None,
ops=None,
url=None,
remote=False):
"""remote KubeFlow pipeline execution
Submit a workflow task to KFP via mlrun API service
:param pipeline KFP pipeline function or path to .yaml/.zip pipeline file
:param arguments pipeline arguments
:param experiment experiment name
:param run optional, run name
:param namespace Kubernetes namespace (if not using default)
:param url optional, url to mlrun API service
:param artifact_path target location/url for mlrun artifacts
:param ops additional operators (.apply() to all pipeline functions)
:param remote use mlrun remote API service vs direct KFP APIs
:return kubeflow pipeline id
"""
namespace = namespace or mlconf.namespace
arguments = arguments or {}
if remote or url:
mldb = get_run_db(url).connect()
if mldb.kind != 'http':
raise ValueError(
'run pipeline require access to remote api-service'
', please set the dbpath url')
id = mldb.submit_pipeline(pipeline,
arguments,
experiment=experiment,
run=run,
namespace=namespace,
ops=ops,
artifact_path=artifact_path)
else:
client = Client(namespace=namespace)
if isinstance(pipeline, str):
experiment = client.create_experiment(name=experiment)
run_result = client.run_pipeline(experiment.id,
run,
pipeline,
params=arguments)
else:
conf = new_pipe_meta(artifact_path, ops)
run_result = client.create_run_from_pipeline_func(
pipeline,
arguments,
run_name=run,
experiment_name=experiment,
pipeline_conf=conf)
id = run_result.run_id
logger.info('Pipeline run id={}, check UI or DB for progress'.format(id))
return id
def main():
args = parse_arguments()
test_cases = []
test_name = args.testname + ' Sample Test'
###### Initialization ######
host = 'ml-pipeline.%s.svc.cluster.local:8888' % args.namespace
client = Client(host=host)
###### Check Input File ######
utils.add_junit_test(test_cases, 'input generated yaml file',
os.path.exists(args.input),
'yaml file is not generated')
if not os.path.exists(args.input):
utils.write_junit_xml(test_name, args.result, test_cases)
print('Error: job not found.')
exit(1)
###### Create Experiment ######
experiment_name = args.testname + ' sample experiment'
response = client.create_experiment(experiment_name)
experiment_id = response.id
utils.add_junit_test(test_cases, 'create experiment', True)
###### Create Job ######
job_name = args.testname + '_sample'
###### Test-specific arguments #######
if args.testname == 'tfx_cab_classification':
params = {
'output': args.output,
'project': 'ml-pipeline-test',
'column-names':
'gs://ml-pipeline-dataset/sample-test/taxi-cab-classification/column-names.json',
'evaluation':
'gs://ml-pipeline-dataset/sample-test/taxi-cab-classification/eval20.csv',
'train':
'gs://ml-pipeline-dataset/sample-test/taxi-cab-classification/train50.csv',
'hidden-layer-size': '5',
'steps': '5'
}
elif args.testname == 'xgboost_training_cm':
params = {
'output': args.output,
'project': 'ml-pipeline-test',
'train-data':
'gs://ml-pipeline-dataset/sample-test/sfpd/train_50.csv',
'eval-data':
'gs://ml-pipeline-dataset/sample-test/sfpd/eval_20.csv',
'schema': 'gs://ml-pipeline-dataset/sample-test/sfpd/schema.json',
'rounds': '20',
'workers': '2'
}
else:
# Basic tests require no additional params.
params = {}
response = client.run_pipeline(experiment_id, job_name, args.input, params)
run_id = response.id
utils.add_junit_test(test_cases, 'create pipeline run', True)
###### Monitor Job ######
try:
start_time = datetime.now()
if args.testname == 'xgboost_training_cm':
response = client.wait_for_run_completion(run_id, 1800)
else:
response = client.wait_for_run_completion(run_id, 1200)
succ = (response.run.status.lower() == 'succeeded')
end_time = datetime.now()
elapsed_time = (end_time - start_time).seconds
utils.add_junit_test(test_cases, 'job completion', succ,
'waiting for job completion failure',
elapsed_time)
finally:
###### Output Argo Log for Debugging ######
workflow_json = client._get_workflow_json(run_id)
workflow_id = workflow_json['metadata']['name']
argo_log, _ = utils.run_bash_command('argo logs -n {} -w {}'.format(
args.namespace, workflow_id))
print('=========Argo Workflow Log=========')
print(argo_log)
if not succ:
utils.write_junit_xml(test_name, args.result, test_cases)
exit(1)
###### Validate the results for specific test cases ######
#TODO: Add result check for tfx-cab-classification after launch.
if args.testname == 'xgboost_training_cm':
cm_tar_path = './confusion_matrix.tar.gz'
utils.get_artifact_in_minio(workflow_json, 'confusion-matrix',
cm_tar_path, 'mlpipeline-ui-metadata')
with tarfile.open(cm_tar_path) as tar_handle:
file_handles = tar_handle.getmembers()
assert len(file_handles) == 1
with tar_handle.extractfile(file_handles[0]) as f:
cm_data = f.read()
utils.add_junit_test(test_cases, 'confusion matrix format',
(len(cm_data) > 0),
'the confusion matrix file is empty')
###### Delete Job ######
#TODO: add deletion when the backend API offers the interface.
###### Write out the test result in junit xml ######
utils.write_junit_xml(test_name, args.result, test_cases)
def main():
args = parse_arguments()
test_cases = []
test_name = 'XGBoost Sample Test'
###### Initialization ######
host = 'ml-pipeline.%s.svc.cluster.local:8888' % args.namespace
client = Client(host=host)
###### Check Input File ######
utils.add_junit_test(test_cases, 'input generated yaml file',
os.path.exists(args.input),
'yaml file is not generated')
if not os.path.exists(args.input):
utils.write_junit_xml(test_name, args.result, test_cases)
print('Error: job not found.')
exit(1)
###### Create Experiment ######
experiment_name = 'xgboost sample experiment'
response = client.create_experiment(experiment_name)
experiment_id = response.id
utils.add_junit_test(test_cases, 'create experiment', True)
###### Create Job ######
job_name = 'xgboost_sample'
params = {
'output': args.output,
'project': 'ml-pipeline-test',
'train-data': 'gs://ml-pipeline-dataset/sample-test/sfpd/train_50.csv',
'eval-data': 'gs://ml-pipeline-dataset/sample-test/sfpd/eval_20.csv',
'schema': 'gs://ml-pipeline-dataset/sample-test/sfpd/schema.json',
'rounds': '20',
'workers': '2'
}
response = client.run_pipeline(experiment_id, job_name, args.input, params)
run_id = response.id
utils.add_junit_test(test_cases, 'create pipeline run', True)
###### Monitor Job ######
start_time = datetime.now()
response = client.wait_for_run_completion(run_id, 1800)
succ = (response.run.status.lower() == 'succeeded')
end_time = datetime.now()
elapsed_time = (end_time - start_time).seconds
utils.add_junit_test(test_cases, 'job completion', succ,
'waiting for job completion failure', elapsed_time)
###### Output Argo Log for Debugging ######
workflow_json = client._get_workflow_json(run_id)
workflow_id = workflow_json['metadata']['name']
argo_log, _ = utils.run_bash_command('argo logs -n {} -w {}'.format(
args.namespace, workflow_id))
print("=========Argo Workflow Log=========")
print(argo_log)
###### If the job fails, skip the result validation ######
if not succ:
utils.write_junit_xml(test_name, args.result, test_cases)
exit(1)
###### Validate the results ######
# confusion matrix should show three columns for the flower data
# target, predicted, count
cm_tar_path = './confusion_matrix.tar.gz'
cm_filename = 'mlpipeline-ui-metadata.json'
utils.get_artifact_in_minio(workflow_json, 'confusion-matrix', cm_tar_path)
tar_handler = tarfile.open(cm_tar_path)
tar_handler.extractall()
with open(cm_filename, 'r') as f:
cm_data = f.read()
utils.add_junit_test(test_cases, 'confusion matrix format',
(len(cm_data) > 0),
'the confusion matrix file is empty')
###### Delete Job ######
#TODO: add deletion when the backend API offers the interface.
###### Write out the test result in junit xml ######
utils.write_junit_xml(test_name, args.result, test_cases)
def main():
args = parse_arguments()
test_cases = []
test_name = 'Resnet CMLE Test'
###### Initialization ######
host = 'ml-pipeline.%s.svc.cluster.local:8888' % args.namespace
client = Client(host=host)
###### Check Input File ######
utils.add_junit_test(test_cases, 'input generated yaml file',
os.path.exists(args.input),
'yaml file is not generated')
if not os.path.exists(args.input):
utils.write_junit_xml(test_name, args.result, test_cases)
print('Error: job not found.')
exit(1)
###### Create Experiment ######
experiment_name = 'resnet cmle sample experiment'
response = client.create_experiment(experiment_name)
experiment_id = response.id
utils.add_junit_test(test_cases, 'create experiment', True)
###### Create Job ######
job_name = 'cmle_sample'
params = {
'output': args.output,
'project_id': 'ml-pipeline-test',
'region': 'us-central1',
'model': 'bolts',
'version': 'beta1',
'tf_version':
'1.9', # Watch out! If 1.9 is no longer supported we need to set it to a newer version.
'train_csv':
'gs://ml-pipeline-dataset/sample-test/bolts/bolt_images_train_sample1000.csv',
'validation_csv':
'gs://ml-pipeline-dataset/sample-test/bolts/bolt_images_validate_sample200.csv',
'labels': 'gs://bolts_image_dataset/labels.txt',
'depth': 50,
'train_batch_size': 32,
'eval_batch_size': 32,
'steps_per_eval': 128,
'train_steps': 128,
'num_train_images': 1000,
'num_eval_images': 200,
'num_label_classes': 10
}
response = client.run_pipeline(experiment_id, job_name, args.input, params)
run_id = response.id
utils.add_junit_test(test_cases, 'create pipeline run', True)
###### Monitor Job ######
try:
start_time = datetime.now()
response = client.wait_for_run_completion(run_id, 1800)
succ = (response.run.status.lower() == 'succeeded')
end_time = datetime.now()
elapsed_time = (end_time - start_time).seconds
utils.add_junit_test(test_cases, 'job completion', succ,
'waiting for job completion failure',
elapsed_time)
finally:
###### Output Argo Log for Debugging ######
workflow_json = client._get_workflow_json(run_id)
workflow_id = workflow_json['metadata']['name']
argo_log, _ = utils.run_bash_command('argo logs -n {} -w {}'.format(
args.namespace, workflow_id))
print("=========Argo Workflow Log=========")
print(argo_log)
if not succ:
utils.write_junit_xml(test_name, args.result, test_cases)
exit(1)
###### Delete Job ######
#TODO: add deletion when the backend API offers the interface.
###### Write out the test result in junit xml ######
utils.write_junit_xml(test_name, args.result, test_cases)
def main():
args = parse_arguments()
test_cases = []
test_name = 'TFX Sample Test'
###### Initialization ######
client = Client(namespace=args.namespace)
###### Check Input File ######
utils.add_junit_test(test_cases, 'input generated yaml file', os.path.exists(args.input), 'yaml file is not generated')
if not os.path.exists(args.input):
utils.write_junit_xml(test_name, args.result, test_cases)
exit()
###### Create Experiment ######
experiment_name = 'TFX sample experiment'
response = client.create_experiment(experiment_name)
experiment_id = response.id
utils.add_junit_test(test_cases, 'create experiment', True)
###### Create Job ######
job_name = 'TFX_sample'
params = {'output': args.output,
'project': 'ml-pipeline-test',
'column-names': 'gs://ml-pipeline-dataset/sample-test/taxi-cab-classification/column-names.json',
'evaluation': 'gs://ml-pipeline-dataset/sample-test/taxi-cab-classification/eval20.csv',
'train': 'gs://ml-pipeline-dataset/sample-test/taxi-cab-classification/train50.csv',
'hidden-layer-size': '5',
'steps': '5'}
response = client.run_pipeline(experiment_id, job_name, args.input, params)
run_id = response.id
utils.add_junit_test(test_cases, 'create pipeline run', True)
###### Monitor Job ######
start_time = datetime.now()
response = client.wait_for_run_completion(run_id, 1200)
succ = (response.run.status.lower()=='succeeded')
end_time = datetime.now()
elapsed_time = (end_time - start_time).seconds
utils.add_junit_test(test_cases, 'job completion', succ, 'waiting for job completion failure', elapsed_time)
if not succ:
utils.write_junit_xml(test_name, args.result, test_cases)
exit()
###### Output Argo Log for Debugging ######
workflow_json = client._get_workflow_json(run_id)
workflow_id = workflow_json['metadata']['name']
argo_log, _ = utils.run_bash_command('argo logs -n {} -w {}'.format(args.namespace, workflow_id))
print("=========Argo Workflow Log=========")
print(argo_log)
###### Validate the results ######
#TODO: enable after launch
# model analysis html is validated
# argo_workflow_id = workflow_json['metadata']['name']
# gcs_html_path = os.path.join(os.path.join(args.output, str(argo_workflow_id)), 'analysis/output_display.html')
# print('Output display HTML path is ' + gcs_html_path)
# utils.run_bash_command('gsutil cp ' + gcs_html_path + './')
# display_file = open('./output_display.html', 'r')
# is_next_line_state = False
# for line in display_file:
# if is_next_line_state:
# state = line.strip()
# break
# if line.strip() == '<script type="application/vnd.jupyter.widget-state+json">':
# is_next_line_state = True
# import json
# state_json = json.loads(state)
# succ = ('state' in state_json and 'version_major' in state_json and 'version_minor' in state_json)
# utils.add_junit_test(test_cases, 'output display html', succ, 'the state json does not contain state, version_major, or version_inor')
###### Delete Job ######
#TODO: add deletion when the backend API offers the interface.
###### Write out the test result in junit xml ######
utils.write_junit_xml(test_name, args.result, test_cases)
def check(self):
"""Run sample test and check results."""
test_cases = []
test_name = self._testname + ' Sample Test'
###### Initialization ######
host = 'ml-pipeline.%s.svc.cluster.local:8888' % self._namespace
client = Client(host=host)
###### Check Input File ######
utils.add_junit_test(test_cases, 'input generated yaml file',
os.path.exists(self._input), 'yaml file is not generated')
if not os.path.exists(self._input):
utils.write_junit_xml(test_name, self._result, test_cases)
print('Error: job not found.')
exit(1)
###### Create Experiment ######
experiment_name = self._testname + ' sample experiment'
response = client.create_experiment(experiment_name)
experiment_id = response.id
utils.add_junit_test(test_cases, 'create experiment', True)
###### Create Job ######
job_name = self._testname + '_sample'
###### Figure out arguments from associated config files. #######
test_args = {}
try:
with open(DEFAULT_CONFIG, 'r') as f:
raw_args = yaml.safe_load(f)
except yaml.YAMLError as yamlerr:
raise RuntimeError('Illegal default config:{}'.format(yamlerr))
except OSError as ose:
raise FileExistsError('Default config not found:{}'.format(ose))
else:
test_timeout = raw_args['test_timeout']
try:
with open(os.path.join(CONFIG_DIR, '%s.config.yaml' % self._testname), 'r') as f:
raw_args = yaml.safe_load(f)
except yaml.YAMLError as yamlerr:
print('No legit yaml config file found, use default args:{}'.format(yamlerr))
except OSError as ose:
print('Config file with the same name not found, use default args:{}'.format(ose))
else:
test_args.update(raw_args['arguments'])
if 'output' in test_args.keys(): # output is a special param that has to be specified dynamically.
test_args['output'] = self._output
if 'test_timeout' in raw_args.keys():
test_timeout = raw_args['test_timeout']
response = client.run_pipeline(experiment_id, job_name, self._input, test_args)
run_id = response.id
utils.add_junit_test(test_cases, 'create pipeline run', True)
###### Monitor Job ######
try:
start_time = datetime.now()
response = client.wait_for_run_completion(run_id, test_timeout)
succ = (response.run.status.lower() == 'succeeded')
end_time = datetime.now()
elapsed_time = (end_time - start_time).seconds
utils.add_junit_test(test_cases, 'job completion', succ,
'waiting for job completion failure', elapsed_time)
finally:
###### Output Argo Log for Debugging ######
workflow_json = client._get_workflow_json(run_id)
workflow_id = workflow_json['metadata']['name']
argo_log, _ = utils.run_bash_command('argo logs -n {} -w {}'.format(
self._namespace, workflow_id))
print('=========Argo Workflow Log=========')
print(argo_log)
if not succ:
utils.write_junit_xml(test_name, self._result, test_cases)
exit(1)
###### Validate the results for specific test cases ######
#TODO: Add result check for tfx-cab-classification after launch.
if self._testname == 'xgboost_training_cm':
# For xgboost sample, check its confusion matrix.
cm_tar_path = './confusion_matrix.tar.gz'
utils.get_artifact_in_minio(workflow_json, 'confusion-matrix', cm_tar_path,
'mlpipeline-ui-metadata')
with tarfile.open(cm_tar_path) as tar_handle:
file_handles = tar_handle.getmembers()
assert len(file_handles) == 1
with tar_handle.extractfile(file_handles[0]) as f:
cm_data = f.read()
utils.add_junit_test(test_cases, 'confusion matrix format',
(len(cm_data) > 0),
'the confusion matrix file is empty')
###### Delete Job ######
#TODO: add deletion when the backend API offers the interface.
###### Write out the test result in junit xml ######
utils.write_junit_xml(test_name, self._result, test_cases)
