Yeah - my preference would be we not be automatically creating stuff by default. I would prefer a -f/--force flag on shed_upload and a separate shed_create action (though we can deal with the separate action post PR). Is this okay with you?

Yep, absolutely. Give me a few minutes and I'll have it updated!

+1 with John, create repos carefully as they cannot be uncreated :)

@jmchilton okay, repo creation is now off by default and optional.

I'll make a separate PR with shed repo creation (though that starts to overlap with parsec)

@jmchilton okay, human names are now in use. .shed.ymls look like:

name: gatk2
owner: iuc
description: The Genome Analysis Toolkit in Version 2
long_description: |
  The Genome Analysis Toolkit or GATK is a software package developed
  at the Broad Institute to analyse next-generation resequencing data. The toolkit
  offers a wide variety of tools, with a primary focus on variant discovery and
  genotyping as well as strong emphasis on data quality assurance. Its robust architecture,
  powerful processing engine and high-performance computing features make it capable
  of taking on projects of any size.

  http://www.broadinstitute.org/gatk/about/

  Repository-Maintainer: Bjoern Gruening, Nicola Soranzo and Jim Johnson
  Repository-Development: https://github.com/galaxy-iuc/tool_shed
type: unrestricted
categories:
    - Sequence Analysis

the category name must match precisely (capitalisation/spaces aren't stripped).

Coverage decreased (-1.53%) to 66.38% when pulling 3b50ddf on erasche:create-repo into a34e5fc on galaxyproject:master.

Oh, lovely, now coveralls is kvetching at me in addition to the linters.

@jmchilton afaict none of the toolshed interactions are actually under test, and I'm not keen to write those...

Ignore coveralls - it is annoying.

Update: Disabled the alerts (I certainly never meant to enable them in the first place :)).

@jmchilton anything else you want for this PR?

Not at first glace - this looks pretty awesome - thanks! I want to find some time to actually play with it before merging though.

@martenson You have thought a lot more than me about this issue - does this seem right to you? Do you feel like the correct things are being stored in .shed.yml in the correct format.

@jmchilton sure, no rush! On that note, there may be a small bug in type, which I need to double check.

@jmchilton @erasche I think it is a good start but we know the target is moving. We are not creating standard yet, just exploring.

@jmchilton okay, that bug in type is actually upstream in galaxyproject/galaxy#31 (so this commit should be technically correct, but my comments on "no rush" still stand)

Crap - I merged the commits with a real merge so the PR would close when I pushed but then I rebased things out of vanity and the merge information got lost. Sorry about that - the commits are nonetheless in there - I hope that is okay.

No worries! Also, thanks for adding the test case. I'll make sure my future PR make some attempts to include those :)