Taverna Tips

Contents

Background

TavernaWorkbench is an open-source workflow tool which provides a workflow language (Scufl) and graphical interface for building, running and editing workflows over distributed computer resources.

• For general and scientific workflow information see WorkflowLinks or WorkFlow#Useful_Links
• For available web services see Duncan Hull's list of Biological Web Services which replaces the out of date BioinformaticsWebServices
• Taverna includes facilities to access services provided byBioMOBY, SeqHound, Biomart, and EMBOSS at EBI.
• Soaplab and its subproject Gowlab can be used to provide web service interfaces to command line applications and web pages respectively.

Recent Developments

WindowsShortcut, ExecutingWorkflowOutsideOfTaverna, UsingRProcessor, TavernaDesiderata, TavernaOMII

• Release1.4 Available 7th June 2006 details at Taverna news blog
• Release1.3.1 Available November 2005 and major revision of Taverna website
• Release1.2 Available 21st June 2005, see Taverna 1.2 Release Notes for details
• Release1.1 Available 28th April 2005
• Release1.0 Available 24th January 2005, see User Manual for details - includes integration with Biomart and more example using Beanshell processors.
• Beta0.12 Available 8th November 2004, many exciting new features descriptions in release notes see what's new in Beta12
• Beta0.11 Offline Mode, Workflow Resource Summary, SeqHound Integration, and new quick start documentation Sept 2004
• Beta0.10 ISMB build - some new improvements from first beta0.10 - uses myGrid LSID authority and store replacing Ouzo prototype - July/Aug 2004
• Beta0.10 Significant UI revisions, data input wizard, support for BioMoby services, (for more details see the Latest Release notes on the Taverna web site) - June 2004
• Significant redesign of web site - June 2004
• Beta0.9 RDF provenance templates, Ouzo LISD authority see Tom's brief notes - March 2004
• Beta0.8 Reliability features, new input window, enhanced results viewing, new look and feel, - February 2004

Documentation

The primary documentation is on the Taverna web site http://taverna.sourceforge.net under the Documentation and old Docs Home links. This includes links to the papers on the myGrid website and many of the resources mentioned below. Remember Taverna develops in response to user requests and the latest version is often ahead of the documentation. If you have a query do try the mailing lists.

For a quick introduction, there is the one page ISMB 2005 demo description, or the Nucleic Acids Research article

Peter Li developed an online Taverna Tutorial as part of the University to Newcastle's MSc course CSC861: e-Science for Bioinformatics (also available on the Taverna web site under Documentation Tutorial)

Papers

• "Taverna: a tool for building and running workflows of services" Nucleic Acids Research 2006 34(Web Server issue):W729-W732 doi:10.1093/nar/gkl320 is a good introduction
• "Taverna: Lessons in creating a workflow environment for the life sciences" published in the Workflow in Grid Systems special issue of Concurrency and Computation: Practice and Experience online article published Dec 2006, paper version Volume 18 Issue 10 (25 August 2006), Pages 1067-1100
  • CCPE Taverna paper online published online Dec 2005, in print Aug 2006.
  • CCPE Taverna paper - preview edition as accepted for Concurrency and Computation: Practice and Experience special issue on Workflow in Grid Systems citation details given above.
• "Panoply of Utilities in Taverna" describes extensions to the early versions of Taverna in response to user feedback Springer online version
• "Taverna: A tool for the composition and enactment of bioinformatics workflows" was the first Taverna journal article, published in the Bioinformatics journal Nolume 20 no 17. (2004) Peter Li did the hard work putting this together.
  • Access to Taverna paper through OUP Bioinformatics site
  • Bioinformatics Advanced Access 16 June doi:10.1093/bioinformatics/bth361
• "Delivering Web Service Coordination Capability to Users" was submitted to www2004, but only accepted as a poster. The 2 page note in the proceedings, concentrating on Scufl is below.
  • p953-oinn.pdf: Scufl 2 pages for www2004 proceedings ( copy on eprints )
• "Experiences with eScience workflow specification and enactment in bioinformatics" at UK e-Science AllHands2003 was the first paper to describe the Taverna, Scufl, Freefluo approach
  • http://www.nesc.ac.uk/events/ahm2003/AHMCD/pdf/108.pdf ( local copy )
• Recycling Services and Workflows through Discovery and Reuse, to appear in Concurrency and Computation: Practice and Experience, particularly looks at the issues of the semantic description of services and workflows
  • http://www.cs.man.ac.uk/~goderisa/wroe+lifecycle-CCPE.pdf

• A paper was presented at the GGF 10 workflow workshop in March 2004. This has been revised to a journal paper to appear in Concurrency and Computation: Practice and Experience see WorkflowAtGGF for details.

Papers that discuss the use of myGrid including Taverna

• Exploring Williams-Beuren Syndrome Using myGrid , ISMB 2004
  • online abstract
• Association of variations on I kappa B-epsilon with Graves' disease using classical and myGrid methodologies, AllHands2004
  • http://www.allhands.org.uk/proceedings/papers/269.pdf
• Exploring Williams-Beuren Syndrome Using myGrid, AllHands2004
  • http://www.allhands.org.uk/proceedings/papers/253.pdf

Papers the discuss Taverna with other Bioinformatics tools

• MOBY Support in Taverna — A powerful front-end to the MOBY interoperability system in Canadian Bioinformatics Help Desk (CBHD) Newsletter descibes Taverna and BioMOBY
• SeqHound: The Next Iteration in Canadian Bioinformatics Help Desk (CBHD) Newsletter mentions Taverna as an interface to SeqHound
• NETTAB 2005 not a paper but workshop on the design and implementation of bioinformatics workflows

See also the folowing sections on the myGrid web site

• publications
• Taverna workbench

Discussion Documents

There is some developer documetation on the Taverna web site Documentation page including the JavaDoc documentation of the Taverna Java APIs.

• myGridWorkflowDebug.doc: Troubleshooting Scufl Workflows

• TavernaIntegrationOptions
• TavernaLabBookStoryBoard
• XScufl language reference from web site - beta10 onwards
  • XScufl specification April 2004 - undergoing further revision
• Drag and Drop discussion note
• Introduction to reliability features beta8
  • Processor state transitions diagram for a detailed understanding of processor execution
  • reliability and fault tolerance discussion note [Apr 04]
  • Reliability features for beta0.8.
• Chris Greenhalgh did an overview of how things link together for Taverna (0.1beta8 Feb04)
• http://www.biomedcentral.com/1471-2105/5/25 is a paper that includes Taverna in its comparison of approaches for composing bioinformatics web services.

A User Guide is available for Release 1.0. Development version - http://www.ebi.ac.uk/~tmo/mygrid/releases/latest/docs.word.html

The old User Guide for beta0.6 (TavernaUserGuide-beta0.6-03Mar.doc (3 Mar 2004) [2.7Mb]) is superceeded. This old version is quite long, only covers the introductory issues, and there were significant problems with the placing of figures when trying to convert it to pdf (see attached example.)

Mail list archives

With its growing user and developer community many issues are discussed on the Taverna mail lists. The archives are available from sourceforge. If you are using Taverna then you will find lots of useful information on the lists, particularly discussion of possible future improvements

• Taverna-users For users of the taverna workflow system
• Taverna-hackers Hacking on taverna, read all about it here - the latest news

Presentations

The myGrid Tutorials and Workshops page includes several Taverna presentations From PresentationStore :

• Tom's Taverna in 2006 EBI Industrial Link presentation, 2006.
• Project team User Day Tutorial February 2006
• Carole's Taverna OMII presentation at GGF 16, February 2006.
• Tom Oinn's myGrid and Taverna for E-BioSci/ORIEL workshop, October 2004
• Carole Goble's GGF12workflowrgmyGrid.ppt GGF12 workflow research group myGrid talk, 22nd September 2004, Brussels
• Chris Wroe's Workflow presentation for biologists and bioinformaticians. July 2004
• Matthew Addis presentation for IXI project on myGrid Workflow, July 2004
• Carole Goble's myGrid tutorial for Grid summer school. July 2004
• Tom Oinn's GGF 10 and LinkUp presentation March 2004, also given by Martin Senger at Biomedicum 2004
• Matthew Addis' myGrid workflow talk at the NeSC Workflow Workshop Dec 2003
• Darren Marvin's workflow presentation at UK eScience AllHands2003

Notes

Extensibility

TavernaExtensibility separate Twiki page for developer information on extending Taverna

Includes notes in the Scufl processor types, the SPI interfaces, running workflows without a gui

Viewing Results

TavernaMimeTypes extended with SeqVista integration [TMO taverna-hackers 16 Oct 2004]

Standalone DataThing viewer available - dataviewer.[bat|sh] in workbench directory. This can be used to view inputs or outputs saved in XML format [TMO taverna-hackers 15 Oct 2004]

Workflow Diagram Problems

Taverna uses an external application, dot - part of GraphViz from At&T, to generate the workflow diagrams. If the dot program is not found then you will be able to load and run workflows, but not to see the diagram. (You will get an error java.io.IOException: dot not found ). Windows users do not normally need to worry as the dot program is included in the Taverna download and the runme.bat file updates the path as required.

There have been some problems, particularly on Linux and Macs, with people not getting the workflow diagrams.

• Error 4 has been reported on beta0.10. This is caused by an array index out of bounds error on the png reader. There is a fix in later releases. The configuration file now allows users to select an alternative, e.g. gif or jpeg, rather than the default png. [taverna-users 15 Jul 2004] See also the Taverna 1.0 user manual http://taverna.sourceforge.net/manual/docs.word.html#_Toc94086048
• I/O error reading PNG header is probably cased by another application called dot.exe on the system path. This is fixed in beta12 [taverna-users 6 Nov 2004]

It is recommended that users having any problems with the diagram ensure that they are on the expected Java 1.4.2.

Building from source

To build from source you will need:

• Maven 2 - we are no longer recommending building with ant.
• Doxygen - not absolutely essential, used in generating the documentation. If ant fails on this, it will already have build the Taverna workbench (or you could edit the build.xml file).

and of course a Java compiler (i.e. the Java 1.4.2 system development kit not just the runtime environment http://java.sun.com/ )

For information on obtaining the latest sources use the CVS link on the sourceforge site

Check out instructions on how to download and compile Taverna on the BuildingTaverna page.

If you're running on windows, this is enough to make it all work. If you are running on some other system you will, in addition, need to install graphviz. This can be done by downloading it from :

http://www.research.att.com/sw/tools/graphviz/download.html

So, to reiterate - the 'workbench' directory is only used as part of the build system, it is not the workbench you run. [Tom Oinn April 2004]

Installation on Macs

Feb 2004 A version of Taverna for Mac OSx is now available from sourceforge. See Tom's email Better tested OSx build available on Taverna-users 15 Feb 2005 for details

Getting the appropriate installation on Macs so that Taverna recognises the external dot application from graphviz has caused some headaches. Some solutions work for one Mac but not others. Taverna-users 10 Dec 2004

Sources of graphviz include:

• http://www.pixelglow.com/graphviz/

Chris Dwan reported that it worked after installing the jpeg libraries and setting the workflow diagram image type in mygrid.properties to jpg. Making sure that dot is on the PATH set of Taverna's runme.sh files is also important.

Doc style web services including NCBI services

There have been some problems reported using doc style web services [remko de knikker Jan 2004]. The problem was due to the way that the Axis invocation framwork parsed the XML documents. However, it should be noted that some DOC services work fine.

There have been some problems using the NCBI document-based services. These problems were fixed for the Taverna 1.0 release. See Taverna-users mail list topic EUtils Examples 19 Jan 2005

In Jan 2005 latest Taverna CVS moved to using a not-yet released version of Axis which included essential bug fixes. The Taverna-users mail list has the latest details e.g. topic NLM eutils, service discovery

Other Links

• Tom Oinn's page http://www.ebi.ac.uk/~tmo/mygrid/
• myGrid project web site http://www.mygrid.org.uk/
  • workflow
• Freefluo http://freefluo.sourceforge.net/
• WebCVS example workflows
  • http://taverna.sourceforge.net/webservices/workflows/ Example Dec 2003

• TavernaWorkflowDiagrams
• TavernaMimeTypes

-- Main.MarkGreenwood - 05 Apr 2006

• scufl_draft2.doc: Introduction note for Link-up meeting