contributors_quick_start_vscode.rst

The outline for this document in GitHub is available at top-right corner button (with 3-dots and 3 lines).

Setup your project

1. Open your IDE or source code editor and select the option to clone the repository

   

2. Paste the copied clone link in the URL field and submit.

   

3. If you use official Python plugin you also have to add "tests" directly of each provider you want to develop as "Extra Paths". This way respective provider tests code can be addressed in imports as from unit.postgres.hooks.test_postgres import ... This is important in Airflow 3.0 we split providers to be separate distributions - each with separate pyproject.toml file. This might improve and might be better automated in the future, but for now you need to do it for each provider separately.

   To do this, open File -> Preferences -> Settings

   

   In Settings tab navigate to Workspace (this will set extra paths only for this project) and go to Extensions -> Pylance section. At Python -> Analysis: Extra Paths add the path to the tests directory of the provider you want to develop.

   

   NB: if you use pyright as LSP with other editor you can set extraPaths the same way in pyrightconfig.json, see pyright configuration docs.

4. Once step 3 is done it is recommended to restart VS Code.

Setting up debugging

1. Configuring Airflow database connection

• Airflow is by default configured to use SQLite database. Configuration can be seen on local machine ~/airflow/airflow.cfg under sql_alchemy_conn.

• Installing required dependency for MySQL connection in airflow-env on local machine.

  $ pyenv activate airflow-env
  $ pip install PyMySQL

• Now set sql_alchemy_conn = mysql+pymysql://root:@127.0.0.1:23306/airflow?charset=utf8mb4 in file ~/airflow/airflow.cfg on local machine.

1. Debugging an example DAG

• In Visual Studio Code open airflow project, directory /files/dags of local machine is by default mounted to docker machine when breeze airflow is started. So any DAG file present in this directory will be picked automatically by scheduler running in docker machine and same can be seen on http://127.0.0.1:28080.

• Copy any example DAG present in the /airflow/example_dags directory to /files/dags/.

• Add a __main__ block at the end of your DAG file to make it runnable. It will run a back_fill job:

  if __name__ == "__main__":
      dag.test()

• Add "AIRFLOW__CORE__EXECUTOR": "DebugExecutor" to the "env" field of Debug configuration.

  • Using the Run view click on Create a launch.json file

    

  • Change "program" to point to an example dag and add "env" and "python" fields to the new Python configuration

    {
        "configurations": [
            "program": "${workspaceFolder}/files/dags/example_bash_operator.py",
            "env": {
                "PYTHONUNBUFFERED": "1",
                "AIRFLOW__CORE__EXECUTOR": "DebugExecutor"
             },
             "python": "${env:HOME}/.pyenv/versions/airflow/bin/python"
        ]
    }

    

• Now Debug an example dag and view the entries in tables such as dag_run, xcom etc in mysql workbench.

Creating a branch

1. Click on the branch symbol in the status bar

   

2. Give a name to a branch and checkout

   

Follow the Quick start for typical development tasks.