Requirements for Taverna 3 data bundles.
For history, see previous specification idea from 2011: 2011-02 Data bundle
Requirements
Taverna Data bundles:
- structured folder layout which form a distributable ZIP archive
- purpose of containing data for or from workflow runs, ie. replace old "baclava"
- should be possible to contain/embed workflow bundle
- should contain provenance of workflow run (ie. be a Workflow Run Bundle)
- should be possible to embed intermediate values
- based on a proposed "RO bundle" (See below).
Proposed products
With estimates in 'ideal developer-time' (Stian).
RO bundle specification: (2d Done)
- a ZIP archive of a Wf4Ever Research Object
- how the ZIP archive is structured, ie. what is the manifest, folders and metadata required.
- Described according to RO model; so supports annotations, provenance etc.
- Modulus of SCUFL2 Workflow Bundle specification.
- Should be written like a lightweight specification (ReSpec HTML?)
- Done: http://purl.org/wf4ever/ro-bundle
RO bundle Java API: (3d 5d)
- based on JSON-LD and Java 7 Filesystem providers
- Browsing and adding of resources in folders
Updated Data Bundle specification: (4d)
- based on the RO bundle
- To be used by developers in BioVel implementing this in Ruby
- Defines similar structures as the old 2011-02 Data bundle document:
- Individual data values (String, bytes, URLs)
- Lists
- Error documents
- Should be written like a lightweight specification, not wiki page
Data Bundle Java API (4d 3d):
- based on RO Bundle Java API and Java 7 FS provider. Just a DataBundle helper, similar to java.nio.Files? (isErrorDocument(), openWorkflowBundle() etc)
- high level exposure of Taverna data bundle structures for creation and browsing
- Might or might not replace or complement the Taverna 3 Data proposal
- Draft: https://github.com/myGrid/databundles
The work of development of API vs. spec is naturally interlinked, and developing the Data Bundle might require fixes to RO Bundle.
Estimate and schedule
Total wild estimate: -13d
Adjusted estimate (2013-04-29): 12d
- Start: 2013-04-30
- Finish: 2013-05-30
Labels