During week 7, I modified the converter tool in Java to have all properties of the nodes (activities/entities) in the Neo4j database and also added two additional properties of runID, wfID to be able to store traces of multiple runs of different workflows. Up to this point we only had provenance traces in the database. This week I modified the code to be able to import the workflow components (modules and the connection between them) into the database. This change makes PBase one step more closer to DPROV model. Now we can answer queries that take both workflow specifications and provenance traces into account, for example, “which modules didn’t work (some of their activities didn’t execute to completion”, …. Based on this, I added some more provenance queries to our list.
For the last two days of week 8 I was on a leave from PBase project to attend BPDM workshop in Chicago. The main focus of this workshop is data mining and knowledge discovery. Next week I will be on a leave for the first half of the week to attend ACM KDD’13. I will post more about my experience of attending this events on my next week’s blog entry.
As usual, comments and feedback are welcome,