determine if Provenance has everything needed by Reporting
Make sure provenance is capturing all data needed by Reporting and check that getting things back out in a form reasonable for Reporting is possible.
E.g. test a variety of workflows to see if provenance is captured things Reporting needs. Look into actors like the Plotters that generate a product from streams of tokens, to see if the product is captured, or can be captured.
#2 Updated by ben leinfelder over 10 years ago
Derik and I were talking about how the schema could accommodate more than just the MOML file. Currently there is a column on the WORKFLOW_EXEC table has a WF_CONTENTS_ID column to point to the MOML as it existed for that run (in the DATA table).
If we do want to store all these auxiliary files in the provenance store then....we'd need to support n* files for each execution. I think it's fine to store them in the same DATA table that is being used for all the other data (and moml file).
Would need to keep track of "file type" - this is more than just "is it a PDF?" - we'd need to know if, say, an XML file was a ROML or a RIO, etc...
If we did have 2 or more Report Layouts (ROML) for a given workflow execution - we'd need a way to distinguish them from each other (perhaps a name or other metadata about the particular layout).
I'm not sure we're completely committed to storing all these artifacts in the provenance store (it seems slightly orthogonal to the intent of provenance).
#3 Updated by ben leinfelder over 10 years ago
we do have all the file artifacts being inserted.
I note the comment about provenance storing things from plotting actors...we have tokens that they might consume, and if they emitted an image then we'd have them...but those actors are very GUI and not so image-file based