Airflow Plugins for Rekcurd Dashboard
- Python 3.6
- apache-airflow >= 1.10.x
- Rekcurd Dashboard >= 1.0.x
- Setup Airflow
- Setup Rekcurd Dashboard and get your access token
- Run the following commands
Use JWT token published by Rekcurd Dashboard as REKCURD_ACCESS_TOKEN for now
# Set the access token and airflow-rekcurd connection to airflow
# Replace the environment variables with your own values.
$ cd $AIRFLOW_HOME
$ airflow variables -s rekcurd_access_token $REKCURD_ACCESS_TOKEN
$ airflow connections -a --conn_id rekcurd_dashboard \
--conn_uri http://$REKCURD_DASHBOARD_HOST:$REKCURD_DASHBOARD_PORT/
$ pip install rekcurd-airflow
You can see example DAGs for the plugins in here
- EvaluationUpdateOperator. (Example DAG)
- ModelDeleteOperator. (Example DAG)
- ModelEvaluateOperator. (Example DAG)
- ModelSwitchOperator. (Example DAG)
- ModelUpdateOperator. (Example DAG)
$ cd airflow-plugin # move to the root directory of this repo
# not necessary if you already setup airflow for this repository
$ export AIRFLOW_HOME=`pwd`
$ pip install -r test-requirements.txt
$ airflow initdb
$ fernet_key=`python -c "from cryptography.fernet import Fernet; print(Fernet.generate_key().decode())"`
$ sed -i -e "s/\(fernet_key =\).*/\1 $fernet_key/" airflow.cfg
$ python -m unittest tests/**/*.py