Merge BLAST XML tool

[pvanheus]

As discussed on galaxyproject/tools-iuc#1893 this turns the merge code from the Galaxy BLAST XML datatype into a tool that can merge a list of datasets (or multiple individual datasets) into a single BLAST XML dataset.

[pvanheus]

BTW this would be made obsolete by galaxyproject/galaxy#5464 if it was achieved.

[peterjc]

Assuming we later have blastxml2 defined as another datatype, is there anything you'd change now to allow this to merge multiple blastxml files OR multiple blastxml2 files? Or would that be a second tool?

(We would not want to try merging a mixture unless someone produces a robust script to interconvert from one format to the other).

[pvanheus]

If blastxml2 is similar enough to be merged by the same tool, I think the simplest solution would be to make the format of the input and in the #for loop check that the inputs were consistent in the type being used (i.e. use a combination of #set, #if and $element.ext).

[pvanheus]

If I read the Travis logs right the error isn't coming from this tool. Am I right?

[peterjc]

Grr. Another new TravisCI failure, but not your fault:

Applying linter tool_urls... FAIL
.. ERROR: Error '404 Client Error: Not Found for url: http://www.jcvi.org/cms/research/projects/tigrfams/overview/' accessing http://www.jcvi.org/cms/research/projects/tigrfams/overview/

Update: Cross reference d31e21d and https://twitter.com/pjacock/status/998941086168076288

[peterjc]

The error I am expecting to see is against the Galaxy master branch only,
galaxyproject/galaxy#6180

[pvanheus]

@peterjc sadly the test against dev failed too - with errors I can't quite make out.

[peterjc]

Hmm. At the very least there were SQLite locking issues, which seems to be an intermittent problem. I've restarted that test on TravisCI, fingers crossed it gets further.

[peterjc]

The dev branch failed again on re-testing.

In good news, the yarn version issue problem has been fixed on the Galaxy master branch, so I've asked TravisCI to retest that.