Nice feature
However, many ontologies use numbers in IRIs, both for classes and relations. This makes the diagrams sometimes hard to read.
It would be nice to enable rendering based on annotations such as rdfs:label or skos:prefLabel.
This mode (label-based rendering) and the order of priority of annotations could be selected by the user in his/her Preferences
I don't know how it was implemented but what would be super super cool is if the graph was driven off a custom stardog graph display ontology like stardog:node, stardog:edge, stardog:nodeLabel, stardog:nodeColor, etc and then use rules to map onto it. That way the user could control how the graph is rendered by simply updating the rules. Since rules are applied at query time it wouldn't add any extra data to your model. You could imagine having a complex set of mappings to render to color, size, etc. It would be something like d3.js but with rules controlling the mapping of data into visual elements.
Relating back to the original post you could control exactly what is used for labels by writing a more complex rule for selecting the label.
This is true, and actually I already do that, for example, to translate units of measure into something which meaningful.
However it would be quite verbous if it had to be done for all classes and relationships.
What I suggested is similar to the the way it is implemented in the Protégé ontology editor.
Thanks for all of the feedback here. This is something we've been thinking about internally (we've already got an internal ticket discussing better user-preference-driven defaults that might be used in visualizations), so stay tuned for updates!
Having said that, I did want to also point out that the visualization in Studio does provide the ability to change which property is used for the display label for nodes of any given type. This may help you to have a better experience with the visualization. To use this feature, you can click on any node type in the top bar of the results area when you've got a visualization active, and then choose (on the bottom bar) which property you'd like to use as a "caption." This is easier to show than describe, so here's a contrived example showing how to do this:
I hope this helps, and thanks again for the feedback.
Jason
Apparently the edges and nodes corresponding to Data properties and their targets are not shown on the graphs. Is there a way to make them visible ?
Best regards
Hi Bernard,
That's correct -- we don't display data properties in the graph so as not to clutter the graph (and because we've heard feedback expressing a desire not to show the data properties there). You can see data properties and their targets for any given node by clicking on the node itself and looking at the bottom bar of the visualization, as shown in the image below (where I've clicked on the ":Empire" node; you can see :name "The Empire Strikes Back", etc.):
Does that help? Do you still have a need to see all of the data properties and their targets inside of the graph visualization itself?
Thanks,
Jason
OK I understand now, the principle is to read at the bottom of the window all data properties attached to the node of interest, and use the graphical display to interact (browse and navigate), basically.
However, I wonder if people would be interested to save an image in .PNG of .JPEG file, then they might be interested to have the data properties in the graph as well. By the way I note that the
'Save layout' capability stores a description of the content of the graph as .JSON, not an image. Is there a possibility to save an image as well ?
Yes, that principle is what we had in mind, glad it makes sense. There is no current capability to save natively as an image (though of course you can screenshot, fullscreen mode will help you get more/better image), but we've heard a few requests for it so it's something we plan to look into, though no specific timeline at the moment. I agree that on export it's even more important to have a way to show data properties, so we plan to explore that at the same time.
Thanks again for all the thoughts and engagement, we appreciate the feedback!
Simon