Thanks all,
  it seems that my real problem is that the audio file (.wav) is not being identified as a valid datatype and ending up as a zero length text file. So, I need to start to explore the world of datatypes. 

Following the docs (https://wiki.galaxyproject.org/Admin/Datatypes/Adding%20Datatypes) I can modify datatypes_conf.xml in my Galaxy sources and add a new datatype for wav files:

    <datatype extension="wav" type="galaxy.datatypes.binary:Binary" display_in_upload="true" mimetype="audio/wav" subclass="True”/>

but, I get a message "The uploaded binary file contains inappropriate content” and a zero length file just as I did before adding this - although the datatype is now set to ‘wav’.  

I didn’t add a sniffer for this and set the datatype explicitly on upload. 

Also, this doesn’t seem like a modular way to add datatypes - how do I include datatypes in my tool definition?  I can see from some other tools that I include a datatypes_conf.xml in my tool folder.   When I try that and test with planemo the new type isn’t found.  

Pointers welcome.

Thanks,

Steve

—
Department of Computing, Macquarie University
http://web.science.mq.edu.au/~cassidy

On 21 Oct. 2016, at 12:58 pm, Peter Cock <p.j.a.cock@googlemail.com> wrote:

Using a soft link for this is a common pattern, and should be followed with &&
(ideally using XML CDATA to avoid escaping everything like &amp; etc),
and quote the filenames just in case there are any spaces. e.g.

https://github.com/galaxyproject/tools-iuc/blob/master/tools/trinity/run_de_analysis.xml#L16


For reference, in tools-iuc there are over 400 soft link examples:

$ grep "ln -s" tools/*/*.xml | wc -l
    446

Peter

On Fri, Oct 21, 2016 at 5:48 PM, Steve Cassidy <steve.cassidy@mq.edu.au> wrote:
Hi,
I’m wrapping a tool that needs it’s input to have a known file extension
(an audio file, eg. .wav).  Since Galaxy stores all data as .dat files the
tool is falling over since it doesn’t know what .dat is.

I thought I’d be able to get around this by hard linking the .dat file to
the same name with a .wav extension (dataset_1.dat.wav), this works when I
try it with the tool on the command line but within Galaxy it fails, here’s
my <command>:

       ln $signal ${signal}.wav &amp;
       /home/maus/maus OUTFORMAT=TextGrid LANGUAGE=$language
       BPF=$bpf INSKANTEXTGRID=$inskantextgrid
INSORTTEXTGRID=$insorttextgrid
       MODUS=$modus MAUSSHIFT=$mausshift MINPAUSLEN=$minpauslen
WEIGHT=$weight
       INSPROB=$insprob NOINITIALFINALSILENCE=$noinitialfinalsilence
OUTSYMBOL=$outsymbol
       OUT=$output SIGNAL=${signal}.wav

resulting in the job command line:

ln /tmp/tmp7AZvx7/files/000/dataset_2.dat
/tmp/tmp7AZvx7/files/000/dataset_2.dat.wav & /home/maus/maus
OUTFORMAT=TextGrid LANGUAGE=aus BPF=/tmp/tmp7AZvx7/files/000/dataset_1.dat
INSKANTEXTGRID=false INSORTTEXTGRID=false MODUS=standard MAUSSHIFT=10
MINPAUSLEN=5 WEIGHT=7.0 INSPROB=0.0 NOINITIALFINALSILENCE=no OUTSYMBOL=sampa
OUT=/tmp/tmp7AZvx7/files/000/dataset_3.dat
SIGNAL=/tmp/tmp7AZvx7/files/000/dataset_2.dat.wav

I’m getting an error message from the tool:

sox FAIL formats: can't open input file
`/tmp/tmp7AZvx7/files/000/dataset_2.dat.wav': WAVE: RIFF header not found

this suggests that the hard link didn’t get made.  I tried copying the file
instead but got the same result.

I could go in and patch the tool script to be more forgiving but it would be
good to find a solution that didn’t require that if possible.

Any pointers appreciated.

Steve
—
Department of Computing, Macquarie University
http://web.science.mq.edu.au/~cassidy


___________________________________________________________
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:
 https://lists.galaxyproject.org/

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