def createTransferLearningDAG(params):
today = datetime,now().strftime("%Y%m%d")
print("Creating Transfer Learning DAG for ", today)
default_cfg = master_env.get('default', {})
env_cfg = master_env.get(params_get('env', 'default'), {})
env = dict_merge(default_cfg, env_cfg)
default_args = {
'owner': 'Dr Cadx',
'depends_on_past': False,
'start_date': datetime(2020,1,1),
'email': cfg.get('email', ''),
'email_on_failure': True,
'email_on_retry': True,
'retries': cfg.get('airflow_retries', 1),
'retry_delay': timedelta(minutes=5),
'provide_context':True
}
dag_name = 'Transfer_Learning_Modeling'
print('Initialize DAG: {}'.format(dag_name))
dag = DAG(dag_name,
default_args=default_args)
dag.catchup = False
print('Retrieve The Transfer Learning Tasks')
dag_tasks = createTransferLearningTasks(params, default_args)
print('Sequencing the Transfer Learning Tasks')
dag >> dag_tasks.transfer_learning
# >> dag_tasks.validation
return dag
params = {
'model_name': Variable,get("model_name"),
'train_path': Variable,get("train_path"),
'test_path': Variable,get("test_path"),
'val_path': Variable,get("val_path")
}
dag = createTransferLearningDAG(params)
def create_dag(self,
dag_id,
schedule_interval,
catchup=None,
is_paused_upon_creation=None,
default_args={},
owner='Dagen',
**options):
default_args['owner'] = owner
default_args['pool'] = options.get('pool',
'default_pool') or 'default_pool'
dag = DAG(
dag_id,
default_args=default_args,
schedule_interval=schedule_interval,
is_paused_upon_creation=is_paused_upon_creation,
)
max_active_runs = options.get('max_active_runs', None)
if max_active_runs is not None:
dag.max_active_runs = max_active_runs
if catchup is not None:
dag.catchup = catchup
return dag
def load_target_db(**kwargs):
logging.info("Loading to target db ...")
# Create DAG
dag = DAG(
'skip_past',
description='Tests bash,operator,email cases',
start_date=datetime(2017, 11, 8, 23, 31, 00)
#,start_date=datetime.now()+timedelta(minutes=-20)
,
schedule_interval="*/5 * * * *"
#,default_args=default_args)
)
dag.catchup = False
def get_num_active_dagruns(dag_id, conn_id='airflow_db'):
# airflow_db = PostgresHook(postgres_conn_id=conn_id)
# conn = airflow_db.get_conn()
# cursor = conn.cursor()
# sql = """
# select count(*)
# from public.dag_run
# where dag_id = '{dag_id}'
# and state in ('running', 'queued', 'up_for_retry')
# """.format(dag_id=dag_id)
# cursor.execute(sql)
# num_active_dagruns = cursor.fetchone()[0]
#return num_active_dagruns
dest_table_countries = "umg-swift.swift_alerts.trending_tracks_countries"
dest_table_regions = "umg-swift.swift_alerts.trending_tracks_regions"
sql_wait = "SELECT * FROM `" + aux_table + "` WHERE _PARTITIONTIME='" + today_prime + "' LIMIT 2"
print("wait_sql=" + sql_wait)
#declare DAG
dagvelocitydaily = DAG(
'velocity_queries_daily',
description='Builds velocity daily by daily increments',
start_date=datetime(2017, 12, 14, 0, 0, 0),
schedule_interval="30 15 * * *"
#,schedule_interval = "None"
,
default_args=default_args)
dagvelocitydaily.catchup = False
task_streams_daily_velocity_base = VelocityStreamByDateOperator(
task_id="create_release_table_daily",
#schema_out = schema_out1,
schema_out=None,
sql=sql1.replace("@datePartition", 'date("{datePartition}")'),
destination_table=aux_table,
start_date=yesterday,
end_date=today,
dag=dagvelocitydaily)
#task_sleep = BashOperator(
# task_id='sleep',
# bash_command='sleep 1200',
# dag=dagvelocitydaily
{'name': 'lean_back_stream_count','type': 'INTEGER'},
{'name': 'users','type': 'INTEGER'},
{'name': 'users_isrc_day_sos','type': 'INTEGER'},
{'name': 'users_isrc_day_lf','type': 'INTEGER'},
{'name': 'users_isrc_day','type': 'INTEGER'},
{'name': 'load_datetime','type': 'TIMESTAMP'}
]
#declare DAG
dagarchivedaily = DAG('archive_first_year_daily'
,description='Builds archive by daily increments'
,start_date=datetime(2017, 12, 20, 0, 0, 0)
#,start_date=datetime.now()+timedelta(hours=-2)
,schedule_interval = "0 16 * * *"
,default_args=default_args)
dagarchivedaily.catchup=False
task_release_daily = CreateReleaseOperator(
task_id="create_release_table_daily",
sql = sql1,
destination_table="umg-dev.swift_alerts.isrc_first_stream_date",
dag=dagarchivedaily
)
# task_streams_daily = StreamByDateOperator(
# task_id="create_stream_by_date_daily",
# sql = sql2,
# destination_table="umg-dev.swift_alerts.track_archive_by_stream_date",
# #start_date="2017-10-26",
# start_date=str(datetime.now(pytz.utc)+timedelta(days=-1))[:10], # for daily processing - run yesterday only
# dag=dagarchivedaily
+ ' --input_file ' + inputfile + "" \
+ ' --temp_directory '+temp_directory +'' \
+ ' --schema track_history_schema.txt'
#loadGCStoBQ = "echo loadGCStoBQ is running"
loadGCStoBQ = loadGCStoBQ_beginning + loadGCStoBQ + t1nn_ending
print loadGCStoBQ
dagTask = DAG('swift_trends_subscriptions',
default_args=default_args,
schedule_interval=schedule,
start_date=datetime(2017, 11, 01, 0, 0, 0)
#,schedule_interval=timedelta(days=1)
)
dagTask.catchup = False
# t1,t2 and t3 are tasks created by instantiating operators
t11 = BashOperator(
task_id='build_email_table_mongodb',
#bash_command='java -cp /opt/app/swift-subscriptions/track-alerts/SwiftTrendSubscriptions-0.1.jar TrackAction.TrackActionSubscription --executionDateTest 2017-09-21 --project umg-dev --runner DataflowRunner --mongodb',
bash_command=step1,
dag=dagTask)
t12 = BashOperator(
task_id='build_email_table_enrichment',
#bash_command='java -cp /opt/app/swift-subscriptions/track-alerts/SwiftTrendSubscriptions-0.1.jar TrackAction.TrackActionSubscription --executionDateTest 2017-09-21 --project umg-dev --runner DataflowRunner --enrichment',
bash_command=step2,
dag=dagTask)
t13 = BashOperator(
task_id='build_email_table_major_sql',
# DAG configuration
#start_date = datetime(2017, 7, 12, 13, 0, 0, tzinfo=pytz.utc)
default_args = {
'owner': 'airflow',
'depends_on_past': False,
'schedule_interval': None,
'email': ['*****@*****.**'],
'email_on_failure': False,
'email_on_retry': False,
'retries': 0,
'retry_delay': timedelta(minutes=5)
}
# declare DAG
dagtest = DAG(
'test_bash_operator',
description='Tests bash,operator,email cases',
start_date=datetime(2017, 11, 03, 0, 10, 0)
#,start_date=datetime.now()+timedelta(hours=-2)
,
schedule_interval="*/5 * * * *",
default_args=default_args)
dagtest.catchup = False
dagtest.catchup_by_default = False
t11 = BashOperator(task_id='catchup_control',
bash_command="echo AAAAAAAAAAAAA BBBBBBBBBB CCCCCCCCCC " +
str(datetime.now()),
dag=dagtest)
