aggregate --> new service and bench action

[mpo-vliz]

reconsidering https://github.com/vliz-be-opsci/rdf-aggregate-action

this basic service should be part of sema

• follow the service pattern we have
• take globs list(str) as argument
• provide cli sema-aggregate sript
• wrap up as available sema-bench service

while at it

• use commons.glob to deal with globbing
• use commons.store to deal with loading into mem-store + applying known cleaning rules --> but have logging about usage of http://schema.org