data-mappings architecture to be dataset-driven instead of topic-driven

[maaikelimper]

Right now the wis2box-architecture processes data based on the directory that is presumed to match the topic-hierarchy.

However users can have different datasets on the same topic-hierarchy (example: Chile with manual and AWS data).

The data-mappings architecture should be dataset-driven instead of topic-driven, which requires a redesign of things currently work.

Related:
#584

[maaikelimper]

proposal:
incoming bucket + dataset_id

[maaikelimper]

I have noted that our auth options are based on locking down topics, I'm updating this to have the command-lines for locking down datasets. This might require changes in wis2box-auth as well. I also note that we are missing automated testing that our command-line options for locking down topics/datasets, I should add this.

[maaikelimper]

This was implemented by #686