View a markdown version of this page

Invoke the DAG - Scene Intelligence with Rosbag on AWS

Invoke the DAG

Complete the following steps to invoke the DAG.

  1. Sign in to the Amazon MWAA console.

  2. Select Environments from the navigation menu.

  3. Select addf-aws-solutions-core-mwaa-environment and choose Open Airflow UI. This opens a new window to the Airflow DAG.

    Example Airflow environments screen listing addf-aws-solutions-core-mwaa-environment and Open Airflow UI link.

    airflow dag

  4. Choose the vsi-image-pipeline DAG.

  5. Choose Play and select Trigger DAG w/ config.

    Example vsi_image_pipeline screen with Trigger Dag w/ config option in upper right corner.

    vsi image pipeline

  6. Provide the paths to the staged rosbag files. The following is an example, where you replace <vehicle-name> with the vehicle name, <bucket> with the S3 bucket name, and <prefix-to-dag> with your prefix:

    {
 "drives_to_process": {
 "<vehicle-name>": {
 "bucket": "<bucket>",
 "prefix": "rosbag-scene-detection/test-vehicle-02/072021/ "
 },
 "<vehicle-name>": {
 "bucket": "<bucket>",
 "prefix": "<prefix-to-dag>"
 }
 }
}
    Note

    The solution requires a minimum of two rosbag entries to invoke the DAG.

The following is an example with these fields completed, where the vehicle names are test-vehicle-01 and test-vehicle-02, the bucket name is addf-aws-solutions-raw-bucket-123456, and the prefix is rosbag-scene-detection/test-vehicle-01/072021/:

{
 "drives_to_process": {
 "test-vehicle-01": {
 "bucket": "addf-aws-solutions-raw-bucket-123456",
 "prefix": "rosbag-scene-detection/test-vehicle-01/072021/"
 },
 "test-vehicle-02": {
 "bucket": "addf-aws-solutions-raw-bucket-123456",
 "prefix": "rosbag-scene-detection/test-vehicle-02/072021/"
 }
 }
}

The following image shows the previous example code in the Airflow DAG window:

Example vsi_image_pipeline screen displaying previous sample code and Trigger button.

vsi image pipeline trigger

  1. Choose Trigger. Wait for the DAG to complete running (approximately 20 minutes).