Closed jdries closed 1 year ago
Just to be clear, what version do you mean (and what does it include)?
It's closest to option 1, but basically would be defined by the user who creates the job. The main problem is that small changes to the parameter in a process graph are not necessarily version changes. For instance, a graph with the same version can be applied to different spatial extents. This interpretation is something that can only be done correctly by the author of the process graph.
Aha, I see.
Some thoughts:
stac
property to the job submission, which shall be merged with the resulting batch job metadata? That could be generally nice that you can customize the STAC output a bit. Then you could set the STAC version
property. That's of course limited to batch jobs.What do you think?
Related issue: https://github.com/Open-EO/openeo-api/issues/472
as already mentioned, I had same kind of question in #472 about adding extra PG metadata (e.g. client software version)
Closing in favor of #472
As part of reproducibility & traceability, users are asking me how to record the version of the process graph itself in the output metadata. We already record backend version and the process graph itself, but for production projects, our users also attach versions to the workflow itself. I see two locations where this can be relevant:
My initial proposal would be to have a top-level property 'version' next to the others like title and description.