Expose element_identifier as a workflow parameter variable.

[jmchilton]

@lparsons and I had a heated debated in IRC as to whether tools should consume this or workflow authors should be setting these up (revolving around Tophat which doesn't have the enhanced read group handling of the bwa/bowtie wrappers (xref https://github.com/galaxyproject/tools-devteam/pull/103/files)). Regardless about what the "best practice" in general should be - there are always going to be cases where it would be good at the workflow level to be able to consume collection metadata.

[mvdbeek]

The element_identifier is a very useful concept, so access in the workflow editor would be great (though I guess using element_identifier in the tool xml covers a lot of ground). 👍

[lparsons]

One instance where I think this would be very helpful is when exporting data from Galaxy. Currently, the files are not named using the element identifiers at all, so allowing users to rename based on this would be great. I might suggest, that, as a default, datasets are named based on the collection name and the element id whenever possible, this would make finding datasets much easier, IMHO.

[lparsons]

I'd like to add one thing to this, it would be very helpful if all datasets had an element identifier regardless of their membership in a collection. To start with, if the element_identifer simply resolved to the dataset name in cases where the dataset is not part of a collection, that would suffice.

[jmchilton]

Within tool wrappers element_identifier does fall back to the name - just be careful about spaces and malicious characters.

    @property
    def element_identifier( self ):
        identifier = self._element_identifier
        if identifier is None:
            identifier = self.name
        return identifier