Skip to content

Commit

Permalink
[AIRFLOW-1273]AIRFLOW-1273] Add Google Cloud ML version and model ope…
Browse files Browse the repository at this point in the history
…rators

Includes Google Cloud ML hooks for version and
model operations,
and their unit tests.

https://issues.apache.org/jira/browse/AIRFLOW-1273

Closes apache#2379 from N3da/master
  • Loading branch information
N3da authored and criccomini committed Jun 27, 2017
1 parent e870a8e commit 534a0e0
Show file tree
Hide file tree
Showing 5 changed files with 605 additions and 0 deletions.
1 change: 1 addition & 0 deletions airflow/contrib/hooks/__init__.py
Original file line number Diff line number Diff line change
Expand Up @@ -40,6 +40,7 @@
'qubole_hook': ['QuboleHook'],
'gcs_hook': ['GoogleCloudStorageHook'],
'datastore_hook': ['DatastoreHook'],
'gcp_cloudml_hook': ['CloudMLHook'],
'gcp_dataproc_hook': ['DataProcHook'],
'gcp_dataflow_hook': ['DataFlowHook'],
'spark_submit_operator': ['SparkSubmitOperator'],
Expand Down
167 changes: 167 additions & 0 deletions airflow/contrib/hooks/gcp_cloudml_hook.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,167 @@
#
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements. See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.


import logging
import random
import time
from airflow import settings
from airflow.contrib.hooks.gcp_api_base_hook import GoogleCloudBaseHook
from apiclient.discovery import build
from apiclient import errors
from oauth2client.client import GoogleCredentials

logging.getLogger('GoogleCloudML').setLevel(settings.LOGGING_LEVEL)


def _poll_with_exponential_delay(request, max_n, is_done_func, is_error_func):

for i in range(0, max_n):
try:
response = request.execute()
if is_error_func(response):
raise ValueError('The response contained an error: {}'.format(response))
elif is_done_func(response):
logging.info('Operation is done: {}'.format(response))
return response
else:
time.sleep((2**i) + (random.randint(0, 1000) / 1000))
except errors.HttpError as e:
if e.resp.status != 429:
logging.info('Something went wrong. Not retrying: {}'.format(e))
raise e
else:
time.sleep((2**i) + (random.randint(0, 1000) / 1000))


class CloudMLHook(GoogleCloudBaseHook):

def __init__(self, gcp_conn_id='google_cloud_default', delegate_to=None):
super(CloudMLHook, self).__init__(gcp_conn_id, delegate_to)
self._cloudml = self.get_conn()

def get_conn(self):
"""
Returns a Google CloudML service object.
"""
credentials = GoogleCredentials.get_application_default()
return build('ml', 'v1', credentials=credentials)

def create_version(self, project_name, model_name, version_spec):
"""
Creates the Version on Cloud ML.
Returns the operation if the version was created successfully and raises
an error otherwise.
"""
parent_name = 'projects/{}/models/{}'.format(project_name, model_name)
create_request = self._cloudml.projects().models().versions().create(
parent=parent_name, body=version_spec)
response = create_request.execute()
get_request = self._cloudml.projects().operations().get(
name=response['name'])

return _poll_with_exponential_delay(
request=get_request,
max_n=9,
is_done_func=lambda resp: resp.get('done', False),
is_error_func=lambda resp: resp.get('error', None) is not None)

def set_default_version(self, project_name, model_name, version_name):
"""
Sets a version to be the default. Blocks until finished.
"""
full_version_name = 'projects/{}/models/{}/versions/{}'.format(
project_name, model_name, version_name)
request = self._cloudml.projects().models().versions().setDefault(
name=full_version_name, body={})

try:
response = request.execute()
logging.info('Successfully set version: {} to default'.format(response))
return response
except errors.HttpError as e:
logging.error('Something went wrong: {}'.format(e))
raise e

def list_versions(self, project_name, model_name):
"""
Lists all available versions of a model. Blocks until finished.
"""
result = []
full_parent_name = 'projects/{}/models/{}'.format(
project_name, model_name)
request = self._cloudml.projects().models().versions().list(
parent=full_parent_name, pageSize=100)

response = request.execute()
next_page_token = response.get('nextPageToken', None)
result.extend(response.get('versions', []))
while next_page_token is not None:
next_request = self._cloudml.projects().models().versions().list(
parent=full_parent_name,
pageToken=next_page_token,
pageSize=100)
response = next_request.execute()
next_page_token = response.get('nextPageToken', None)
result.extend(response.get('versions', []))
time.sleep(5)
return result

def delete_version(self, project_name, model_name, version_name):
"""
Deletes the given version of a model. Blocks until finished.
"""
full_name = 'projects/{}/models/{}/versions/{}'.format(
project_name, model_name, version_name)
delete_request = self._cloudml.projects().models().versions().delete(
name=full_name)
response = delete_request.execute()
get_request = self._cloudml.projects().operations().get(
name=response['name'])

return _poll_with_exponential_delay(
request=get_request,
max_n=9,
is_done_func=lambda resp: resp.get('done', False),
is_error_func=lambda resp: resp.get('error', None) is not None)

def create_model(self, project_name, model):
"""
Create a Model. Blocks until finished.
"""
assert model['name'] is not None and model['name'] is not ''
project = 'projects/{}'.format(project_name)

request = self._cloudml.projects().models().create(
parent=project, body=model)
return request.execute()

def get_model(self, project_name, model_name):
"""
Gets a Model. Blocks until finished.
"""
assert model_name is not None and model_name is not ''
full_model_name = 'projects/{}/models/{}'.format(
project_name, model_name)
request = self._cloudml.projects().models().get(name=full_model_name)
try:
return request.execute()
except errors.HttpError as e:
if e.resp.status == 404:
logging.error('Model was not found: {}'.format(e))
return None
raise e
178 changes: 178 additions & 0 deletions airflow/contrib/operators/cloudml_operator.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,178 @@
#
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements. See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the 'License'); you may not use this file except in compliance with
# the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an 'AS IS' BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.


import logging
from airflow import settings
from airflow.contrib.hooks.gcp_cloudml_hook import CloudMLHook
from airflow.operators import BaseOperator
from airflow.utils.decorators import apply_defaults

logging.getLogger('GoogleCloudML').setLevel(settings.LOGGING_LEVEL)


class CloudMLVersionOperator(BaseOperator):
"""
Operator for managing a Google Cloud ML version.
:param model_name: The name of the Google Cloud ML model that the version
belongs to.
:type model_name: string
:param project_name: The Google Cloud project name to which CloudML
model belongs.
:type project_name: string
:param version: A dictionary containing the information about the version.
If the `operation` is `create`, `version` should contain all the
information about this version such as name, and deploymentUrl.
If the `operation` is `get` or `delete`, the `version` parameter
should contain the `name` of the version.
If it is None, the only `operation` possible would be `list`.
:type version: dict
:param gcp_conn_id: The connection ID to use when fetching connection info.
:type gcp_conn_id: string
:param operation: The operation to perform. Available operations are:
'create': Creates a new version in the model specified by `model_name`,
in which case the `version` parameter should contain all the
information to create that version
(e.g. `name`, `deploymentUrl`).
'get': Gets full information of a particular version in the model
specified by `model_name`.
The name of the version should be specified in the `version`
parameter.
'list': Lists all available versions of the model specified
by `model_name`.
'delete': Deletes the version specified in `version` parameter from the
model specified by `model_name`).
The name of the version should be specified in the `version`
parameter.
:type operation: string
:param delegate_to: The account to impersonate, if any.
For this to work, the service account making the request must have
domain-wide delegation enabled.
:type delegate_to: string
"""


template_fields = [
'_model_name',
'_version',
]

@apply_defaults
def __init__(self,
model_name,
project_name,
version=None,
gcp_conn_id='google_cloud_default',
operation='create',
delegate_to=None,
*args,
**kwargs):

super(CloudMLVersionOperator, self).__init__(*args, **kwargs)
self._model_name = model_name
self._version = version
self._gcp_conn_id = gcp_conn_id
self._delegate_to = delegate_to
self._project_name = project_name
self._operation = operation

def execute(self, context):
hook = CloudMLHook(
gcp_conn_id=self._gcp_conn_id, delegate_to=self._delegate_to)

if self._operation == 'create':
assert self._version is not None
return hook.create_version(self._project_name, self._model_name,
self._version)
elif self._operation == 'set_default':
return hook.set_default_version(
self._project_name, self._model_name,
self._version['name'])
elif self._operation == 'list':
return hook.list_versions(self._project_name, self._model_name)
elif self._operation == 'delete':
return hook.delete_version(self._project_name, self._model_name,
self._version['name'])
else:
raise ValueError('Unknown operation: {}'.format(self._operation))


class CloudMLModelOperator(BaseOperator):
"""
Operator for managing a Google Cloud ML model.
:param model: A dictionary containing the information about the model.
If the `operation` is `create`, then the `model` parameter should
contain all the information about this model such as `name`.
If the `operation` is `get`, the `model` parameter
should contain the `name` of the model.
:type model: dict
:param project_name: The Google Cloud project name to which CloudML
model belongs.
:type project_name: string
:param gcp_conn_id: The connection ID to use when fetching connection info.
:type gcp_conn_id: string
:param operation: The operation to perform. Available operations are:
'create': Creates a new model as provided by the `model` parameter.
'get': Gets a particular model where the name is specified in `model`.
:param delegate_to: The account to impersonate, if any.
For this to work, the service account making the request must have
domain-wide delegation enabled.
:type delegate_to: string
"""

template_fields = [
'_model',
]

@apply_defaults
def __init__(self,
model,
project_name,
gcp_conn_id='google_cloud_default',
operation='create',
delegate_to=None,
*args,
**kwargs):
super(CloudMLModelOperator, self).__init__(*args, **kwargs)
self._model = model
self._operation = operation
self._gcp_conn_id = gcp_conn_id
self._delegate_to = delegate_to
self._project_name = project_name

def execute(self, context):
hook = CloudMLHook(
gcp_conn_id=self._gcp_conn_id, delegate_to=self._delegate_to)
if self._operation == 'create':
hook.create_model(self._project_name, self._model)
elif self._operation == 'get':
hook.get_model(self._project_name, self._model['name'])
else:
raise ValueError('Unknown operation: {}'.format(self._operation))
4 changes: 4 additions & 0 deletions airflow/utils/db.py
Original file line number Diff line number Diff line change
Expand Up @@ -128,6 +128,10 @@ def initdb():
conn_id='presto_default', conn_type='presto',
host='localhost',
schema='hive', port=3400))
merge_conn(
models.Connection(
conn_id='google_cloud_default', conn_type='google_cloud_platform',
schema='default',))
merge_conn(
models.Connection(
conn_id='hive_cli_default', conn_type='hive_cli',
Expand Down
Loading

0 comments on commit 534a0e0

Please sign in to comment.