Allow 'planemo run' to stage in existing datasets without reuploading

[simonbray]

Fixes galaxyproject/planemo#1080. This was much easier than I expected.

The idea is that the job or test yaml file can contain galaxy_ids for any File or Collection instead of a local path. Assuming these are valid for a particular galaxy server, the user can then use planemo run and execute the workflow using the existing datasets/collections and therefore can skip the upload step.

mgf_input_list:
  class: Collection
  galaxy_id: 1e94904b7ba7b9cc
metapeptides:
  class: File
  galaxy_id: 11ac94870d0bb33a71934a4bcff13734
gene_ontology_terms:
  class: File
  galaxy_id: 11ac94870d0bb33aa74376be3811a4fe

I tested this with both files and collections and it seems to work fine. Also, specifying galaxy_id for some files and location for others (i.e. so some files are uploaded and others not) works.

[jmchilton]

I'm glad that was so simple! This seems awesome to me.