recordr
recordr copied to clipboard
Include provone:Execution attributes in lineage info
When recordr::publishRun() is called to publish a DataPackage, a DataONE compliant resource map is created for the package to be uploaded. The resmap includes the provenance relationships for the provone:Execution, but currently no information describing the Execution is sent with the DataPackage.
Information collected for a Recordr run can be included in the resmap as properties linked to the provone:Execution. The following items are recorded for a run:
| Recordr run property (R Execution slot name) | Description | Proposed RDF representation | Current RDF representation |
|---|---|---|---|
| executionId | execution uuid | provone:Execution | |
| metadataId | metadata uuid | ||
| tag | text string associated with this run | ||
| datapackageId | DataONE package uuid (resmap uuid) | ||
| user | local user name | prov:wasAssociatedWith -> prov:Agent (foaf:Account, rdf:type) | |
| subject | DataONE user id | ||
| hostId | DataONE host id | ||
| startTime | Execution start time | prov:startedAtTime | |
| operatingSystem | local OS | ||
| runtime | local execution env (e.g. "R 3.5.5") | ||
| softwareApplication | local application (e.g. "myProg.R") | provone:qualifiedAssociation -> prov:Association -> provone:hadPlan | |
| moduleDependencies | system modules used during run | ||
| endTime | execution ending time | prov:endedAtTime | |
| errorMessage | execution error message | ||
| publishTime | time DataPackage was uploaded | ||
| console | was this run performed with 'startRecord()/endRecord()' | ||
| publishNodeId | DataONE node id DataPackage was published to | ||
| seq | local unique identifier for a run |
Which of these attributes that does not have a current or proposed RDF representation do we want to create one for? If no current W3C PROV or DataONE PROV relationships or entities exist for these, should they be added?
@mbjones what are your thoughts on this?