As explained here, a user can create an EO processing job and share it with other users. We can also help MEP-TPS users exposing an existing Spark job as an on-demand service, similar to the N-daily compositor service.

In order to request the setup of an on-demand service, contact the helpdesk and provide the following information:

  • Spark job documentation and a pointer to the Spark job you want to expose (e.g. location on the cluster, Git, etc ...). Of course, you will have to create this job yourself. Of course, we will provide support if needed.
  • input data needs (name of the collection, can the service be triggered by a subscription, etc ...).
  • list of job input parameters which can be modified by the end user.
  • mockup of the UI form used for submitting the job; we can provide a dedicated user interface allowing the user to specify the job parameter values to submit the processing job.
  • describe how you want to make the job results available to the end user; currently, we can distribute the results via FTP, but we are working on pushing results to GeoServer, allowing WMS/WCS support.
  • access restrictions (if needed)

After evaluating your request, we will further guide you through the process of exposing your service.