This is a package for enabling access to an Airflow (https://airflow.apache.org/) service via a RESTful interface. It is currently only functional with a LocalExecutor.

In order to integrate this package with an existing Airflow service you can either:

1. Copy the entirety of the source code at airflowapi and the plugins.py file into a folder at $AIRFLOW_HOME/plugins in the airflow service.

   mkdir -p $AIRFLOW_HOME/plugins && cp -r airflowapi $AIRFLOW_HOME/plugins && cp plugins.py $AIRFLOW_HOME/plugins
   

2. Pip install the airflowapi package into the same python distribution that is running the airflow service and copy the plugins.py file into a folder at $AIRFLOW_HOME/plugins in the airflow service

   pip install . && mkdir -p $AIRFLOW_HOME/plugins && cp plugins.py $AIRFLOW_HOME/plugins
   

The api's documentation can be found at the api/v1/doc route of the airflow service.

In order to do development you will need a python 3.6 environment set up as your base python installation.

We encourage developers to understand and leverage either of the following to create a clean environment for development:

• conda via Anaconda or Miniconda
• virtualenv

To run all tests simply run the following command:

./bin/run_tests.sh

This assumes you are developing on Mac OSX System. Development on other systems is currently not tested or documented.

You will need docker installed on your system in order to run these tests.

You will first need to build the base image used in testing. This helps speed up development by establishing system values which aren't often changed. You can do so by running the following command:

docker build -f AirflowTestEnvDockerfile -t "airflow-api:base-image" .

You will need to add the following to your /etc/hosts file:

127.0.0.1       airflow-webserver

This can be done easily bu running the following command:

sudo echo "127.0.0.1 airflow-webserver" >> /etc/hosts 

To run all tests simply run the following command:

./bin/run_integration_tests.sh

If you would like to run tests in debug mode you can run the following to spin up the docker images needed:

./docker-compose-wrapper.sh up

When you are done with the images you can spin them down with this command:

./docker-compose-wrapper.sh down

The following are the tested application configurations which are confirmed to work: