The first version of the Taverna 2 (T2) command line application should accommodate as much of the behavour as the Taverna 1 (T2) executeworkflow.sh script, and it should remain familiar.
Details of the T1 command line tool can be found at: http://www.mygrid.org.uk/usermanual1.7/executeworkflow.html and is summarised below:
The workflow filename can either be a URL (file:// or http://) or a file path (/tmp/myworkflow.sh).
The XML file format used is the standard Baclava format used by both T1 and T2 for importing and exporting data, which can be viewed graphically using teh Data Viewer application. This needs to be bundled with T2 together with the command line script if it is not done so already.
In the first instance the command line tool will run In Memory and with Provenance switched off. Further command line options can be provided later to support defining a database to connect to to run backed with a data source and with Provenance.
Much of the existing executeworkflow.sh source-code can be re-used, but changes will be required to handle:
- Registering and Resolving data with the Reference Manager, and translation to and from the Baclava XML format.
- Initialising and configuring the Plugin Manager & Raven as this has changed for T2 since T1. The configured modules should remain external to the application code so that it can easily be reconfigured for newer components.