I can run the single file as a dataflow job in cloud composer but when i run it as a package it fails .
pipeline_jobs/
-- __init__.py
-- run.py (main file)
-- setup.py
-- data_pipeline/
----- __init__.py
----- tasks.py
----- transform.py
----- util.py
I'm getting this error:
WARNING - File "/tmp/dataflowd232f-run.py", line 14, in <module
{gcp_dataflow_hook.py:120} WARNING - from data_pipeline.tasks import task
WARNING - ImportError: No module named data_pipeline.tasks.
This is the dag configuration:
from datetime import datetime, timedelta
from airflow import DAG
from airflow.contrib.operators.dataflow_operator import DataFlowPythonOperator
default_args = {
'owner': 'airflow',
'depends_on_past': False,
'start_date': datetime.strptime("2017-11-01","%Y-%m-%d"),
'py_options': [],
'dataflow_default_options': {
'start-date': '20171101',
'end-date': '20171101',
'project': '<project-id>',
'region': '<location>',
'temp_location': 'gs://<bucket>/flow/tmp',
'staging_location': 'gs://<bucket>/flow/staging',
'setup_file': 'gs://<bucket>/dags/pipeline_jobs/setup.py',
'runner': 'DataFlowRunner',
'job_name': 'job_name_lookup',
'task-id': 'run_pipeline'
},
}
dag = DAG(
dag_id='pipeline_01',
default_args=default_args,
max_active_runs=1,
concurrency =1
)
task_1 = DataFlowPythonOperator(
py_file = 'gs://<bucket>/dags/pipeline_jobs/run.py',
gcp_conn_id='google_cloud_default',
task_id='run_job',
dag=dag)
I tried putting run.py into dags folder but still getting same error. Any kind of suggestions would be really helpful.
Tried doing this as well: from pipeline_jobs .data_pipeline.tasks import task but still the same issue.
