def infer_dataset_data(s3_path, max_lines):
"""
:type s3_path: str
:param s3_path: The full s3 path to a sample delimited dataset
file.
:type max_lines: int
:param max_lines: The maximum number of lines to peek.
"""
guesses, has_header, compression, dialect = get_guesses(s3_path, max_lines)
columns = columns_with_best_guess(guesses, has_header)
location = guess_location(s3_path)
return Dataset(data=DatasetData(name='guessed_dataset',
table_name='public',
data_format=DataFormat(
file_format=FileFormat.TEXTFILE,
row_format=RowFormat.DELIMITED,
delimited_by=dialect.delimiter,
escaped_by=dialect.escapechar,
quoted_by=dialect.quotechar,
),
location=location,
columns=columns,
compression=compression,
load_type=LoadType.INSERT))
def test_crud(self):
columns = [
Column('c1', DataType.VARCHAR, 50),
Column('c2', DataType.BIGINT)
]
df = DataFormat(FileFormat.PARQUET, RowFormat.NONE)
ds = Dataset(
data=DatasetData(NoOpActionTypes.action_that_succeeds.name,
NoOpActionTypes.action_that_succeeds.name,
's3://bucket/prefix',
df,
columns,
tags=['foo']))
posted_dataset = self.dart.save_dataset(ds)
self.assertEqual(posted_dataset.data.to_dict(), ds.data.to_dict())
dataset = self.dart.get_dataset(posted_dataset.id)
self.assertEqual(posted_dataset.to_dict(), dataset.to_dict())
dataset.data.compression = Compression.GZIP
put_dataset = self.dart.save_dataset(dataset)
self.assertEqual(put_dataset.data.compression, Compression.GZIP)
self.assertNotEqual(posted_dataset.to_dict(), put_dataset.to_dict())
self.dart.delete_dataset(dataset.id)
try:
self.dart.get_dataset(dataset.id)
except DartRequestException as e:
self.assertEqual(e.response.status_code, 404)
return
self.fail('dataset should have been missing after delete!')
def test_dataset_schema(self):
columns = [Column('c1', DataType.VARCHAR, 50), Column('c2', DataType.BIGINT)]
num_header_rows = None
df = DataFormat(FileFormat.PARQUET, RowFormat.NONE, num_header_rows)
ds = Dataset(data=DatasetData('test-dataset', 'test_dataset_table', 's3://bucket/prefix', df, columns))
obj_before = ds.to_dict()
obj_after = default_and_validate(ds, dataset_schema()).to_dict()
# num_header_rows should have been defaulted to 0, making these unequal
self.assertNotEqual(obj_before, obj_after)
def test_dataset_schema_invalid(self):
with self.assertRaises(DartValidationException) as context:
columns = [Column('c1', DataType.VARCHAR, 50), Column('c2', DataType.BIGINT)]
df = DataFormat(FileFormat.PARQUET, RowFormat.NONE)
location = None
ds = Dataset(data=DatasetData('test-dataset', 'test_dataset_table', location, df, columns))
# should fail because location is required
default_and_validate(ds, dataset_schema())
self.assertTrue(isinstance(context.exception, DartValidationException))
def test_impala_table_definition_step(self):
ds = Dataset(data=DatasetData(
name='weblogs_v01',
table_name='weblogs_parquet',
location='s3://wsm-log-servers/weblogs/www.retailmenot.com/ec2/',
data_format=DataFormat(
file_format=FileFormat.PARQUET,
row_format=RowFormat.NONE,
),
columns=[
Column('ip', 'STRING'),
Column('user', 'STRING'),
Column('requestDate', 'TIMESTAMP', date_pattern='dd/MMM/yyyy:HH:mm:ss Z'),
Column('httpMethod', 'STRING'),
Column('urlPath', 'STRING'),
Column('queryString', 'STRING'),
Column('httpVersion', 'STRING'),
Column('statusCode', 'STRING'),
Column('bytesSent', 'INT'),
Column('referrer', 'STRING'),
Column('userAgent', 'STRING'),
Column('responseTime', 'BIGINT'),
Column('hostname', 'STRING'),
Column('userFingerprint', 'STRING'),
Column('userId', 'STRING'),
Column('sessionId', 'STRING'),
Column('requestId', 'STRING'),
Column('visitorId', 'STRING'),
Column('vegSlice', 'STRING'),
Column('fruitSlice', 'STRING'),
Column('cacheHitMiss', 'STRING'),
],
compression='GZIP',
partitions=[
Column('year', 'STRING'),
Column('week', 'STRING'),
Column('day', 'STRING'),
],
))
call('mkdir -p /tmp/dart-emr-test/impala/')
this_path = os.path.dirname(os.path.abspath(__file__))
shutil.copyfile(this_path + '/../../../engine/emr/steps/impala/copy_to_table.sql', '/tmp/dart-emr-test/impala/copy_to_table.sql')
impala_copy_to_table(ds, 'weblogs_stage', ds, 'weblogs_parquet', 's3://test', '/tmp/dart-emr-test/', 'actionid123', 1, 1)
with open(os.path.join(this_path, 'copy_to_table_weblogs_parquet.sql')) as f:
expected_contents = f.read()
with open('/tmp/dart-emr-test/impala/copy_to_table_weblogs_parquet.sql') as f:
actual_contents = f.read()
self.assertEqual(expected_contents, actual_contents)
def setUp(self):
dart = Dart(host='localhost', port=5000)
""" :type dart: dart.client.python.dart_client.Dart """
self.dart = dart
cs = [
Column('c1', DataType.VARCHAR, 50),
Column('c2', DataType.BIGINT)
]
df = DataFormat(FileFormat.TEXTFILE, RowFormat.DELIMITED)
dataset_data = DatasetData(
name='test-dataset',
table_name='test_dataset_table',
load_type=LoadType.INSERT,
location=('s3://' + os.environ['DART_TEST_BUCKET'] + '/impala'),
data_format=df,
columns=cs,
tags=[])
self.dataset = self.dart.save_dataset(Dataset(data=dataset_data))
start = 's3://' + os.environ['DART_TEST_BUCKET'] + '/impala/impala'
end = 's3://' + os.environ['DART_TEST_BUCKET'] + '/impala/install'
regex = '.*\\.rpm'
ds = Subscription(data=SubscriptionData(
'test-subscription', self.dataset.id, start, end, regex))
self.subscription = self.dart.save_subscription(ds)
dst_args = {'action_sleep_time_in_seconds': 0}
dst = Datastore(data=DatastoreData('test-datastore',
'no_op_engine',
args=dst_args,
state=DatastoreState.TEMPLATE))
self.datastore = self.dart.save_datastore(dst)
wf = Workflow(data=WorkflowData(
'test-workflow', self.datastore.id, state=WorkflowState.ACTIVE))
self.workflow = self.dart.save_workflow(wf, self.datastore.id)
a_args = {'subscription_id': self.subscription.id}
a0 = Action(data=ActionData(NoOpActionTypes.action_that_succeeds.name,
NoOpActionTypes.action_that_succeeds.name,
state=ActionState.TEMPLATE))
a1 = Action(data=ActionData(NoOpActionTypes.consume_subscription.name,
NoOpActionTypes.consume_subscription.name,
a_args,
state=ActionState.TEMPLATE))
self.action0, self.action1 = self.dart.save_actions(
[a0, a1], workflow_id=self.workflow.id)
data=(DatasetData(
name='beacon_native_app_parsed_gzipped_v03',
table_name='beacon_native_app',
location=
's3://example-bucket/prd/beacon/native_app/v3/dwh-delimited/gzipped',
load_type=LoadType.INSERT,
distribution_keys=['created'],
sort_keys=['created', 'eventtype'],
hive_compatible_partition_folders=True,
data_format=DataFormat(
FileFormat.TEXTFILE,
RowFormat.DELIMITED,
delimited_by='\t',
quoted_by='"',
escaped_by='\\',
null_string='NULL',
),
compression=Compression.GZIP,
partitions=[Column('createdpartition', DataType.STRING)],
columns=[
Column('logfileid', DataType.INT),
Column('linenumber', DataType.INT),
Column('created',
DataType.TIMESTAMP,
date_pattern="yyyy-MM-dd HH:mm:ss"),
Column('remoteip', DataType.VARCHAR, 500),
Column('useragent', DataType.VARCHAR, 2500),
Column('eventtype', DataType.VARCHAR, 255),
Column('appversion', DataType.VARCHAR, 255),
Column('advertiserid', DataType.VARCHAR, 2048),
Column('couponsonpage', DataType.INT),
Column('coupons', DataType.VARCHAR, 10000),
Column('channel', DataType.VARCHAR, 128),
Column('geocouponcount', DataType.BIGINT),
Column('geofence', DataType.VARCHAR, 255),
Column('geofencetimespent',
DataType.NUMERIC,
precision=14,
scale=5),
Column('loginstatus', DataType.VARCHAR, 25),
Column('products', DataType.VARCHAR, 2500),
Column('session', DataType.VARCHAR, 5000),
Column('systemname', DataType.VARCHAR, 2500),
Column('systemversion', DataType.VARCHAR, 2500),
Column('udid', DataType.VARCHAR, 128),
Column('userqualifier', DataType.VARCHAR, 64),
Column('url', DataType.VARCHAR, 5000),
Column('user_uuid', DataType.VARCHAR, 64),
Column('userid', DataType.INT),
Column('searchtype', DataType.VARCHAR, 128),
Column('searchlistterm', DataType.VARCHAR, 512),
Column('searchterm', DataType.VARCHAR, 512),
Column('emailuuid', DataType.VARCHAR, 128),
Column('userfingerprint', DataType.VARCHAR, 64),
Column('locationstatus', DataType.VARCHAR, 128),
Column('pushnotificationstatus', DataType.VARCHAR, 128),
Column('placement', DataType.VARCHAR, 1024),
Column('loc', DataType.VARCHAR, 128),
Column('ppoi0', DataType.VARCHAR, 128),
Column('ppoi1', DataType.VARCHAR, 128),
Column('ppoi2', DataType.VARCHAR, 128),
Column('ppoi3', DataType.VARCHAR, 128),
Column('ppoi4', DataType.VARCHAR, 128),
Column('applaunchnotificationtype', DataType.VARCHAR, 128),
Column('scenarioname', DataType.VARCHAR, 128),
Column('behaviorname', DataType.VARCHAR, 128),
Column('coupontype', DataType.VARCHAR, 128),
Column('couponposition', DataType.VARCHAR, 128),
Column('hasqsrcontent', DataType.VARCHAR, 128),
Column('promptname', DataType.VARCHAR, 128),
Column('locationpermissionchanage', DataType.VARCHAR, 128),
Column('couponproblemtype', DataType.VARCHAR, 128),
Column('storetitle', DataType.VARCHAR, 128),
Column('mallname', DataType.VARCHAR, 128),
Column('restaurantname', DataType.VARCHAR, 128),
Column('milesaway', DataType.VARCHAR, 128),
Column('menuitem', DataType.VARCHAR, 128),
Column('toolname', DataType.VARCHAR, 128),
Column('toolaction', DataType.VARCHAR, 128),
Column('toolstep', DataType.VARCHAR, 128),
Column('mallposition', DataType.VARCHAR, 128),
Column('recommendstorename', DataType.VARCHAR, 128),
Column('recommendstoreposition', DataType.VARCHAR, 128),
Column('favoritestorename', DataType.VARCHAR, 128),
Column('favoritestoreaction', DataType.VARCHAR, 128),
Column('favoritestoreposition', DataType.VARCHAR, 128),
Column('favoritesiteid', DataType.VARCHAR, 128),
Column('receivername', DataType.VARCHAR, 128),
Column('outclickbuttonprompt', DataType.VARCHAR, 128),
Column('datasource', DataType.VARCHAR, 1024),
Column('searchresultcount', DataType.VARCHAR, 128),
Column('searchresultposition', DataType.VARCHAR, 128),
Column('sharetype', DataType.VARCHAR, 128),
Column('daysuntilexpiration', DataType.VARCHAR, 128),
Column('firedate', DataType.VARCHAR, 128),
Column('settingschangevalue', DataType.VARCHAR, 128),
Column('settingschangetype', DataType.VARCHAR, 128),
Column('settingschangelocation', DataType.VARCHAR, 128),
Column('clickaction', DataType.VARCHAR, 128),
Column('tnt', DataType.VARCHAR, 128),
Column('previouspage', DataType.VARCHAR, 2500),
Column('clickpage', DataType.VARCHAR, 2500),
Column('launchreason', DataType.VARCHAR, 128),
Column('taplyticsData', DataType.VARCHAR, 150),
Column('appCampaign', DataType.VARCHAR, 50),
Column('accountMethod', DataType.VARCHAR, 60),
Column('appState', DataType.VARCHAR, 100),
Column('btStatus', DataType.BOOLEAN),
Column('btBeaconId', DataType.VARCHAR, 500),
Column('btBeaconFactoryId', DataType.VARCHAR, 500),
Column('btBeaconName', DataType.VARCHAR, 500),
Column('btTimeSpent', DataType.VARCHAR, 50),
Column('purchaseId', DataType.VARCHAR, 500),
Column('transactionId', DataType.VARCHAR, 500),
Column('outclickLink', DataType.VARCHAR, 1000),
Column('outclickPage', DataType.VARCHAR, 300),
Column('featuredCouponPosition', DataType.INT),
Column('commentCount', DataType.INT),
Column('mallCount', DataType.INT),
Column('clickCount', DataType.INT),
Column('merchantName', DataType.VARCHAR, 100),
Column('merchantPosition', DataType.INT),
Column('couponUuids', DataType.VARCHAR, 10000),
Column('favoriteSiteUuid', DataType.VARCHAR, 50),
Column('deepLinkType', DataType.VARCHAR, 40),
Column('adUnitUuid', DataType.VARCHAR, 50),
],
))))
Dataset(data=(DatasetData(
name='owen_eu_DW-3411_v1',
table_name='owen_eu',
location='s3://example-bucket/prd/inbound/overlord/eu-all-events',
load_type=LoadType.MERGE,
data_format=DataFormat(
file_format=FileFormat.TEXTFILE,
row_format=RowFormat.JSON,
),
compression=Compression.GZIP,
partitions=[
Column('year', DataType.STRING),
Column('month', DataType.STRING),
Column('day', DataType.STRING),
],
columns=[
Column('host', DataType.STRING, path='metadata.host'),
Column('referer', DataType.STRING, path='metadata.referer'),
Column(
'eventName', DataType.STRING, path='owen.event.eventName'),
Column('eventVersion',
DataType.STRING,
path='owen.event.eventVersion'),
Column('eventPlatform',
DataType.STRING,
path='owen.event.eventPlatform'),
Column('eventInstanceUuid',
DataType.STRING,
path='owen.event.eventInstanceUuid'),
Column('eventCategory',
DataType.STRING,
path='owen.event.eventCategory'),
Column('eventTimestamp',
DataType.TIMESTAMP,
path='owen.event.eventTimestamp',
date_pattern="yyyy-MM-dd'T'HH:mm:ss'Z'"),
Column('eventTarget',
DataType.STRING,
path='owen.event.eventTarget'),
Column('eventAction',
DataType.STRING,
path='owen.event.eventAction'),
Column('eventPlatformVersion',
DataType.STRING,
path='owen.event.eventPlatformVersion'),
Column('osName', DataType.STRING, path='owen.context.osName'),
Column('loggedInFlag',
DataType.BOOLEAN,
path='owen.context.loggedInFlag'),
Column('custom', DataType.STRING, path='owen.context.custom'),
Column('browserVersion',
DataType.STRING,
path='owen.context.browserVersion'),
Column(
'referrer', DataType.STRING, path='owen.context.referrer'),
Column('previousPageName',
DataType.STRING,
path='owen.context.previousPageName'),
Column('screenHeight',
DataType.STRING,
path='owen.context.screenHeight'),
Column('breadCrumb',
DataType.STRING,
path='owen.context.breadCrumb'),
Column(
'pageName', DataType.STRING, path='owen.context.pageName'),
Column('country', DataType.STRING,
path='owen.context.country'),
Column('propertyName',
DataType.STRING,
path='owen.context.propertyName'),
Column('launchCount',
DataType.STRING,
path='owen.context.launchCount'),
Column('viewInstanceUuid',
DataType.STRING,
path='owen.context.viewInstanceUuid'),
Column('osVersion',
DataType.STRING,
path='owen.context.osVersion'),
Column('connectionType',
DataType.STRING,
path='owen.context.connectionType'),
Column('partialSearchTerm',
DataType.STRING,
path='owen.context.partialSearchTerm'),
Column('carrier', DataType.STRING,
path='owen.context.carrier'),
Column('longitude',
DataType.STRING,
path='owen.context.longitude'),
Column(
'productSectionPosition_0',
DataType.STRING,
path='owen.context.inventory[0].productSectionPosition'),
Column('savedFlag_0',
DataType.BOOLEAN,
path='owen.context.inventory[0].savedFlag'),
Column('position_0',
DataType.STRING,
path='owen.context.inventory[0].position'),
Column('brand_0',
DataType.STRING,
path='owen.context.inventory[0].brand'),
Column('affiliateNetwork_0',
DataType.STRING,
path='owen.context.inventory[0].affiliateNetwork'),
Column('deepLinkUrl_0',
DataType.STRING,
path='owen.context.inventory[0].deepLinkUrl'),
Column('conquestingFlag_0',
DataType.BOOLEAN,
path='owen.context.inventory[0].conquestingFlag'),
Column('originalPrice_0',
DataType.STRING,
path='owen.context.inventory[0].originalPrice'),
Column('adUnitUuid_0',
DataType.STRING,
path='owen.context.inventory[0].adUnitUuid'),
Column('startDate_0',
DataType.TIMESTAMP,
path='owen.context.inventory[0].startDate',
date_pattern="yyyy-MM-dd'T'HH:mm:ss'Z'"),
Column('proximityUnit_0',
DataType.STRING,
path='owen.context.inventory[0].proximityUnit'),
Column('commentsCount_0',
DataType.STRING,
path='owen.context.inventory[0].commentsCount'),
Column('outRedirectUrl_0',
DataType.STRING,
path='owen.context.inventory[0].outRedirectUrl'),
Column('productCardPosition_0',
DataType.STRING,
path='owen.context.inventory[0].productCardPosition'),
Column('productSectionUuid_0',
DataType.STRING,
path='owen.context.inventory[0].productSectionUuid'),
Column('lastVerifiedDate_0',
DataType.TIMESTAMP,
path='owen.context.inventory[0].lastVerifiedDate',
date_pattern="yyyy-MM-dd'T'HH:mm:ss'Z'"),
Column('productCardUuid_0',
DataType.STRING,
path='owen.context.inventory[0].productCardUuid'),
Column('redemptionChannel_0',
DataType.STRING,
path='owen.context.inventory[0].redemptionChannel'),
Column('noVotes_0',
DataType.STRING,
path='owen.context.inventory[0].noVotes'),
Column('retailCategory_0',
DataType.STRING,
path='owen.context.inventory[0].retailCategory'),
Column('couponRank_0',
DataType.STRING,
path='owen.context.inventory[0].couponRank'),
Column('inventoryChannel_0',
DataType.STRING,
path='owen.context.inventory[0].inventoryChannel'),
Column('yesVotes_0',
DataType.STRING,
path='owen.context.inventory[0].yesVotes'),
Column('inventorySource_0',
DataType.STRING,
path='owen.context.inventory[0].inventorySource'),
Column('inventoryName_0',
DataType.STRING,
path='owen.context.inventory[0].inventoryName'),
Column('monetizableFlag_0',
DataType.BOOLEAN,
path='owen.context.inventory[0].monetizableFlag'),
Column('recommendedFlag_0',
DataType.BOOLEAN,
path='owen.context.inventory[0].recommendedFlag'),
Column('expirationDate_0',
DataType.TIMESTAMP,
path='owen.context.inventory[0].expirationDate',
date_pattern="yyyy-MM-dd'T'HH:mm:ss'Z'"),
Column('clickLocation_0',
DataType.STRING,
path='owen.context.inventory[0].clickLocation'),
Column('finalPrice_0',
DataType.STRING,
path='owen.context.inventory[0].finalPrice'),
Column('usedByCount_0',
DataType.STRING,
path='owen.context.inventory[0].usedByCount'),
Column('proximity_0',
DataType.STRING,
path='owen.context.inventory[0].proximity'),
Column('inventoryUuid_0',
DataType.STRING,
path='owen.context.inventory[0].inventoryUuid'),
Column('siteUuid_0',
DataType.STRING,
path='owen.context.inventory[0].siteUuid'),
Column('outclickUuid_0',
DataType.STRING,
path='owen.context.inventory[0].outclickUuid'),
Column('adUnitType_0',
DataType.STRING,
path='owen.context.inventory[0].adUnitType'),
Column('exclusivityFlag_0',
DataType.BOOLEAN,
path='owen.context.inventory[0].exclusivityFlag'),
Column('inventoryType_0',
DataType.STRING,
path='owen.context.inventory[0].inventoryType'),
Column('successPercentage_0',
DataType.STRING,
path='owen.context.inventory[0].successPercentage'),
Column('claimUuid_0',
DataType.STRING,
path='owen.context.inventory[0].claimUuid'),
Column('region', DataType.STRING, path='owen.context.region'),
Column('session', DataType.STRING,
path='owen.context.session'),
Column('content',
DataType.STRING,
path='owen.context.marketing.content'),
Column('marketingVendor',
DataType.STRING,
path='owen.context.marketing.vendor'),
Column('campaign',
DataType.STRING,
path='owen.context.marketing.campaign'),
Column('adGroup',
DataType.STRING,
path='owen.context.marketing.adGroup'),
Column('campaignUuid',
DataType.STRING,
path='owen.context.marketing.campaignUuid'),
Column('campaignSendCount',
DataType.STRING,
path='owen.context.marketing.campaignSendCount'),
Column('source',
DataType.STRING,
path='owen.context.marketing.source'),
Column('term',
DataType.STRING,
path='owen.context.marketing.term'),
Column('channel',
DataType.STRING,
path='owen.context.marketing.channel'),
Column('medium',
DataType.STRING,
path='owen.context.marketing.medium'),
Column('cdRank',
DataType.STRING,
path='owen.context.marketing.cdRank'),
Column('notificationUuid',
DataType.STRING,
path='owen.context.marketing.notificationUuid'),
Column('inventoryCount',
DataType.STRING,
path='owen.context.inventoryCount'),
Column('favoriteFlag',
DataType.BOOLEAN,
path='owen.context.favoriteFlag'),
Column(
'pageType', DataType.STRING, path='owen.context.pageType'),
Column('bluetoothBeaconType',
DataType.STRING,
path='owen.context.bluetoothBeaconType'),
Column('variation_0',
DataType.STRING,
path='owen.context.experiment[0].variation'),
Column('campaign_0',
DataType.STRING,
path='owen.context.experiment[0].campaign'),
Column('locationEnabledFlag',
DataType.BOOLEAN,
path='owen.context.locationEnabledFlag'),
Column('macAddress',
DataType.STRING,
path='owen.context.macAddress'),
Column('browserFamily',
DataType.STRING,
path='owen.context.browserFamily'),
Column('geofenceUuid',
DataType.STRING,
path='owen.context.geofenceUuid'),
Column('mobileDeviceMake',
DataType.STRING,
path='owen.context.mobileDeviceMake'),
Column('vendor_0',
DataType.STRING,
path='owen.context.vendor[0].vendor'),
Column('vendorClickUuid_0',
DataType.STRING,
path='owen.context.vendor[0].vendorClickUuid'),
Column('udid', DataType.STRING, path='owen.context.udid'),
Column(
'latitude', DataType.STRING, path='owen.context.latitude'),
Column('bluetoothEnabledFlag',
DataType.BOOLEAN,
path='owen.context.bluetoothEnabledFlag'),
Column('environment',
DataType.STRING,
path='owen.context.environment'),
Column('city', DataType.STRING, path='owen.context.city'),
Column(
'userUuid', DataType.STRING, path='owen.context.userUuid'),
Column('dma', DataType.STRING, path='owen.context.dma'),
Column('testUuid',
DataType.STRING,
path='owen.context.test.testUuid'),
Column('userAgent',
DataType.STRING,
path='owen.context.userAgent'),
Column('previousViewInstanceUuid',
DataType.STRING,
path='owen.context.previousViewInstanceUuid'),
Column(
'language', DataType.STRING, path='owen.context.language'),
Column('deviceCategory',
DataType.STRING,
path='owen.context.deviceCategory'),
Column('bluetoothBeaconId',
DataType.STRING,
path='owen.context.bluetoothBeaconId'),
Column('screenWidth',
DataType.STRING,
path='owen.context.screenWidth'),
Column('personalizationFlag',
DataType.BOOLEAN,
path='owen.context.personalizationFlag'),
Column('appForegroundFlag',
DataType.BOOLEAN,
path='owen.context.appForegroundFlag'),
Column('mobileDeviceModel',
DataType.STRING,
path='owen.context.mobileDeviceModel'),
Column('userQualifier',
DataType.STRING,
path='owen.context.userQualifier'),
Column('deviceFingerprint',
DataType.STRING,
path='owen.context.deviceFingerprint'),
Column('ipAddress',
DataType.STRING,
path='owen.context.ipAddress'),
Column(
'osFamily', DataType.STRING, path='owen.context.osFamily'),
Column('advertiserUuid',
DataType.STRING,
path='owen.context.advertiserUuid'),
Column('notificationEnabledFlag',
DataType.BOOLEAN,
path='owen.context.notificationEnabledFlag'),
Column('inventory',
DataType.STRING,
path='owen.context.inventory'),
Column('vendor', DataType.STRING, path='owen.context.vendor'),
Column('experiment',
DataType.STRING,
path='owen.context.experiment'),
],
))))
Dataset(data=(DatasetData(
name='beacon_native_app_parsed_v01',
table_name='beacon_native_app',
location='s3://example-bucket/nb.retailmenot.com/parsed_logs',
load_type=LoadType.INSERT,
data_format=DataFormat(FileFormat.TEXTFILE,
RowFormat.DELIMITED,
delimited_by='\t',
quoted_by='"',
escaped_by='\\',
null_string='NULL',
num_header_rows=1),
compression=Compression.NONE,
partitions=[
Column('year', DataType.STRING),
Column('week', DataType.STRING),
],
columns=[
Column('logFileId', DataType.BIGINT),
Column('lineNumber', DataType.INT),
Column('created',
DataType.TIMESTAMP,
date_pattern="yyyy-MM-dd HH:mm:ss"),
Column('remoteip', DataType.STRING),
Column('useragent', DataType.STRING),
Column('eventType', DataType.STRING),
Column('appVersion', DataType.STRING),
Column('advertiserID', DataType.STRING),
Column('couponsOnPage', DataType.INT),
Column('coupons', DataType.STRING),
Column('channel', DataType.STRING),
Column('geoCouponCount', DataType.STRING),
Column('geofence', DataType.STRING),
Column('geofenceTimeSpent', DataType.STRING),
Column('loginStatus', DataType.STRING),
Column('products', DataType.STRING),
Column('session', DataType.STRING),
Column('systemName', DataType.STRING),
Column('systemVersion', DataType.STRING),
Column('udid', DataType.STRING),
Column('userQualifier', DataType.STRING),
Column('url', DataType.STRING),
Column('user_uuid', DataType.STRING),
Column('userId', DataType.STRING),
Column('searchType', DataType.STRING),
Column('searchListTerm', DataType.STRING),
Column('searchTerm', DataType.STRING),
Column('emailUUId', DataType.STRING),
Column('userFingerprint', DataType.STRING),
Column('locationStatus', DataType.STRING),
Column('pushNotificationStatus', DataType.BOOLEAN),
Column('placement', DataType.STRING),
Column('loc', DataType.STRING),
Column('ppoi0', DataType.STRING),
Column('ppoi1', DataType.STRING),
Column('ppoi2', DataType.STRING),
Column('ppoi3', DataType.STRING),
Column('ppoi4', DataType.STRING),
Column('appLaunchNotificationType', DataType.STRING),
Column('scenarioName', DataType.STRING),
Column('behaviorName', DataType.STRING),
Column('couponType', DataType.STRING),
Column('couponPosition', DataType.STRING),
Column('hasQSRContent', DataType.BOOLEAN),
Column('promptName', DataType.STRING),
Column('locationPermissionChanage', DataType.STRING),
Column('couponProblemType', DataType.STRING),
Column('storeTitle', DataType.STRING),
Column('mallName', DataType.STRING),
Column('restaurantName', DataType.STRING),
Column('milesAway', 'float'),
Column('menuItem', DataType.STRING),
Column('toolName', DataType.STRING),
Column('toolAction', DataType.STRING),
Column('toolStep', DataType.STRING),
Column('mallPosition', DataType.INT),
Column('recommendStoreName', DataType.STRING),
Column('recommendStorePosition', DataType.INT),
Column('favoriteStoreName', DataType.STRING),
Column('favoriteStoreAction', DataType.STRING),
Column('favoriteStorePosition', DataType.INT),
Column('favoriteSiteId', DataType.STRING),
Column('receiverName', DataType.STRING),
Column('outclickButtonPrompt', DataType.STRING),
Column('dataSource', DataType.STRING),
Column('searchResultCount', DataType.INT),
Column('searchResultPosition', DataType.INT),
Column('shareType', DataType.STRING),
Column('daysUntilExpiration', DataType.INT),
Column('fireDate', DataType.BIGINT),
Column('settingsChangeValue', DataType.STRING),
Column('settingsChangeType', DataType.STRING),
Column('settingsChangeLocation', DataType.STRING),
Column('clickAction', DataType.STRING),
Column('tnt', DataType.STRING),
Column('previousPage', DataType.STRING),
Column('clickPage', DataType.STRING),
Column('launchReason', DataType.STRING),
Column('taplyticsData', DataType.STRING),
Column('appCampaign', DataType.STRING),
Column('accountMethod', DataType.STRING),
Column('appState', DataType.STRING),
Column('btStatus', DataType.BOOLEAN),
Column('btBeaconId', DataType.STRING),
Column('btBeaconFactoryId', DataType.STRING),
Column('btBeaconName', DataType.STRING),
Column('btTimeSpent', DataType.STRING),
Column('purchaseId', DataType.STRING),
Column('transactionId', DataType.STRING),
Column('outclickLink', DataType.STRING),
Column('outclickPage', DataType.STRING),
Column('featuredCouponPosition', DataType.INT),
Column('commentCount', DataType.INT),
Column('mallCount', DataType.INT),
Column('clickCount', DataType.INT),
Column('merchantName', DataType.STRING),
Column('merchantPosition', DataType.INT),
],
))))
def test_hive_table_definition_step(self):
ds = Dataset(data=DatasetData(
name='owen_eu_v01',
table_name='owen_eu',
location='s3://s3-rpt-uss-dat-warehouse/prd/inbound/overlord/eu-all-events',
data_format=DataFormat(
file_format=FileFormat.TEXTFILE,
row_format=RowFormat.JSON,
),
columns=[
Column('host', 'STRING', path='metadata.host'),
Column('pageName', 'STRING', path='owen.context.pageName'),
Column('viewInstanceUuid', 'STRING', path='owen.context.viewInstanceUuid'),
Column('previousPageName', 'STRING', path='owen.context.previousPageName'),
Column('previousViewInstanceUuid', 'STRING', path='owen.context.previousViewInstanceUuid'),
Column('session', 'STRING', path='owen.context.session'),
Column('pageType', 'STRING', path='owen.context.pageType'),
Column('propertyName', 'STRING', path='owen.context.propertyName'),
Column('enviroment', 'STRING', path='owen.context.environment'),
Column('appForegroundFlag', 'BOOLEAN', path='owen.context.appForegroundFlag'),
Column('bluetoothEnabledFlag', 'BOOLEAN', path='owen.context.bluetoothEnabledFlag'),
Column('favoriteFlag', 'BOOLEAN', path='owen.context.favoriteFlag'),
Column('locationEnabledFlag', 'BOOLEAN', path='owen.context.locationEnabledFlag'),
Column('loggedInFlag', 'BOOLEAN', path='owen.context.loggedInFlag'),
Column('notificationEnabledFlag', 'BOOLEAN', path='owen.context.notificationEnabledFlag'),
Column('personalizationFlag', 'BOOLEAN', path='owen.context.personalizationFlag'),
Column('advertiserUuid', 'STRING', path='owen.context.advertiserUuid'),
Column('udid', 'STRING', path='owen.context.udid'),
Column('userQualifier', 'STRING', path='owen.context.userQualifier'),
Column('userId', 'STRING', path='owen.context.custom.legacy.userId'),
Column('userUuid', 'STRING', path='owen.context.userUuid'),
Column('macAddress', 'STRING', path='owen.context.macAddress'),
Column('ipAddress', 'STRING', path='owen.context.ipAddress'),
Column('osVersion', 'STRING', path='owen.context.osVersion'),
Column('osFamily', 'STRING', path='owen.context.osFamily'),
Column('osName', 'STRING', path='owen.context.osName'),
Column('browserFamily', 'STRING', path='owen.context.browserFamily'),
Column('deviceCategory', 'STRING', path='owen.context.deviceCategory'),
Column('deviceMake', 'STRING', path='owen.context.mobileDeviceMake'),
Column('deviceModel', 'STRING', path='owen.context.mobileDeviceModel'),
Column('connectionType', 'STRING', path='owen.context.connectionType'),
Column('userAgent', 'STRING', path='owen.context.userAgent'),
Column('geofenceId', 'STRING', path='owen.context.custom.legacy.geofenceId'),
Column('eventTimestamp', 'TIMESTAMP', path='owen.event.eventTimestamp', date_pattern="yyyy-MM-dd'T'HH:mm:ssZ"),
Column('eventInstanceUuid', 'STRING', path='owen.event.eventInstanceUuid'),
Column('eventPlatformVersion', 'STRING', path='owen.event.eventPlatformVersion'),
Column('eventVersion', 'STRING', path='owen.event.eventVersion'),
Column('eventCategory', 'STRING', path='owen.event.eventCategory'),
Column('eventName', 'STRING', path='owen.event.eventName'),
Column('eventAction', 'STRING', path='owen.event.eventAction'),
Column('eventPlatform', 'STRING', path='owen.event.eventPlatform'),
Column('testUnixTimestampSecondsPattern', 'TIMESTAMP', path='some.fake.path.testUnixTimestampSecondsPattern', date_pattern='UNIX_TIMESTAMP_SECONDS'),
Column('testUnixTimestampMillisPattern', 'TIMESTAMP', path='some.fake.path.testUnixTimestampMillisPattern', date_pattern='UNIX_TIMESTAMP_MILLIS'),
],
compression='GZIP',
partitions=[
Column('year', 'STRING'),
Column('week', 'STRING'),
Column('day', 'STRING'),
],
))
call('mkdir -p /tmp/dart-emr-test/hive/')
this_path = os.path.dirname(os.path.abspath(__file__))
shutil.copyfile(this_path + '/../../../engine/emr/steps/hive/copy_to_table.hql', '/tmp/dart-emr-test/hive/copy_to_table.hql')
action_id = 'actionid123'
target_dataset = Dataset.from_dict(ds.to_dict())
target_dataset.data.data_format.num_header_rows = 0
target_dataset.data.data_format = DataFormat(FileFormat.RCFILE, RowFormat.NONE)
stage_dataset = Dataset.from_dict(ds.to_dict())
assert isinstance(stage_dataset, Dataset)
for c in stage_dataset.data.columns:
c.data_type = DataType.STRING
hive_copy_to_table(stage_dataset, 'owen_eu_stage', target_dataset, 'owen_eu', 's3://test', '/tmp/dart-emr-test/', action_id, None, 1, 1)
with open(os.path.join(this_path, 'copy_to_table_owen_eu.hql')) as f:
expected_contents = f.read()
with open('/tmp/dart-emr-test/hive/copy_to_table_owen_eu.hql') as f:
actual_contents = f.read()
self.assertEqual(expected_contents, actual_contents)
dataset = dart.save_dataset(Dataset(data=DatasetData(
name='weblogs_v01',
table_name='weblogs',
location='s3://example-bucket/weblogs/www.retailmenot.com/ec2/',
data_format=DataFormat(
file_format=FileFormat.TEXTFILE,
row_format=RowFormat.REGEX,
regex_input="(?<ip>^(?:(?:unknown(?:,\\s)?|(?:\\d+\\.\\d+\\.\\d+\\.\\d+(?:,\\s)?))+)|\\S*)\\s+\\S+\\s+(?<userIdentifier>(?:[^\\[]+|\\$\\S+\\['\\S+'\\]|\\[username\\]))\\s*\\s+\\[(?<requestDate>[^\\]]+)\\]\\s+\"(?<httpMethod>(?:GET|HEAD|POST|PUT|DELETE|TRACE))\\s(?<urlPath>(?:[^ ?]+))(?:\\?(?<queryString>(?:[^ ]+)))?\\sHTTP/(?<httpVersion>(?:[\\d\\.]+))\"\\s+(?<statusCode>[0-9]+)\\s+(?<bytesSent>\\S+)\\s+\"(?<referrer>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<userAgent>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+(?<responseTime>[-0-9]*)\\s+\"(?<hostName>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<userFingerprint>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<userId>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<sessionId>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<requestId>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<visitorId>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<vegSlice>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<fruitSlice>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<cacheHitMiss>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s*\\Z",
regex_output="%1$s %2$s %3$s %4$s %5$s %6$s %7$s %8$s %9$s %10$s %11$s %12$s %13$s %14$s %15$s %16$s %17$s %18$s %19$s %20$s %21s",
),
columns=[
Column('ip', DataType.STRING),
Column('user', DataType.STRING),
Column('requestDate', DataType.TIMESTAMP, date_pattern='dd/MMM/yyyy:HH:mm:ss Z'),
Column('httpMethod', DataType.STRING),
Column('urlPath', DataType.STRING),
Column('queryString', DataType.STRING),
Column('httpVersion', DataType.STRING),
Column('statusCode', DataType.STRING),
Column('bytesSent', DataType.INT),
Column('referrer', DataType.STRING),
Column('userAgent', DataType.STRING),
Column('responseTime', DataType.BIGINT),
Column('hostname', DataType.STRING),
Column('userFingerprint', DataType.STRING),
Column('userId', DataType.STRING),
Column('sessionId', DataType.STRING),
Column('requestId', DataType.STRING),
Column('visitorId', DataType.STRING),
Column('vegSlice', DataType.STRING),
Column('fruitSlice', DataType.STRING),
Column('cacheHitMiss', DataType.STRING),
],
compression=Compression.BZ2,
partitions=[
Column('year', DataType.STRING),
Column('week', DataType.STRING),
],
)))
def test_hive_table_definition_step(self):
ds = Dataset(data=DatasetData(
name='weblogs_v01',
table_name='weblogs',
location='s3://wsm-log-servers/weblogs/www.retailmenot.com/ec2/',
data_format=DataFormat(
file_format=FileFormat.TEXTFILE,
row_format=RowFormat.REGEX,
regex_input=
"(?<ip>^(?:(?:unknown(?:,\\s)?|(?:\\d+\\.\\d+\\.\\d+\\.\\d+(?:,\\s)?))+)|\\S*)\\s+\\S+\\s+(?<userIdentifier>(?:[^\\[]+|\\$\\S+\\['\\S+'\\]|\\[username\\]))\\s*\\s+\\[(?<requestDate>[^\\]]+)\\]\\s+\"(?<httpMethod>(?:GET|HEAD|POST|PUT|DELETE|TRACE))\\s(?<urlPath>(?:[^ ?]+))(?:\\?(?<queryString>(?:[^ ]+)))?\\sHTTP/(?<httpVersion>(?:[\\d\\.]+))\"\\s+(?<statusCode>[0-9]+)\\s+(?<bytesSent>\\S+)\\s+\"(?<referrer>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<userAgent>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+(?<responseTime>[-0-9]*)\\s+\"(?<hostName>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<userFingerprint>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<userId>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<sessionId>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<requestId>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<visitorId>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<vegSlice>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<fruitSlice>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s+\"(?<cacheHitMiss>(?:[^\"\\\\]*(?:\\\\.[^\"\\\\]*)*))\"\\s*\\Z",
regex_output=
"%1$s %2$s %3$s %4$s %5$s %6$s %7$s %8$s %9$s %10$s %11$s %12$s %13$s %14$s %15$s %16$s %17$s %18$s %19$s %20$s %21s",
),
columns=[
Column('ip', 'STRING'),
Column('user', 'STRING'),
Column('requestDate',
'TIMESTAMP',
date_pattern='dd/MMM/yyyy:HH:mm:ss Z'),
Column('httpMethod', 'STRING'),
Column('urlPath', 'STRING'),
Column('queryString', 'STRING'),
Column('httpVersion', 'STRING'),
Column('statusCode', 'STRING'),
Column('bytesSent', 'INT'),
Column('referrer', 'STRING'),
Column('userAgent', 'STRING'),
Column('responseTime', 'BIGINT'),
Column('hostname', 'STRING'),
Column('userFingerprint', 'STRING'),
Column('userId', 'STRING'),
Column('sessionId', 'STRING'),
Column('requestId', 'STRING'),
Column('visitorId', 'STRING'),
Column('vegSlice', 'STRING'),
Column('fruitSlice', 'STRING'),
Column('cacheHitMiss', 'STRING'),
],
compression='GZIP',
partitions=[
Column('year', 'STRING'),
Column('week', 'STRING'),
Column('day', 'STRING'),
],
))
call('mkdir -p /tmp/dart-emr-test/hive/')
this_path = os.path.dirname(os.path.abspath(__file__))
shutil.copyfile(
this_path + '/../../../engine/emr/steps/hive/copy_to_table.hql',
'/tmp/dart-emr-test/hive/copy_to_table.hql')
hive_copy_to_table(ds, 'weblogs_stage', ds, 'weblogs', 's3://test',
'/tmp/dart-emr-test/', 'actionid123', None, 1, 1)
with open(os.path.join(this_path, 'copy_to_table_weblogs.hql')) as f:
expected_contents = f.read()
with open('/tmp/dart-emr-test/hive/copy_to_table_weblogs.hql') as f:
actual_contents = f.read()
self.assertEqual(expected_contents, actual_contents)
Dataset(data=(DatasetData(
name='owen_eu_DW-3213_v3',
table_name='owen_eu',
location='s3://example-bucket/prd/inbound/overlord/eu-all-events',
load_type=LoadType.MERGE,
data_format=DataFormat(
file_format=FileFormat.TEXTFILE,
row_format=RowFormat.JSON,
),
compression=Compression.GZIP,
partitions=[
Column('year', DataType.STRING),
Column('month', DataType.STRING),
Column('day', DataType.STRING),
],
columns=[
Column('host', DataType.STRING, path='metadata.host'),
Column('referer', DataType.STRING, path='metadata.referer'),
Column('userAgent',
DataType.STRING,
path='owen.context.userAgent'),
Column('ipAddress',
DataType.STRING,
path='owen.context.ipAddress'),
Column('session', DataType.STRING,
path='owen.context.session'),
Column('propertyName',
DataType.STRING,
path='owen.context.propertyName'),
Column(
'pageName', DataType.STRING, path='owen.context.pageName'),
Column('previousPageName',
DataType.STRING,
path='owen.context.previousPageName'),
Column('viewInstanceUuid',
DataType.STRING,
path='owen.context.viewInstanceUuid'),
Column('previousViewInstanceUuid',
DataType.STRING,
path='owen.context.previousViewInstanceUuid'),
Column(
'pageType', DataType.STRING, path='owen.context.pageType'),
Column('udid', DataType.STRING, path='owen.context.udid'),
Column('advertiserUuid',
DataType.STRING,
path='owen.context.advertiserUuid'),
Column(
'osFamily', DataType.STRING, path='owen.context.osFamily'),
Column(
'latitude', DataType.STRING, path='owen.context.latitude'),
Column('longitude',
DataType.STRING,
path='owen.context.longitude'),
Column('userId',
DataType.STRING,
path='owen.context.custom.legacy.userId'),
Column('geofenceId',
DataType.STRING,
path='owen.context.custom.legacy.geofenceId'),
Column(
'userUuid', DataType.STRING, path='owen.context.userUuid'),
Column('offerId',
DataType.STRING,
path='owen.context.inventory[0].inventoryUuid'),
Column('inventorySource',
DataType.STRING,
path='owen.context.inventory[0].inventorySource'),
Column('expirationDate',
DataType.STRING,
path='owen.context.inventory[0].expirationDate'),
Column('position',
DataType.STRING,
path='owen.context.inventory[0].position'),
Column('offerType',
DataType.STRING,
path='owen.context.inventory[0].inventoryType'),
Column('eventInstanceUuid',
DataType.STRING,
path='owen.event.eventInstanceUuid'),
Column('eventTimestamp',
DataType.TIMESTAMP,
path='owen.event.eventTimestamp',
date_pattern="yyyy-MM-dd'T'HH:mm:ss'Z'"),
Column('eventPlatform',
DataType.STRING,
path='owen.event.eventPlatform'),
Column('eventCategory',
DataType.STRING,
path='owen.event.eventCategory'),
Column('eventAction',
DataType.STRING,
path='owen.event.eventAction'),
Column(
'eventName', DataType.STRING, path='owen.event.eventName'),
Column('eventTarget',
DataType.STRING,
path='owen.event.eventTarget'),
Column('eventVersion',
DataType.STRING,
path='owen.event.eventVersion'),
Column('userQualifier',
DataType.STRING,
path='owen.context.userQualifier'),
Column('outclickUuid',
DataType.STRING,
path='owen.context.inventory[0].outclickUuid'),
Column('inventoryName',
DataType.STRING,
path='owen.context.inventory[0].inventoryName'),
Column('enviroment',
DataType.STRING,
path='owen.context.environment'),
Column('loggedInFlag',
DataType.STRING,
path='owen.context.loggedInFlag'),
Column('eventPlatformVersion',
DataType.STRING,
path='owen.event.eventPlatformVersion'),
Column('appForegroundFlag',
DataType.BOOLEAN,
path='owen.context.appForegroundFlag'),
Column('bluetoothEnabledFlag',
DataType.BOOLEAN,
path='owen.context.bluetoothEnabledFlag'),
Column('favoriteFlag',
DataType.BOOLEAN,
path='owen.context.favoriteFlag'),
Column('locationEnabledFlag',
DataType.BOOLEAN,
path='owen.context.locationEnabledFlag'),
Column('notificationEnabledFlag',
DataType.BOOLEAN,
path='owen.context.notificationEnabledFlag'),
Column('personalizationFlag',
DataType.BOOLEAN,
path='owen.context.personalizationFlag'),
Column('macAddress',
DataType.STRING,
path='owen.context.macAddress'),
Column('osVersion',
DataType.STRING,
path='owen.context.osVersion'),
Column('osName', DataType.STRING, path='owen.context.osName'),
Column('browserFamily',
DataType.STRING,
path='owen.context.browserFamily'),
Column('deviceCategory',
DataType.STRING,
path='owen.context.deviceCategory'),
Column('deviceMake',
DataType.STRING,
path='owen.context.mobileDeviceMake'),
Column('deviceModel',
DataType.STRING,
path='owen.context.mobileDeviceModel'),
Column('connectionType',
DataType.STRING,
path='owen.context.connectionType'),
Column('browserVersion',
DataType.STRING,
path='owen.context.browserVersion'),
Column('city', DataType.STRING, path='owen.context.city'),
Column('country', DataType.STRING,
path='owen.context.country'),
Column('region', DataType.STRING, path='owen.context.region'),
Column('partialSearchTerm',
DataType.STRING,
path='owen.context.partialSearchTerm'),
Column('outclickURL',
DataType.STRING,
path='owen.context.inventory[0].outRedirectUrl'),
Column('clickLocation',
DataType.STRING,
path='owen.context.inventory[0].clickLocation'),
Column('inventoryChannel',
DataType.STRING,
path='owen.context.inventory[0].inventoryChannel'),
Column('brand',
DataType.STRING,
path='owen.context.inventory[0].brand'),
Column('commentsCount',
DataType.INT,
path='owen.context.inventory[0].commentsCount'),
Column('legacyOfferId',
DataType.STRING,
path='owen.context.custom.legacy.offerIds.offerId'),
Column('pageViewHash',
DataType.STRING,
path='owen.context.custom.legacy.pageViewHash'),
Column('vIdInt',
DataType.STRING,
path='owen.context.custom.legacy.vIdInt'),
Column('merchantId',
DataType.STRING,
path='owen.context.custom.legacy.merchantId'),
Column('facebookConnect',
DataType.STRING,
path='owen.context.custom.facebookConnect'),
Column('schemaKey', DataType.STRING, path='schema.key'),
],
))))
dataset = dart.save_dataset(Dataset(data=(DatasetData(
name='beacon_native_app_v02',
table_name='beacon_native_app',
location='s3://example-bucket/prd/beacon/native_app/v2/parquet/snappy',
hive_compatible_partition_folders=True,
load_type=LoadType.INSERT,
data_format=DataFormat('parquet'),
columns=[
Column('logFileId', DataType.BIGINT),
Column('lineNumber', DataType.INT),
Column('created', DataType.BIGINT),
Column('remoteip', DataType.STRING),
Column('useragent', DataType.STRING),
Column('eventType', DataType.STRING),
Column('appVersion', DataType.STRING),
Column('advertiserID', DataType.STRING),
Column('couponsOnPage', DataType.INT),
Column('coupons', DataType.STRING),
Column('channel', DataType.STRING),
Column('geoCouponCount', DataType.STRING),
Column('geofence', DataType.STRING),
Column('geofenceTimeSpent', DataType.STRING),
Column('loginStatus', DataType.STRING),
Column('products', DataType.STRING),
Column('session', DataType.STRING),
Column('systemName', DataType.STRING),
Column('systemVersion', DataType.STRING),
Column('udid', DataType.STRING),
Column('userQualifier', DataType.STRING),
Column('url', DataType.STRING),
Column('user_uuid', DataType.STRING),
Column('userId', DataType.STRING),
Column('searchType', DataType.STRING),
Column('searchListTerm', DataType.STRING),
Column('searchTerm', DataType.STRING),
Column('emailUUId', DataType.STRING),
Column('userFingerprint', DataType.STRING),
Column('locationStatus', DataType.STRING),
Column('pushNotificationStatus', DataType.BOOLEAN),
Column('placement', DataType.STRING),
Column('loc', DataType.STRING),
Column('ppoi0', DataType.STRING),
Column('ppoi1', DataType.STRING),
Column('ppoi2', DataType.STRING),
Column('ppoi3', DataType.STRING),
Column('ppoi4', DataType.STRING),
Column('appLaunchNotificationType', DataType.STRING),
Column('scenarioName', DataType.STRING),
Column('behaviorName', DataType.STRING),
Column('couponType', DataType.STRING),
Column('couponPosition', DataType.STRING),
Column('hasQSRContent', DataType.BOOLEAN),
Column('promptName', DataType.STRING),
Column('locationPermissionChanage', DataType.STRING),
Column('couponProblemType', DataType.STRING),
Column('storeTitle', DataType.STRING),
Column('mallName', DataType.STRING),
Column('restaurantName', DataType.STRING),
Column('milesAway', 'float'),
Column('menuItem', DataType.STRING),
Column('toolName', DataType.STRING),
Column('toolAction', DataType.STRING),
Column('toolStep', DataType.STRING),
Column('mallPosition', DataType.INT),
Column('recommendStoreName', DataType.STRING),
Column('recommendStorePosition', DataType.INT),
Column('favoriteStoreName', DataType.STRING),
Column('favoriteStoreAction', DataType.STRING),
Column('favoriteStorePosition', DataType.INT),
Column('favoriteSiteId', DataType.STRING),
Column('receiverName', DataType.STRING),
Column('outclickButtonPrompt', DataType.STRING),
Column('dataSource', DataType.STRING),
Column('searchResultCount', DataType.INT),
Column('searchResultPosition', DataType.INT),
Column('shareType', DataType.STRING),
Column('daysUntilExpiration', DataType.INT),
Column('fireDate', DataType.BIGINT),
Column('settingsChangeValue', DataType.STRING),
Column('settingsChangeType', DataType.STRING),
Column('settingsChangeLocation', DataType.STRING),
Column('clickAction', DataType.STRING),
Column('tnt', DataType.STRING),
Column('previousPage', DataType.STRING),
Column('clickPage', DataType.STRING),
Column('launchReason', DataType.STRING),
Column('taplyticsData', DataType.STRING),
Column('appCampaign', DataType.STRING),
Column('accountMethod', DataType.STRING),
Column('appState', DataType.STRING),
Column('btStatus', DataType.BOOLEAN),
Column('btBeaconId', DataType.STRING),
Column('btBeaconFactoryId', DataType.STRING),
Column('btBeaconName', DataType.STRING),
Column('btTimeSpent', DataType.STRING),
Column('purchaseId', DataType.STRING),
Column('transactionId', DataType.STRING),
Column('outclickLink', DataType.STRING),
Column('outclickPage', DataType.STRING),
Column('featuredCouponPosition', DataType.INT),
Column('commentCount', DataType.INT),
Column('mallCount', DataType.INT),
Column('clickCount', DataType.INT),
Column('merchantName', DataType.STRING),
Column('merchantPosition', DataType.INT),
],
compression=Compression.SNAPPY,
partitions=[Column('createdpartition', DataType.STRING)],
))))
def setUp(self):
dart = Dart(host='localhost', port=5000)
""" :type dart: dart.client.python.dart_client.Dart """
self.dart = dart
cs = [Column('c1', DataType.VARCHAR, 50), Column('c2', DataType.BIGINT)]
df = DataFormat(FileFormat.PARQUET, RowFormat.NONE)
dataset_data = DatasetData('test-dataset0', 'test_dataset_table0', 's3://test/dataset/0/%s' + random_id(), df, cs)
self.dataset0 = self.dart.save_dataset(Dataset(data=dataset_data))
cs = [Column('c1', DataType.VARCHAR, 50), Column('c2', DataType.BIGINT)]
df = DataFormat(FileFormat.PARQUET, RowFormat.NONE)
dataset1_location = 's3://test/dataset/1/%s' + random_id()
dataset_data = DatasetData('test-dataset1', 'test_dataset_table1', dataset1_location, df, cs)
self.dataset1 = self.dart.save_dataset(Dataset(data=dataset_data))
cs = [Column('c1', DataType.VARCHAR, 50), Column('c2', DataType.BIGINT)]
df = DataFormat(FileFormat.PARQUET, RowFormat.NONE)
dataset_data = DatasetData('test-dataset2-no-show', 'test_dataset_table2', 's3://test/dataset/2/%s' + random_id(), df, cs)
self.dataset2 = self.dart.save_dataset(Dataset(data=dataset_data))
s = Subscription(data=SubscriptionData('test-subscription0', self.dataset0.id))
self.subscription0 = self.dart.save_subscription(s)
s = Subscription(data=SubscriptionData('test-subscription2-no-show', self.dataset2.id))
self.subscription2 = self.dart.save_subscription(s)
dst_args = {'action_sleep_time_in_seconds': 0}
dst = Datastore(data=DatastoreData('test-datastore0', 'no_op_engine', args=dst_args, state=DatastoreState.TEMPLATE))
self.datastore0 = self.dart.save_datastore(dst)
dst = Datastore(data=DatastoreData('test-datastore1', 'no_op_engine', args=dst_args, state=DatastoreState.TEMPLATE))
self.datastore1 = self.dart.save_datastore(dst)
dst = Datastore(data=DatastoreData('test-datastore2-no-show', 'no_op_engine', args=dst_args, state=DatastoreState.ACTIVE))
self.datastore2 = self.dart.save_datastore(dst)
wf0 = Workflow(data=WorkflowData('test-workflow0', self.datastore0.id, state=WorkflowState.ACTIVE))
self.workflow0 = self.dart.save_workflow(wf0, self.datastore0.id)
wf1 = Workflow(data=WorkflowData('test-workflow1', self.datastore1.id, state=WorkflowState.ACTIVE))
self.workflow1 = self.dart.save_workflow(wf1, self.datastore1.id)
wf2 = Workflow(data=WorkflowData('test-workflow2-no-show', self.datastore2.id, state=WorkflowState.ACTIVE))
self.workflow2 = self.dart.save_workflow(wf2, self.datastore2.id)
a_args = {'source_hdfs_path': 'hdfs:///user/hive/warehouse/test', 'destination_s3_path': dataset1_location}
a00 = Action(data=ActionData(NoOpActionTypes.action_that_succeeds.name, NoOpActionTypes.action_that_succeeds.name, state=ActionState.TEMPLATE))
a01 = Action(data=ActionData(NoOpActionTypes.consume_subscription.name, NoOpActionTypes.consume_subscription.name, {'subscription_id': self.subscription0.id}, state=ActionState.TEMPLATE))
a02 = Action(data=ActionData(NoOpActionTypes.action_that_succeeds.name, NoOpActionTypes.action_that_succeeds.name, state=ActionState.TEMPLATE))
a03 = Action(data=ActionData(NoOpActionTypes.copy_hdfs_to_s3_action.name, NoOpActionTypes.copy_hdfs_to_s3_action.name, a_args, state=ActionState.TEMPLATE))
a04 = Action(data=ActionData(NoOpActionTypes.action_that_succeeds.name, NoOpActionTypes.action_that_succeeds.name, state=ActionState.TEMPLATE))
self.action00, self.action01, self.action02, self.action03, self.action04 = \
self.dart.save_actions([a00, a01, a02, a03, a04], workflow_id=self.workflow0.id)
a10 = Action(data=ActionData(NoOpActionTypes.load_dataset.name, NoOpActionTypes.load_dataset.name, {'dataset_id': self.dataset1.id}, state=ActionState.TEMPLATE))
self.action10 = self.dart.save_actions([a10], workflow_id=self.workflow1.id)
a20 = Action(data=ActionData(NoOpActionTypes.action_that_succeeds.name, NoOpActionTypes.action_that_succeeds.name, state=ActionState.HAS_NEVER_RUN))
a21 = Action(data=ActionData(NoOpActionTypes.load_dataset.name, NoOpActionTypes.load_dataset.name, {'dataset_id': self.dataset2.id}, state=ActionState.TEMPLATE))
self.action20 = self.dart.save_actions([a20], datastore_id=self.datastore2.id)
self.action21 = self.dart.save_actions([a21], workflow_id=self.workflow2.id)
self.event1 = self.dart.save_event(Event(data=EventData('test-event1', state=EventState.ACTIVE)))
self.event2 = self.dart.save_event(Event(data=EventData('test-event2-no-show', state=EventState.ACTIVE)))
tr_args = {'event_id': self.event1.id}
tr = Trigger(data=TriggerData('test-event-trigger1', 'event', [self.workflow1.id], tr_args, TriggerState.ACTIVE))
self.event_trigger1 = self.dart.save_trigger(tr)
tr_args = {'event_id': self.event2.id}
tr = Trigger(data=TriggerData('test-event-trigger2-no-show', 'event', [self.workflow2.id], tr_args, TriggerState.ACTIVE))
self.event_trigger2 = self.dart.save_trigger(tr)
st_args = {'fire_after': 'ALL', 'completed_trigger_ids': [self.event_trigger1.id]}
st = Trigger(data=TriggerData('test-super-trigger1', 'super', None, st_args, TriggerState.ACTIVE))
self.super_trigger1 = self.dart.save_trigger(st)
st_args = {'fire_after': 'ANY', 'completed_trigger_ids': [self.super_trigger1.id]}
st = Trigger(data=TriggerData('test-super-trigger2', 'super', [self.workflow1.id], st_args, TriggerState.ACTIVE))
self.super_trigger2 = self.dart.save_trigger(st)
Dataset(data=(DatasetData(
name='owen_outclick_us_v02',
description=
'Owen outclick data, based on overlord schema version. Considered a replacement for outclick events.',
table_name='outclick',
location=
's3://example-bucket/prd/inbound/overlord/raw-firehose-02/rmn-outclicks',
load_type=LoadType.MERGE,
data_format=DataFormat(
file_format=FileFormat.TEXTFILE,
row_format=RowFormat.JSON,
),
compression=Compression.GZIP,
partitions=[
Column('year', DataType.STRING),
Column('month', DataType.STRING),
Column('day', DataType.STRING),
],
primary_keys=['eventInstanceUuid'],
merge_keys=['eventInstanceUuid'],
sort_keys=[
'eventTimestamp', 'eventInstanceUuid', 'derivedEventInstanceId'
],
distribution_keys=['eventInstanceUuid'],
batch_merge_sort_keys=['owenProcessed DESC'],
columns=[
Column('advertiserUuid',
DataType.VARCHAR,
length=2048,
path='owen.context.advertiserUuid'),
Column('appBadgeCount',
DataType.INT,
path='owen.context.appBadgeCount'),
Column('appForegroundFlag',
DataType.BOOLEAN,
path='owen.context.appForegroundFlag'),
Column('bluetoothBeaconId',
DataType.VARCHAR,
length=50,
path='owen.context.bluetoothBeaconId'),
Column('bluetoothBeaconType',
DataType.VARCHAR,
length=25,
path='owen.context.bluetoothBeaconType'),
Column('bluetoothEnabledFlag',
DataType.BOOLEAN,
path='owen.context.bluetoothEnabledFlag'),
Column('breadCrumb',
DataType.VARCHAR,
length=2048,
path='owen.context.breadCrumb'),
Column('browserFamily',
DataType.VARCHAR,
length=50,
path='owen.context.browserFamily'),
Column('browserVersion',
DataType.VARCHAR,
length=50,
path='owen.context.browserVersion'),
Column('carrier',
DataType.VARCHAR,
length=25,
path='owen.context.carrier'),
Column('city',
DataType.VARCHAR,
length=75,
path='owen.context.city'),
Column('connectionType',
DataType.VARCHAR,
length=25,
path='owen.context.connectionType'),
Column('country',
DataType.VARCHAR,
length=2,
path='owen.context.country'),
Column('custom', DataType.VARCHAR, path='owen.context.custom'),
Column('deviceCategory',
DataType.VARCHAR,
length=2048,
path='owen.context.deviceCategory'),
Column('deviceFingerprint',
DataType.VARCHAR,
length=26,
path='owen.context.deviceFingerprint'),
Column('dma', DataType.INT, path='owen.context.dma'),
Column('environment',
DataType.VARCHAR,
length=2048,
path='owen.context.environment'),
Column('experimentObject',
DataType.VARCHAR,
length=1024,
path='owen.context.experiment'),
Column('failureFlag',
DataType.BOOLEAN,
path='owen.context.failureFlag'),
Column('failureReason',
DataType.VARCHAR,
length=2048,
path='owen.context.failureReason'),
Column('favoriteFlag',
DataType.BOOLEAN,
path='owen.context.favoriteFlag'),
Column('featureFlags',
DataType.VARCHAR,
path='owen.context.featureFlags'),
Column('geofenceUuid',
DataType.VARCHAR,
length=2048,
path='owen.context.geofenceUuid'),
Column('inventoryCount',
DataType.INT,
path='owen.context.inventoryCount'),
Column('inventory_affiliateNetwork',
DataType.VARCHAR,
length=50,
path='owen.context.inventory[0].affiliateNetwork'),
Column('inventory_brand',
DataType.VARCHAR,
length=100,
path='owen.context.inventory[0].brand'),
Column('inventory_claimUuid',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].claimUuid'),
Column('inventory_clickLocation',
DataType.VARCHAR,
length=100,
path='owen.context.inventory[0].clickLocation'),
Column('inventory_commentsCount',
DataType.INT,
path='owen.context.inventory[0].commentsCount'),
Column('inventory_conquestingFlag',
DataType.BOOLEAN,
path='owen.context.inventory[0].conquestingFlag'),
Column('inventory_couponRank',
DataType.NUMERIC,
precision=18,
scale=4,
path='owen.context.inventory[0].couponRank'),
Column('inventory_deepLinkUrl',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].deepLinkUrl'),
Column('inventory_deepLinkUrlScheme',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].deepLinkUrlScheme'),
Column('inventory_exclusivityFlag',
DataType.BOOLEAN,
path='owen.context.inventory[0].exclusivityFlag'),
Column('inventory_expirationDate',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].expirationDate'),
Column('inventory_finalPrice',
DataType.NUMERIC,
precision=18,
scale=4,
path='owen.context.inventory[0].finalPrice'),
Column('inventory_instoreType',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].instoreType'),
Column('inventory_inventoryChannel',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].inventoryChannel'),
Column('inventory_inventoryName',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].inventoryName'),
Column('inventory_inventorySource',
DataType.VARCHAR,
length=50,
path='owen.context.inventory[0].inventorySource'),
Column('inventory_inventoryType',
DataType.VARCHAR,
length=25,
path='owen.context.inventory[0].inventoryType'),
Column('inventory_inventoryUuid',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].inventoryUuid'),
Column('inventory_lastVerifiedDate',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].lastVerifiedDate'),
Column('inventory_monetizableFlag',
DataType.BOOLEAN,
path='owen.context.inventory[0].monetizableFlag'),
Column('inventory_noVotes',
DataType.INT,
path='owen.context.inventory[0].noVotes'),
Column('inventory_onlineType',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].onlineType'),
Column('inventory_originalPrice',
DataType.NUMERIC,
precision=18,
scale=4,
path='owen.context.inventory[0].originalPrice'),
Column('inventory_outRedirectUrl',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].outRedirectUrl'),
Column('inventory_outclickUuid',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].outclickUuid'),
Column('inventory_parentInventoryUuid',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].parentInventoryUuid'),
Column('inventory_personalizationFlag',
DataType.BOOLEAN,
path='owen.context.inventory[0].personalizationFlag'),
Column('inventory_position',
DataType.INT,
path='owen.context.inventory[0].position'),
Column('inventory_proximity',
DataType.NUMERIC,
precision=18,
scale=4,
path='owen.context.inventory[0].proximity'),
Column('inventory_proximityUnit',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].proximityUnit'),
Column('inventory_recommendedFlag',
DataType.BOOLEAN,
path='owen.context.inventory[0].recommendedFlag'),
Column('inventory_redemptionChannel',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].redemptionChannel'),
Column('inventory_retailCategory',
DataType.VARCHAR,
length=75,
path='owen.context.inventory[0].retailCategory'),
Column('inventory_savedFlag',
DataType.BOOLEAN,
path='owen.context.inventory[0].savedFlag'),
Column('inventory_siteUuid',
DataType.VARCHAR,
length=26,
path='owen.context.inventory[0].siteUuid'),
Column('inventory_startDate',
DataType.VARCHAR,
length=2048,
path='owen.context.inventory[0].startDate'),
Column('inventory_successPercentage',
DataType.NUMERIC,
precision=18,
scale=4,
path='owen.context.inventory[0].successPercentage'),
Column('inventory_usedByCount',
DataType.INT,
path='owen.context.inventory[0].usedByCount'),
Column('inventory_yesVotes',
DataType.INT,
path='owen.context.inventory[0].yesVotes'),
Column('ipAddress',
DataType.VARCHAR,
length=45,
path='owen.context.ipAddress'),
Column('language',
DataType.VARCHAR,
length=6,
path='owen.context.language'),
Column('latitude',
DataType.NUMERIC,
precision=18,
scale=4,
path='owen.context.latitude'),
Column('locationEnabledFlag',
DataType.BOOLEAN,
path='owen.context.locationEnabledFlag'),
Column('loggedInFlag',
DataType.BOOLEAN,
path='owen.context.loggedInFlag'),
Column('longitude',
DataType.NUMERIC,
precision=18,
scale=4,
path='owen.context.longitude'),
Column('macAddress',
DataType.VARCHAR,
length=2048,
path='owen.context.macAddress'),
Column('marketing_adGroup',
DataType.VARCHAR,
length=2048,
path='owen.context.marketing.adGroup'),
Column('marketing_campaign',
DataType.VARCHAR,
length=50,
path='owen.context.marketing.campaign'),
Column('marketing_campaignSendCount',
DataType.INT,
path='owen.context.marketing.campaignSendCount'),
Column('marketing_campaignUuid',
DataType.VARCHAR,
length=2048,
path='owen.context.marketing.campaignUuid'),
Column('marketing_cdRank',
DataType.INT,
path='owen.context.marketing.cdRank'),
Column('marketing_channel',
DataType.VARCHAR,
length=50,
path='owen.context.marketing.channel'),
Column('marketing_content',
DataType.VARCHAR,
length=2048,
path='owen.context.marketing.content'),
Column('marketing_medium',
DataType.VARCHAR,
length=50,
path='owen.context.marketing.medium'),
Column('marketing_notificationUuid',
DataType.VARCHAR,
length=2048,
path='owen.context.marketing.notificationUuid'),
Column('marketing_source',
DataType.VARCHAR,
length=100,
path='owen.context.marketing.source'),
Column('marketing_term',
DataType.VARCHAR,
length=2048,
path='owen.context.marketing.term'),
Column('marketing_vendor',
DataType.VARCHAR,
length=25,
path='owen.context.marketing.vendor'),
Column('mobileDeviceMake',
DataType.VARCHAR,
length=25,
path='owen.context.mobileDeviceMake'),
Column('mobileDeviceModel',
DataType.VARCHAR,
length=50,
path='owen.context.mobileDeviceModel'),
Column('notificationEnabledFlag',
DataType.BOOLEAN,
path='owen.context.notificationEnabledFlag'),
Column('osFamily',
DataType.VARCHAR,
length=25,
path='owen.context.osFamily'),
Column('osName',
DataType.VARCHAR,
length=2048,
path='owen.context.osName'),
Column('osVersion',
DataType.VARCHAR,
length=2048,
path='owen.context.osVersion'),
Column('pageName',
DataType.VARCHAR,
length=2048,
path='owen.context.pageName'),
Column('pageType',
DataType.VARCHAR,
length=100,
path='owen.context.pageType'),
Column('partialSearchTerm',
DataType.VARCHAR,
length=2048,
path='owen.context.partialSearchTerm'),
Column('personalizationFlag',
DataType.BOOLEAN,
path='owen.context.personalizationFlag'),
Column('previousPageName',
DataType.VARCHAR,
length=2048,
path='owen.context.previousPageName'),
Column('previousViewInstanceUuid',
DataType.VARCHAR,
length=2048,
path='owen.context.previousViewInstanceUuid'),
Column('promptName',
DataType.VARCHAR,
length=2048,
path='owen.context.promptName'),
Column('propertyName',
DataType.VARCHAR,
length=20,
path='owen.context.propertyName'),
Column('referrer',
DataType.VARCHAR,
length=2048,
path='owen.context.referrer'),
Column('region',
DataType.VARCHAR,
length=25,
path='owen.context.region'),
Column('screenHeight',
DataType.INT,
path='owen.context.screenHeight'),
Column('screenWidth',
DataType.INT,
path='owen.context.screenWidth'),
Column('session',
DataType.VARCHAR,
length=2048,
path='owen.context.session'),
Column('test_testUuid',
DataType.VARCHAR,
length=26,
path='owen.context.test.testUuid'),
Column('udid',
DataType.VARCHAR,
length=40,
path='owen.context.udid'),
Column('userAgent',
DataType.VARCHAR,
length=2048,
path='owen.context.userAgent'),
Column('userQualifier',
DataType.VARCHAR,
length=26,
path='owen.context.userQualifier'),
Column('userUuid',
DataType.VARCHAR,
length=2048,
path='owen.context.userUuid'),
Column('vendorObject',
DataType.VARCHAR,
length=512,
path='owen.context.vendor'),
Column('viewInstanceUuid',
DataType.VARCHAR,
length=128,
path='owen.context.viewInstanceUuid'),
Column('eventAction',
DataType.VARCHAR,
length=2048,
path='owen.event.eventAction'),
Column('eventCategory',
DataType.VARCHAR,
length=25,
path='owen.event.eventCategory'),
Column('eventInstanceUuid',
DataType.VARCHAR,
length=26,
path='owen.event.eventInstanceUuid'),
Column('eventName',
DataType.VARCHAR,
length=50,
path='owen.event.eventName'),
Column('eventPlatform',
DataType.VARCHAR,
length=25,
path='owen.event.eventPlatform'),
Column('eventPlatformVersion',
DataType.VARCHAR,
length=25,
path='owen.event.eventPlatformVersion'),
Column('eventTarget',
DataType.VARCHAR,
length=2048,
path='owen.event.eventTarget'),
Column('eventVersion',
DataType.VARCHAR,
length=25,
path='owen.event.eventVersion'),
Column('eventTimestamp',
DataType.DATETIME,
date_pattern="yyyy-MM-dd'T'HH:mm:ss'Z'",
path='owen.event.eventTimestamp'),
Column('derivedEventInstanceId',
DataType.VARCHAR,
length=64,
path='metadata.derivedEventInstanceId'),
Column('owenProcessed',
DataType.DATETIME,
date_pattern="yyyy-MM-dd'T'HH:mm:ss'Z'",
path='metadata.analyticsTopologyFinishTime'),
],
))))
def _get_engineless_static_subgraphs_by_related_type(graph_entity_service):
sub_graph_map = {}
d_entity_models = graph_entity_service.to_entity_models_with_randomized_ids(
[
Dataset(id=Ref.dataset(1),
data=DatasetData(None,
None,
None,
None,
None,
columns=[],
partitions=[]))
])
e_entity_models = graph_entity_service.to_entity_models_with_randomized_ids(
[Event(id=Ref.event(1), data=EventData('event'))])
sub_graph_map[None] = [
SubGraph(
name='dataset',
description='create a new dataset entity',
related_type=None,
related_is_a=None,
graph=graph_entity_service.to_graph(None, d_entity_models),
entity_map=graph_entity_service.to_entity_map(d_entity_models),
icon='⬟',
),
SubGraph(
name='event',
description='create a new event entity',
related_type=None,
related_is_a=None,
graph=graph_entity_service.to_graph(None, e_entity_models),
entity_map=graph_entity_service.to_entity_map(e_entity_models),
icon='★',
),
]
entity_models = graph_entity_service.to_entity_models_with_randomized_ids([
Subscription(id=Ref.subscription(1),
data=SubscriptionData('subscription', Ref.parent()))
])
sub_graph_map[EntityType.dataset] = [
SubGraph(
name='subscription',
description='create a new subscription entity',
related_type=EntityType.dataset,
related_is_a=Relationship.PARENT,
graph=graph_entity_service.to_graph(None, entity_models),
entity_map=graph_entity_service.to_entity_map(entity_models),
icon='⬢',
),
]
entity_models = graph_entity_service.to_entity_models_with_randomized_ids([
Trigger(id=Ref.trigger(1),
data=TriggerData(name='%s_trigger' % event_trigger.name,
trigger_type_name=event_trigger.name,
state=TriggerState.INACTIVE,
workflow_ids=[],
args={'event_id': Ref.parent()}))
])
sub_graph_map[EntityType.event] = [
SubGraph(
name='event trigger',
description='create a new event trigger entity',
related_type=EntityType.event,
related_is_a=Relationship.PARENT,
graph=graph_entity_service.to_graph(None, entity_models),
entity_map=graph_entity_service.to_entity_map(entity_models),
icon='▼',
),
]
entity_models = graph_entity_service.to_entity_models_with_randomized_ids([
Trigger(id=Ref.trigger(1),
data=TriggerData(
name='%s_trigger' % subscription_batch_trigger.name,
trigger_type_name=subscription_batch_trigger.name,
state=TriggerState.INACTIVE,
workflow_ids=[],
args={
'subscription_id': Ref.parent(),
'unconsumed_data_size_in_bytes': 1000000
}))
])
sub_graph_map[EntityType.subscription] = [
SubGraph(
name='subscription batch trigger',
description='create a new subscription batch trigger entity',
related_type=EntityType.subscription,
related_is_a=Relationship.PARENT,
graph=graph_entity_service.to_graph(None, entity_models),
entity_map=graph_entity_service.to_entity_map(entity_models),
icon='▼',
),
]
return sub_graph_map