AgriculturalSemantics
Create subset exports
@celineaubert is beginning to create subsets using the in_subset AP. We should start exporting these for easy consumption by developers. Our recipe from ENVO (which will be a bit different from what we need here) is summarised below:
- First, we create a "simple" version of the ontology as a sort of subset in itself using owltools. The Makefile drives this here. We should try to avoid the need to create an OBO version (here) and make the point below only rely on the simplified OWL file. @cmungall is that possible?
- With the basic file (currently an OBO file in ENVO), we can use owltools' subset extraction to pull out named subsets. See these lines
@marieALaporte, perhaps you could try to implement this?
@marieALaporte points out that there may be no need for OWL subsets at this stage.We agree that generating flat files of all classes in each subset may have more immediate use for the developers of the fieldbook.
Thus, (correct me if I'm wrong @marieALaporte) we'll fuse the idea of subsets and reports (#11) and export flat files containing the IDs, labels, and definitions of classes in subsets.
Since the "subset reports" will be SPARQL-driven, perhaps the other reports from #11 can be done en passant:
Over in ENVO, we create simple tabular files for non-owl-inclined software developers to work with. These are dubbed "reports" and are simplistic exports of the ontology content. As this is SPARQL-driven (via arq and jena), perhaps you'd like to take a crack at impementing (and improving?) this @marieALaporte?
Our sparql files produces reports. The Makefile drives this in these lines
If you do want subsets, then another method is to keep a list of URIs in a file or spreadsheet and then MIREOT the subset. YMMV.
For SPARQL, we have an open ticket to do this from Robot https://github.com/ontodev/robot/issues/24 But for now arq is probably easiest. One extra moving piece, mildly annoying.
