Taverna 2.3 and later has support for exporting the provenance of a workflow run in two different formats. The provenance contains details about the execution of the workflow, in particular it can show the origin of each piece of output data through the intermediate values and service executions.
 | This page describes the deprecated provenance export that is built-in to Taverna 2.3 and 2.4, generating traces in the OPM and Janus formats. Rather than using this mechanism, users are recommended to install and use the updated Taverna-PROV plugin which produces PROV-O traces and includes the data values. The Taverna-PROV traces are more complete, more are "correct" and address many of the known issues in OPM/Janus. |
Exporting provenance from the Workbench
From the Taverna workbench, go to the Result perspective, select a finished workflow run (which has been run with provenance enabled), make sure the workflow output ports are selected (not a particular service) and click the Save All button.
This brings up the Workflow run data saver dialogue:
Click Save as OPM (experimental) or Save as Janus (experimental) depending on your preferred format. For OPM and Janus the tickboxes of input/output ports are ignored (all ports and services are included).
Select a folder and filename - the extension is .rdf for both formats.
Exporting provenance from the Command Line
Using the executeworkflow command line tool, use the option -opm FILE and/or -janus together with the database option -embedded or -clientserver to save the provenance traces. The -provenance option is implied by either -opm or -janus.
You can customize the provenance filenames:
Open Provenance Model - OPM
The Open Provenance Model is a standard specification and format for sharing provenance information across multiple tool. Taverna supports exporting provenance in a subset of OPM to a RDF/XML format according to the OPM OWL ontology.
Example provenance-opm.rdf - from running lineage-example.t2flow:
OPM example
| These URIs, in particular under the http://ns.taverna.org.uk/ domain, are only used as unique identifiers, the actual data and workflow definitions are not available by following those links. The links are created to be compatible with the SCUFL2 identifiers. |
In the example above, the data item (Artifact) identified as
http://ns.taverna.org.uk/2011/data/6bf06207-8c57-46e8-88de-2492e3814ca4/ref/9849dbf6-6280-4dbb-b8a2-5100d8c56e64
was Used by
http://ns.taverna.org.uk/2010/workflow/e024ea01-89fd-4f93-b6ae-ad6da4a6df08/processor/P1/
.. that is processor (ie service) "P1" in the workflow with the identifier "e024ea01-89fd-4f93-b6ae-ad6da4a6df08".
This data item was Generated by the process
http://ns.taverna.org.uk/2010/workflow/e024ea01-89fd-4f93-b6ae-ad6da4a6df08/processor/P0
By following this trace you are able to see which intermediate values contributed to the calculations of a particular output value.
Converted to the N3/Turtle format using a tool like cwm it might be easier to follow the file manually:
From provenance-opm.n3:
Known issues with the OPM support
OPM export is currently experimental. Please contact myGrid for any comments and bug reports.
Please see the list of known provenance issues.
Janus
Janus is an ontology modelled at myGrid by Paolo Missier to be closer to Taverna's internal execution model, and allow attaching third-party annotations on services.
Taverna 2.3 and later supports exporting provenance as an RDF/XML document using the Janus ontology.
See:
- Janus: from Workflows to Semantic Provenance and Linked Open Data. Missier, P.; Sahoo, S. S.; Zhao, J.; Sheth, A.; and Goble, C. 2010. In Procs. IPAW 2010, Troy, NY.
- Janus Presentation
- Janus Ontology - http://purl.org/net/taverna/janus
Example from provenance-janus.rdf from running lineage-example.t2flow::
Reformatted into N3/Turtle (from provenance-opm.n3):
Janus includes an outline of the workflow structure in addition to the provenance of the workflow run. For instance we see that the processor ../P0/ has the parameter .../out/Y and is part of the workflow e024ea01-89fd-4f93-b6ae-ad6da4a6df08, which has been executed as run 6bf06207-8c57-46e8-88de-2492e3814ca4. The port has had a value binding to http://ns.taverna.org.uk/2011/data/6bf06207-8c57-46e8-88de-2492e3814ca4/ref/9849dbf6-6280-4dbb-b8a2-5100d8c56e64 - which we can see later has the textual value "a1" (noted as an rdfs:comment)
Known issues in Janus
Janus export and the Janus ontology itself is currently experimental. Please contact myGrid for any comments and bug reports.
Please see the list of known provenance issues.
Labels