Utiliser un remote_id pour l'identification des ressources moissonnées

[maudetes]

Follows on the idea introduced in this discussion.

We could use a dedicated remote_id to identify harvested resources. It would tackle the issue opendatateam/udata-ckan#217.
We would need to have a relevant id for all harvest backends.

• ckan : use the UUID provided by ckan
• ods : the url (based on dataset_id + format) ? -> same behaviour as actual
• dcat : either the url OR the uri used to describe the distribution in the graph -> now stored in harvest.uri if an URIRef
• maaf ?

We could also consider removing resources with a remote_id set, but that is no more on the remote platform?

[abulte]

We could also consider removing resources with a remote_id set, but that is no more on the remote platform?

pretty neat!

[agarrone]

@maudetes still relevant ?