Closed gothub closed 7 years ago
Provenance can be asserted between two packages (say package A
and package B
by using the insertDerivation()
method and specifying an object from package A
as a source and an object
from package B
as a derivation. In this example, package B
is the current package represented by object dp
:
dp <- new("DataPackage")
dp <- insertDerivation(dp, sources="1234", derivations="5678")
where PID 1234
is from package A
and PID 5678
is from package B
.
This call to insertDerivation()
will cause a prov:wasDerivedFrom
relationship to be entered into the current package dp
.
If the program
argument is also used, then a prov:used
relationship will be added, which will link the program (and execution) from package B
to the source from package A
A user has requested that datapack support the ability to assert provenance relationships between DataPackages, e.g. DataPackage A contains source data and DataPackage B contains products derived from DataPackage A. Of course these relationships are useful after the packages have been uploaded to a data repository, so possibly a repository based tool for establishing these relationships makes more sense that an R based solution.