On 10/21/11 12:29 PM, James Taylor wrote:
Excerpts from Jim Johnson's message of 2011-10-21 17:13:02 +0000:
I put the gmap tool suite in the galaxy Tool Shed, let me know if there is more I should do.
I added a requirement tag for the datatypes to the tool-configs:
% grep 'requirement.*datatype' *.xml gmap_build.xml:<requirement type="datatype">gmapdb</requirement>
Requirement tags for datatypes are an interesting idea, but I'm wondering if this is something we should require? It seems like all this information is implicit -- a tool requires a datatype if it has an input or output parameter that references that type. Is there other information that should go in the requirement tag?
That is certainly correct that the tag would be redundant, the tool config parser could identify the list of datatype formats.
I was just trying to think of some way to indicate that additional datatypes were required above those in the central distribution. My goal would be to have the installation of tools from the Tool Shed also be able to install the extra datatypes that those tools require.
Having datatypes specified separately in the Tool Shed from tools would hopefully promote less redundancy of datatypes and better interoperability among developers tools. For example the metagenomics applications mothur and qiime have many specific formats that are internal to their tools, but also a few that might be used to migrate data between those applications. We'd need a way to avoid name clashes, perhaps adopting a namespace pattern for the file_ext attribute.