Easier deployment of sample resources, metadata, and annotations

[manthey]

Currently, the provisioning script will install a fixed set of samples if requested. It would be nice if this could be specified more explicitly. Something like:

initial-data:
  - source-host: <girder api url>
    source-resource-path: </collection/sample/folder>
    destination-resource-path: </collection/sample/folder>
    destination-owner: <user id>
    include-metadata: <boolean>
    include-annotations: <boolean>

[manthey]

@dgutman @cooperlab

I was thinking we should have 2 or more sample data sets for the DSA that are quasi-real that would include images, metadata, and annotations. They should be too big, but should be big enough to show off some of the features. Maybe something like a half dozen of the TCGA images from the crowd source paper including annotations and marked up with the TCGA source metadata, plus an example that would show off brain samples or something.

I wrote a simple little script that will package everything from a folder and its subfolders into a zip file and extract the same. In the ideal world, there would be a process with only a few steps to let a user get these samples and add them to a collection (DSA Demos?) so they could play with them. I have a provisional workflow which would be (a) install a slicer_cli_web algorithm plugin that has a "Get Demo Set" task, (b) Navigate to the "Get Demo Set" task, (c) paste a URL to a zip file, (d) click "Run". This still feels like too many steps for the novice. I'm always torn between having a magical "get data" method and showing someone how something actually works.

[cooperlab]

This is a great idea and for this one I would lean towards the magical "get data" method.