Skip to main content

Manually invoking ingestion from an AWS S3 bucket

Ingestion of robolog metadata can either be triggered manually by POST request to our /api/ingest/v1/trigger endpoint, or automatically by registering your bucket with us as outlined in Automating Ingestion. While our recommendation is for you to create a workflow that enables automatic ingestion, it is important that you also know how to manually trigger ingestion.

Step 1. Determine the robolog type​

Model-Prime currently supports ingestion of ROS 2 metadata .yaml and .mcap files (ROSBag2) and ROS 1 .bag files (ROSBag1).

Step 2. Determine the file location for the metadata.​

For manual ingestion to work, you will need to make a POST request to our /api/ingest/v1/trigger endpoint with the following information:

  • The AWS Account ID for the owner of the bucket
  • The log type
  • The location of the metadata file to be ingested

For this example, we will assume you have an AWS account ID of 123456789000 and you want to ingest metadata of the type ROSBag2 from a bucket named modelprime.example.bucket where the metadata is located at /robodata/d006b838-ee73-4430-bde3-2d308c64631d/metadata.yaml.

Step 3. Trigger the ingest manually​

In the Model-Prime web app, navigate to the Ingest Dashboard. Click on the widget that says "Manually ingest a robolog from AWS S3."