I suggested augmenting the tool_conf syntax as part of the DataCollection 
development.

To replace the need for the the multiple output determined at runtime,
I suggest being able to declare data collections within the outputs  tags,  and 
being able to use regular expressions in the from_work_dir  param to populate 
the collections.
In workflows, one would want to be able to hook a data collection output to a 
data input.


Mothur Metagenomics tool that has an output per distance label and calculator 
method

An example of declaring a list of outputs, which will determined at run time 
based on from_work_dir regular expression:
<tool id="mothur_classify_otu" name="Classify.otu" version="1.20.0" 
force_history_refresh="True">
  ...
  <outputs>
    <dataset_collection type="list" label="${tool.name} on ${on_string} consensus 
taxonomies">
      <data format="cons.taxonomy" name="splicing_diff" label="${tool.name} on ${on_string}: 
${file_name}" from_work_dir="^\S+?\.(unique|[0-9.]*\.cons\.taxonomy)$" />
    </dataset_collection>
    <dataset_collection type="list" label="{tool.name} on ${on_string} taxomy 
summaries">
      <data format="cons.taxonomy" name="splicing_diff" label="${tool.name} on ${on_string}: 
${file_name}" from_work_dir="^\S+?\.(unique|[0-9.]*\.cons\.tax\.summary)$" />
    </dataset_collection>
  </outputs>


Hey Peter,

Have you seen this solution?

https://wiki.galaxyproject.org/Admin/Tools/Multiple%20Output%20Files#Number_of_Output_datasets_cannot_be_determined_until_tool_run

It always seems to get mentioned when this topic is brought up. It has
serious limitations in terms or workflow running, but it can probably
be made to work for individual tool executions.

Otherwise I would wait for future feature sets, maybe other people
have some good ideas however.

-John

On Wed, Feb 19, 2014 at 7:16 AM, Peter Cock<p.j.a.c...@googlemail.com>  wrote:
Hi all,

I'm looking for examples of tools which take multiple input
files (one or more, determined at run time) and produce
multiple output files (one for each input file). Any
specific suggestions?

I have a number of sequence filtering/renaming tools
where this might be useful - in some cases taking
multiple input files and producing a single output is
fine, but in general I'd like to know how to preserve a
one to one mapping from input files to output files.

I realise this may overlap slightly with the work John is
doing on dataset collections, but for now I'd like to target
the current Galaxy feature set.

In some of the simpler cases, if I have N input datasets
and want N output files, I can just run the tool N times .
This means more steps in the Galaxy GUI, but it isn't
very complicated.

However, for the current problem I need access to all
the inputs at once for setting overall data derived
parameters.

Regards,

Peter
___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
   http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
   http://galaxyproject.org/search/mailinglists/



--
James E. Johnson, Minnesota Supercomputing Institute, University of Minnesota
___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
 http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
 http://galaxyproject.org/search/mailinglists/

Reply via email to