[MOBY-l] MOBY tutorial - Taverna
Paul Gordon
gordonp at ucalgary.ca
Tue Nov 21 19:17:22 UTC 2006
Hi all,
I finally found some time to look at this Taverna/MobyServlet
incompatability.
It seems that Taverna does not include an XML declaration (<?xml...?>)
at the start of the MOBY payload. Because the payload did not have
this, MobyServlet tried as a last ditch effort to Base64 decode the
contents (some old servers sent data this way). That's why you got
garbage. Now, between not finding the XML declaration and attempting a
base64 decoding, I've inserted a check for the MOBY root tag. Now
Taverna jobs work, and will produce the following message in the
catalina log:
Warning: The MOBY contents was missing an XML declaration, but it is
required by the MOBY API, and may stop working in the future without
it. Please contact the client's provider to correct this.
Perhaps I am wrong, but the MOBY API specifically says that this
declaration is required, no? That's why we have to string-encode or
base64 encode, etc. the SOAP contents (since processing instructions
starting with "xml" are not allowed in the contents of an XML
document)? If not, I'll get rid of the warning message...
With kind regards,
Paul
P.S. It may take a few hours for the Web site MobyServlet.war to be
mirrored, but the one from CVS is updated.
> (Thanks to Paul for the test client fix.)
>
> I'm having trouble using the ConvertAAtoFASTA_AA servlet example from
> within Taverna. I made a simple workflow to test this, and everything
> works perfectly if I send the biomoby service a fully formed XML input
> but not if I build the input using biomoby processor objects. The
> workflow runs both cases in parallel:
>
> Workflow: http://nassar.europa.renci.org/biomoby_aa2fasta.xml
> Inputs: http://nassar.europa.renci.org/biomoby_aa2fasta_input.xml
> Error detail: http://nassar.europa.renci.org/biomoby_aa2fasta_error.xml
>
> The error is "Task failed due to problem invoking biomoby service. /
> SAXParseException: Premature end of file." Tomcat's log reports
> non-XML in the SOAP message (and it actually appears to be non-ASCII).
> I also see differences in the input XML constructed by the
> AminoAcidSequence object, such as missing articleName='inseq' in
> element moby:Simple.
>
> I've tried this with the Taverna 1.4 release, the stable snapshot from
> October 2, and recent builds from cvs (hopefully including the newer
> biomoby plug-in module), and it seems I get the same results in all
> cases.
>
> Maybe I'm building the input objects incorrectly?
>
> Nassib
> _______________________________________________
> moby-l mailing list
> moby-l at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/moby-l
>
>
>
More information about the moby-l
mailing list