def setUp(self):
# Create a temporary directory
self.config_path = tempfile.mkdtemp()
self.config_file = 'config.yaml'
self.config_fil_full_path = os.path.join(self.config_path,
self.config_file)
self.config_content = open(
get_resource_file_path('default-config.yaml')).read()
with open(self.config_fil_full_path, 'w') as stream:
stream.write(self.config_content)
# Creating prescience config
self.prescience_config = PrescienceConfig(config_path=self.config_path,
config_file=self.config_file)
# Loading configuration
self.prescience_config.load()
# Setting environnement variables
os.environ['PRESCIENCE_DEFAULT_PROJECT'] = 'project-default'
os.environ['PRESCIENCE_DEFAULT_TOKEN'] = 'token-default'
os.environ[
'PRESCIENCE_DEFAULT_API_URL'] = 'https://default-api.ai.ovh.net'
os.environ[
'PRESCIENCE_DEFAULT_WEBSOCKET_URL'] = 'wss://default-websocket.ai.ovh.net'
os.environ[
'PRESCIENCE_DEFAULT_SERVING_URL'] = 'https://default-serving.ai.ovh.net'
def test_scenario(self):
print('01 - Creating local input...')
input_path = get_resource_file_path('test.csv')
input_file = prescience.csv_local_file_input(filepath=input_path, headers=True)
print('02 - Launching parse task...')
parse_task = input_file.parse(source_id='my-source-id').watch()
print('03 - Showing sources...')
prescience.sources().show()
source = parse_task.source()
print('04 - Showing source schema...')
source.schema().show()
print('05 - Launching preprocess task...')
preprocess_task = source.preprocess(dataset_id='my-dataset_id').watch()
print('06 - Showing datasets...')
prescience.datasets().show()
dataset = preprocess_task.dataset()
print('07 - Showing dataset schema...')
dataset.schema().show()
print('08 - Launching optimize task...')
dataset.optimize().watch()
evaluation_results = dataset.evaluation_results()
print('09 - Showing evaluation results...')
evaluation_results.show()
single_eval_result = evaluation_results.content[0]
print('10 - Launching train task...')
single_eval_result.train('my-model').watch()
prescience.models().show()
def setUp(self):
# Create a temporary directory
self.config_path = tempfile.mkdtemp()
self.config_file = 'config.yaml'
self.config_fil_full_path = os.path.join(self.config_path,
self.config_file)
self.config_content = open(
get_resource_file_path('default-config.yaml')).read()
with open(self.config_fil_full_path, 'w') as stream:
stream.write(self.config_content)
# Creating prescience config
self.prescience_config = PrescienceConfig(config_path=self.config_path,
config_file=self.config_file)
# Loading configuration
self.prescience_config.load()
# Creating prescience client
self.presience_client = PrescienceClient(
prescience_config=self.prescience_config)
def test_parse(self):
"""
Test the launch of a simple Parse
"""
# Init
output = {
'uuid': 'parse-task-uuid',
'type': 'parse',
'status': 'PENDING'
}
self.presience_client.call = MagicMock(return_value=(200, output, {}))
# Test
csv_path = get_resource_file_path('test.csv')
input_file = self.presience_client.csv_local_file_input(
filepath=csv_path, headers=True)
task = input_file.parse('my-source-id')
self.assertEqual('parse-task-uuid', task.uuid())
self.assertEqual('parse', task.type())
self.assertIsInstance(task, ParseTask)
self.assertEqual(Status.PENDING, task.status())
expected_parse_payload = {
'source_id': 'my-source-id',
'type': 'CSV',
'headers': True,
'separator': 'COMMA'
}
self.presience_client.call.assert_called_with(
method='POST',
path='/ml/upload/source',
content_type='multipart/form-data',
multipart=[('input',
(pycurl.FORM_CONTENTS,
json.dumps(expected_parse_payload),
pycurl.FORM_CONTENTTYPE, 'application/json')),
('input-file', (pycurl.FORM_FILE, csv_path))],
data=None,
filepath=None,
query_parameters=None,
call_type=PrescienceWebService.API)
def test_scenario(self):
print('01 - Creating local input...')
input_path = get_resource_file_path('test.csv')
input_file = prescience.csv_local_file_input(filepath=input_path,
headers=True)
print('02 - Launching parse task...')
input_file.parse(source_id='my-source-id').watch()
print('03 - Showing sources...')
prescience.sources().show()
source = prescience.source('my-source-id')
print('04 - Showing source schema...')
source.schema().show()
print('05 - Launching preprocess task...')
preprocess_task = source.preprocess(dataset_id='my-dataset_id').watch()
print('06 - Showing datasets...')
prescience.datasets().show()
dataset = preprocess_task.dataset()
print('07 - Showing dataset schema...')
dataset.schema().show()
print('08 - Launching optimize task...')
dataset.optimize().watch()
evaluation_results = dataset.evaluation_results()
print('09 - Showing evaluation results...')
evaluation_results.show()
single_eval_result = evaluation_results.content[0]
print('10 - Launching train task...')
single_eval_result.train('my-model').watch()
prescience.models().show()
print('10 - Evaluate model...')
evaluation_payload_input = prescience.model(
'my-model').get_model_evaluation_payload(
evaluation_id='my-evaluation',
arguments={
'hours-per-week': 40,
'capital-gain': 2174,
'education-num': 13,
'random-bool': 'True',
'marital-status': 'Never-married',
'age': 39,
'sex': 'Male',
'relationship': 'Not-in-family',
'education': 'Bachelors',
'race': 'White',
'native-country': 'United-States',
'fnlwgt': 77516,
'workclass': 'State-gov',
'capital-loss': 0,
'occupation': 'Adm-clerical'
})
evaluation_payload_input.show()
validation_result, _ = evaluation_payload_input.validate()
self.assertEqual(False, validation_result)
evaluation_payload_input.add_payload_argument('random-bool', True)
evaluation_payload_input.show()
validation_result, _ = evaluation_payload_input.validate()
self.assertEqual(True, validation_result)
evaluation_payload_output = evaluation_payload_input.evaluate()
evaluation_payload_output.show()
final_label = evaluation_payload_output.get_result_label()
self.assertIn(final_label, ['<=50K', '>50K'])