However, we would like to clarify that infores:cam-kp should be treated as an intermediate aggregator knowledge source between the primary KS and Automat. So we would like the output to look like this:
I've tried to do that in this PR, but I have three questions:
Is this the correct way to use kgxedge.aggregator_knowledge_sources? I've tried to follow the usage in e.g. the MonarchKG parser, but I might have got something wrong.
Do I need to specify infores:automat-cam-kp as an aggregator source, or is that added automatically?
Is it possible to get each aggregator to reference the previous one in the upstream_resource_ids as I've specified above? Or will they all have "upstream_resource_ids": ["infores:go-cam"] as the sole primary knowledge source?
This was included in a different PR, and is in graphs now. The automat- infores is appended in plater, so we only needed to add infores:cam-kp as an aggregator in the parser.
Automat-CAM-KP sources currently look like this:
However, we would like to clarify that
infores:cam-kp
should be treated as an intermediate aggregator knowledge source between the primary KS and Automat. So we would like the output to look like this:I've tried to do that in this PR, but I have three questions:
infores:automat-cam-kp
as an aggregator source, or is that added automatically?upstream_resource_ids
as I've specified above? Or will they all have"upstream_resource_ids": ["infores:go-cam"]
as the sole primary knowledge source?You can see our discussion about this issue at https://github.com/ExposuresProvider/cam-pipeline/issues/118#issuecomment-1908569248