Taverna Open Development workshop, Manchester, 2014-10-30
This work is licensed under a Creative Commons Attribution 4.0 International License.
OSGi-based Java API
Constructing and running workflows
Workflow language: scufl2
Workflow input/output/run: databundle
Taverna Platform API
Taverna Engine implementation
Taverna common activities
Taverna plugin activities
Feature-complete
Used by T3 Command line and T3 Workbench
Most core activities migrated (e.g. wsdl, rest, xpath, beanshell, spreadsheet)
Next:
Migrate remaining core activities (tool, interaction)
Migrate plugin activities (helio/vamdc/astrotaverna)
Simplify 3rd-party usage (common pom incl. activities)
Documentation of platform-api
Reorganize source code/build under taverna-engine-*
Taverna Server Client as alternative execution platform
Executes Taverna 2/3 workflows
(Almost) same options as in Taverna 2.5
Workflow Data Bundle replaces Baclava
$ sh executeworkflow.sh
usage: executeworkflow [options] [workflow]
-bundle <bundle> Save outputs to a new Workflow
Run Bundle (zip).
-clientserver Connect as a client to a derby
server instance.
-cmdir <directory path> Absolute path to a directory
where Credential Manager's files
(keystore and truststore) are
located.
-cmpassword Indicate that the master password
for Credential Manager will be
provided on standard input.
-dbproperties <filename> Load a properties file to
configure the database.
-embedded Connect to an embedded Derby
database. This can prevent
mulitple invocations.
-help Display comprehensive help
information.
-inmemory Run the workflow with data stored
in-memory rather than in a
database (this is the default
option). This can give
performance inprovements, at the
cost of overall memory usage.
-inputdelimiter <inputname delimiter> Cause an inputvalue or inputfile
to be split into a list according
to the delimiter. The associated
workflow input must be expected
to receive a list.
-inputdoc <document> Load inputs from a Baclava
document.
-inputfile <inputname filename> Load the named input from file or
URL.
-inputvalue <inputname value> Directly use the value for the
named input.
-logfile <filename> The logfile to which more verbose
logging will be written to.
-outputdir <directory> Save outputs as files in
directory, default is to make a
new directory
workflowName_output.
-port <portnumber> The port that the database is
running on. If set requested to
start its own internal server,
this is the start port that will
be used.
-provenance Generate provenance information
and store it in the database.
-startdb Automatically start an internal
Derby database server.
Two beta releases (+ Apache Taverna CL beta 3)
"Compare run between 2.5 and 3.0" testing
Not yet production-ready
Next:
Complete data bundle & provenance support
Installer as in 2.5 (is install4j OK?)
Documentation: How to install plugins?
Bug fixes: https, logging, error handling, configuration
Code reorganizing: Confusing module names taverna-command-line, taverna3-commandline, taverna-commandline-common.
Execute Taverna 2/3 workflows remotely
REST/SOAP API
Security separation and workflow restrictions
Same WSDL/REST API as in Taverna Server 2.5
.. except now can also run Taverna 3 wfbundle
Executes Taverna Command Line 3
Next:
Migrate databundle support (already in 2.5)
Use Taverna Platform directly
Alternative workflow execution environments
Workflow status reports during run
Dynamic plugin installation
Design and run workflows
Look & Feel almost like 2.5
Workbench code fully based on Taverna Platform
Load/Save/Share workflow runs
Open Taverna 2/3 workflows
Feature complete for workflow design/run
Removed perspectives: myExperiment, BioCatalogue
Plugin perspective: bioswr
Next:
Installer as in 2.5
Optional plugins and editions - ask on startup
Iteration strategy & looping - simplify UI
Create nested workflow/component from selection
Standalone application to view databundles as Results?
Bugfixes: rest config, logging