What is indexed, and what SHOULD be indexed

[markwilkinson]

Hi all!

After today's Pillar 2 general meeting, I was exploring the idea from Chris that we could automate the deployment of FDPs for Biotools. that got me thinking... the way this should be modeled is, perhaps, that each Biotool becomes a "catalog" entry within a single FDP deployment, or that each BioTool becomes a 'dataset' entry within a single catalog 'BioTools'

However... I believe that the VP Index will only capture the top-level FDP itself (is that correct?) so onboarding these resources in this way will leave then not visibly onboarded.

Is it perhaps more sensible that the catalog level is what is indexed? Or even the dataset level?

I would love to have this conversation ASAP to get these things onboarded before the end of June...

Mark

[luizbonino]

Hi Mark, no, the index does not only capture the top-level FDP metadata, it captures catalogs, datasets, biobanks, etc. This is the index. The VP Portal, however, only gets the entities (catalogs, datasets, etc.) that have either VPDiscoverable or VPQueryable as either its theme or keyword. The SPARQL that is used to select the content to be used by the portal is (example for VPDiscoverable):

SELECT DISTINCT ?resource ?resource_name ?resource_type ?created_time ?updated_time  ?home_page ?theme ?resource_description ?resource_logo ?keyword {

    ?resource a ?resource_type;
              rdfs:label ?resource_name;
              fdp-o:metadataIssued ?created_time;
              fdp-o:metadataModified ?updated_time.
    {?resource dcat:theme ?theme, ejprd:VPDiscoverable.} UNION {?resource dcat:keyword ?keyword,?s FILTER(?s = str(ejprd:VPDiscoverable)).}

    FILTER (?resource_type != dcat:Resource)

    OPTIONAL {?resource dct:description ?resource_description}
    OPTIONAL {?resource dcat:landingPage ?home_page}
    OPTIONAL {?resource dct:publisher [foaf:logo ?resource_logo]}
} ORDER BY ?resource