This week I was mostly working on the convertor (from PROV-XML to Cypher scripts) code to make some changes in the way the traces are stored in PBase repository. Here is a quick look at the current structure of PBase repository:
The blue boxes show the properties of each node. The subgraph on the right shows data and activity nodes and the edges between them (trace information), as well as, modules and the connection between them (workflow specification). The small subgraph on the left shows each workflow as a node and connects it to all of its runs.
Using this structure we can answering the queries that take both trace information and workflow specification into account (for example, “Find which modules did not execute to completion within a trace”).
Next week I am going to mostly concentrate on putting some documents for PBase together and prepare a summary of the project’s results in form of a poster.