def tag_data_package(publisher, package):
"""
DPR metadata put operation.
This API is responsible for tagging data package
---
tags:
- package
parameters:
- in: path
name: publisher
type: string
required: true
description: publisher name
- in: path
name: package
type: string
required: true
description: package name
- in: body
name: version
type: string
required: true
description: version value
responses:
400:
description: JWT is invalid or req body is not valid
401:
description: Invalid Header for JWT
403:
description: User not allowed for operation
404:
description: User not found
500:
description: Internal Server Error
200:
description: Success Message
schema:
id: put_package_success
properties:
status:
type: string
description: Status of the operation
default: OK
"""
try:
data = request.get_json()
if 'version' not in data:
return handle_error('ATTRIBUTE_MISSING', 'version not found', 400)
bitstore = BitStore(publisher, package)
status_db = MetaDataDB.create_or_update_version(publisher, package, data['version'])
status_bitstore = bitstore.copy_to_new_version(data['version'])
if status_db is False or status_bitstore is False:
raise Exception("failed to tag data package")
return jsonify({"status": "OK"}), 200
except Exception as e:
app.logger.error(e)
return handle_error('GENERIC_ERROR', e.message, 500)
def test_return_none_if_no_readme_found(self):
with self.app.app_context():
bit_store = BitStore('test_pub', 'test_package')
s3 = boto3.client('s3')
bucket_name = self.app.config['S3_BUCKET_NAME']
s3.create_bucket(Bucket=bucket_name)
read_me_key = bit_store.build_s3_key('test.md')
s3.put_object(Bucket=bucket_name, Key=read_me_key, Body='')
self.assertEqual(bit_store.get_readme_object_key(), None)
def get_s3_signed_url():
"""
This API is responsible for generate signed url to post data to S3
---
tags:
- auth
parameters:
- in: body
name: publisher
type: string
required: true
description: publisher name
- in: body
name: package
type: string
required: true
description: package name
- in: body
name: path
type: string
required: true
description: relative path of the resources
responses:
200:
description: Success
schema:
id: get_signed_url
properties:
key:
type: string
description: signed url for post data to S3
400:
description: Publisher or package can not be empty
500:
description: Internal Server Error
"""
try:
data = request.get_json()
publisher = data.get('publisher', None)
package = data.get('package', None)
path = data.get('path', None)
md5 = data.get('md5', None)
if publisher is None or package is None:
return handle_error('INVALID_INPUT',
'publisher or package can not be empty',
400)
if md5 is None:
return handle_error('INVALID_INPUT',
'md5 hash can not be empty',
400)
metadata = BitStore(publisher=publisher, package=package)
url = metadata.generate_pre_signed_put_obj_url(path, md5)
return jsonify({'key': url}), 200
except Exception as e:
app.logger.error(e)
return handle_error('GENERIC_ERROR', e.message, 500)
def finalize_metadata(publisher, package):
"""
DPR metadata finalize operation.
This API is responsible for getting data from S3 and push it to RDS.
---
tags:
- package
parameters:
- in: path
name: publisher
type: string
required: true
description: publisher name
- in: path
name: package
type: string
required: true
description: package name
responses:
200:
description: Data transfer complete
400:
description: JWT is invalid
401:
description: Invalid Header for JWT
403:
description: User name and publisher not matched
404:
description: User not found
500:
description: Internal Server Error
"""
try:
user = _request_ctx_stack.top.current_user
user_id = user['user']
user = User.query.filter_by(id=user_id).first()
if user is not None:
if user.name == publisher:
bit_store = BitStore(publisher, package)
body = bit_store.get_metadata_body()
if body is not None:
readme = bit_store.get_s3_object(
bit_store.get_readme_object_key())
MetaDataDB.create_or_update(name=package,
publisher_name=publisher,
descriptor=body,
readme=readme)
return jsonify({"status": "OK"}), 200
raise Exception("Failed to get data from s3")
return handle_error('NOT_PERMITTED',
'user name and publisher not matched', 403)
return handle_error('USER_NOT_FOUND', 'user not found', 404)
except Exception as e:
app.logger.error(e)
return handle_error('GENERIC_ERROR', e.message, 500)
def test_get_metadata_body(self):
with self.app.app_context():
s3 = boto3.client('s3')
bucket_name = self.app.config['S3_BUCKET_NAME']
s3.create_bucket(Bucket=bucket_name)
metadata = BitStore(publisher="pub_test",
package="test_package",
body='hi')
s3.put_object(Bucket=bucket_name,
Key=metadata.build_s3_key('datapackage.json'),
Body=metadata.body)
self.assertEqual(metadata.body, metadata.get_metadata_body())
def test_get_empty_metadata_name_for_publisher(self):
with self.app.app_context():
s3 = boto3.client('s3')
bucket_name = self.app.config['S3_BUCKET_NAME']
s3.create_bucket(Bucket=bucket_name)
metadata = BitStore(publisher="pub_test",
package="test_package",
body='hi')
s3.put_object(Bucket=bucket_name,
Key='test/key.json',
Body=metadata.body)
self.assertEqual(
0, len(metadata.get_all_metadata_name_for_publisher()))
def test_save(self):
with self.app.app_context():
s3 = boto3.client('s3')
bucket_name = self.app.config['S3_BUCKET_NAME']
s3.create_bucket(Bucket=bucket_name)
metadata = BitStore(publisher="pub_test",
package="test_package",
body='hi')
key = metadata.build_s3_key('datapackage.json')
metadata.save()
obs_list = list(s3.list_objects(Bucket=bucket_name, Prefix=key).\
get('Contents'))
self.assertEqual(1, len(obs_list))
self.assertEqual(key, obs_list[0]['Key'])
def test_generate_pre_signed_put_obj_url(self):
with self.app.app_context():
s3 = boto3.client('s3')
bucket_name = self.app.config['S3_BUCKET_NAME']
s3.create_bucket(Bucket=bucket_name)
metadata = BitStore(publisher="pub_test",
package="test_package",
body='hi')
url = metadata.generate_pre_signed_put_obj_url(
'datapackage.json', 'm')
parsed = urlparse(url)
self.assertEqual(
parsed.netloc, 's3-{region}.amazonaws.com'.format(
region=self.app.config['AWS_REGION']))
def populate_data(publisher_name):
data = json.loads(open('fixtures/datapackage.json').read())
data_csv = open('fixtures//data/demo-resource.csv').read()
readme = open('fixtures/README.md').read()
publisher = Publisher.query.filter_by(name=publisher_name).one()
metadata = MetaDataDB(name="demo-package")
metadata.descriptor, metadata.status, metadata.private, metadata.readme \
= json.dumps(data), 'active', False, readme
publisher.packages.append(metadata)
db.session.add(publisher)
db.session.commit()
bitstore = BitStore(publisher_name, package='demo-package')
key = bitstore.build_s3_key('demo-resource.csv')
bucket_name = app.config['S3_BUCKET_NAME']
s3_client = app.config['S3']
s3_client.put_object(Bucket=bucket_name, Key=key, Body=data_csv)
def purge_data_package(publisher, package):
"""
DPR data package hard delete operation.
This API is responsible for deletion of data package
---
tags:
- package
parameters:
- in: path
name: publisher
type: string
required: true
description: publisher name
- in: path
name: package
type: string
required: true
description: package name
responses:
500:
description: Internal Server Error
200:
description: Success Message
schema:
id: put_package_success
properties:
status:
type: string
default: OK
"""
try:
bitstore = BitStore(publisher=publisher, package=package)
status_acl = bitstore.delete_data_package()
status_db = MetaDataDB.delete_data_package(publisher, package)
if status_acl and status_db:
return jsonify({"status": "OK"}), 200
if not status_acl:
raise Exception('Failed to delete from s3')
if not status_db:
raise Exception('Failed to delete from db')
except Exception as e:
app.logger.error(e)
return handle_error('GENERIC_ERROR', e.message, 500)
def test_change_acl(self):
with self.app.app_context():
bit_store = BitStore('test_pub', 'test_package')
s3 = boto3.client('s3')
bucket_name = self.app.config['S3_BUCKET_NAME']
s3.create_bucket(Bucket=bucket_name)
read_me_key = bit_store.build_s3_key('test.md')
data_key = bit_store.build_s3_key('data.csv')
metadata_key = bit_store.build_s3_key('datapackage.json')
s3.put_object(Bucket=bucket_name, Key=read_me_key, Body='')
s3.put_object(Bucket=bucket_name, Key=data_key, Body='')
s3.put_object(Bucket=bucket_name, Key=metadata_key, Body='')
bit_store.change_acl("private")
res = s3.get_object_acl(Bucket=bucket_name, Key=read_me_key)
grant = res['Grants'][0]['Permission']
self.assertEqual(grant, 'FULL_CONTROL')
def test_throw_403_if_not_owner_or_member_of_publisher(self):
s3 = boto3.client('s3')
s3.create_bucket(Bucket=self.bucket_name)
bit_store = BitStore('test_pub', 'test_package')
read_me_key = bit_store.build_s3_key('test.md')
data_key = bit_store.build_s3_key('data.csv')
metadata_key = bit_store.build_s3_key('datapackage.json')
s3.put_object(Bucket=self.bucket_name, Key=read_me_key, Body='readme')
s3.put_object(Bucket=self.bucket_name, Key=data_key, Body='data')
s3.put_object(Bucket=self.bucket_name,
Key=metadata_key,
Body='metedata')
response = self.client.post(self.jwt_url,
data=json.dumps({
'username': self.user_not_allowed_name,
'secret': 'super_secret'
}),
content_type='application/json')
data = json.loads(response.data)
jwt_not_allowed = data['token']
auth_not_allowed = "bearer %s" % jwt_not_allowed
response = self.client.post(
self.url,
data=json.dumps({'version': 'tag_one'}),
content_type='application/json',
headers=dict(Authorization=auth_not_allowed))
self.assertEqual(response.status_code, 403)
with self.app.app_context():
data_latest = MetaDataDB.query.join(Publisher). \
filter(Publisher.name == self.publisher_name,
MetaDataDB.name == self.package).all()
self.assertEqual(1, len(data_latest))
bit_store_tagged = BitStore('test_pub', 'test_package', 'tag_one')
objects_nu = s3.list_objects(
Bucket=self.bucket_name,
Prefix=bit_store_tagged.build_s3_versioned_prefix())
self.assertTrue('Contents' not in objects_nu)
def test_delete_data_package(self):
with self.app.app_context():
bit_store = BitStore('test_pub', 'test_package')
s3 = boto3.client('s3')
bucket_name = self.app.config['S3_BUCKET_NAME']
s3.create_bucket(Bucket=bucket_name)
read_me_key = bit_store.build_s3_key('test.md')
data_key = bit_store.build_s3_key('data.csv')
metadata_key = bit_store.build_s3_key('datapackage.json')
s3.put_object(Bucket=bucket_name, Key=read_me_key, Body='readme')
s3.put_object(Bucket=bucket_name, Key=data_key, Body='data')
s3.put_object(Bucket=bucket_name,
Key=metadata_key,
Body='metedata')
status = bit_store.delete_data_package()
read_me_res = s3.list_objects(Bucket=bucket_name,
Prefix=read_me_key)
self.assertTrue('Contents' not in read_me_res)
data_res = s3.list_objects(Bucket=bucket_name, Prefix=data_key)
self.assertTrue('Contents' not in data_res)
self.assertTrue(status)
def test_metadata_s3_key(self):
metadata = BitStore(publisher="pub_test", package="test_package")
expected = "{t}/pub_test/test_package/_v/latest/datapackage.json".\
format(t=metadata.prefix)
self.assertEqual(expected, metadata.build_s3_key('datapackage.json'))
def save_metadata(publisher, package):
"""
DPR metadata put operation.
This API is responsible for pushing datapackage.json to S3.
---
tags:
- package
parameters:
- in: path
name: publisher
type: string
required: true
description: publisher name
- in: path
name: package
type: string
required: true
description: package name
responses:
400:
description: JWT is invalid or req body is not valid
401:
description: Invalid Header for JWT
403:
description: User name and publisher not matched
404:
description: User not found
500:
description: Internal Server Error
200:
description: Success Message
schema:
id: put_package_success
properties:
status:
type: string
description: Status of the operation
default: OK
"""
try:
user = _request_ctx_stack.top.current_user
user_id = user['user']
user = User.query.filter_by(id=user_id).first()
if user is not None:
if user.name == publisher:
metadata = BitStore(publisher=publisher,
package=package,
body=request.data)
is_valid = metadata.validate()
if not is_valid:
return handle_error('INVALID_DATA',
'Missing required field in metadata',
400)
metadata.save()
return jsonify({"status": "OK"}), 200
return handle_error('NOT_PERMITTED',
'user name and publisher not matched', 403)
return handle_error('USER_NOT_FOUND', 'user not found', 404)
except Exception as e:
app.logger.error(e)
return handle_error('GENERIC_ERROR', e.message, 500)
def get_resource(publisher, package, resource):
"""
DPR resource get operation.
This API is responsible for getting resource from S3.
---
tags:
- package
parameters:
- in: path
name: publisher
type: string
required: true
description: publisher name
- in: path
name: package
type: string
required: true
description: package name - to retrieve the data package metadata
- in: path
name: resource
type: string
required: true
description: resource index or name
responses:
200:
description: Get Data package for one key
schema:
id: get_data_package
properties:
data:
type: string
description: The resource
500:
description: Internal Server Error
"""
try:
path = request.path
metadata = BitStore(publisher, package)
if path.endswith('csv'):
resource_key = metadata.build_s3_key(resource + '.csv')
data = metadata.get_s3_object(resource_key)
def generate():
for row in data.splitlines():
yield row + '\n'
return Response(generate()), 200
else:
resource_key = metadata.build_s3_key(resource + '.csv')
data = metadata.get_s3_object(resource_key)
data = csv.DictReader(data.splitlines())
# taking first and adding at the end to avoid last comma
first_row = next(data)
def generate():
yield '['
for row in data:
yield json.dumps(row) + ','
yield json.dumps(first_row) + ']'
return Response(generate(), content_type='application/json'), 200
except Exception as e:
return handle_error('GENERIC_ERROR', e.message, 500)
def test_should_copy_all_object_from_latest_to_tag(self):
numeric_version = 0.8
with self.app.app_context():
bit_store = BitStore('test_pub', 'test_package')
s3 = boto3.client('s3')
bucket_name = self.app.config['S3_BUCKET_NAME']
s3.create_bucket(Bucket=bucket_name)
read_me_key = bit_store.build_s3_key('test.md')
data_key = bit_store.build_s3_key('data.csv')
metadata_key = bit_store.build_s3_key('datapackage.json')
s3.put_object(Bucket=bucket_name, Key=read_me_key, Body='readme')
s3.put_object(Bucket=bucket_name, Key=data_key, Body='data')
s3.put_object(Bucket=bucket_name,
Key=metadata_key,
Body='metedata')
bit_store.copy_to_new_version(numeric_version)
bit_store_numeric = BitStore('test_pub', 'test_package',
numeric_version)
objects_nu = s3.list_objects(
Bucket=bucket_name,
Prefix=bit_store_numeric.build_s3_versioned_prefix())
objects_old = s3.list_objects(
Bucket=bucket_name,
Prefix=bit_store.build_s3_versioned_prefix())
self.assertEqual(len(objects_nu['Contents']),
len(objects_old['Contents']))
def test_metadata_s3_prefix(self):
metadata = BitStore(publisher="pub_test", package="test_package")
expected = "{t}/pub_test".format(t=metadata.prefix)
self.assertEqual(expected, metadata.build_s3_prefix())