ncbo / bioportal-project

Serves to consolidate (in Zenhub) all public issues in BioPortal
BSD 2-Clause "Simplified" License
7 stars 5 forks source link

describe ontologies with schema.org #93

Open graybeal opened 6 years ago

graybeal commented 6 years ago

Create a dataset description for ontologies in BioPortal, include on ontologies summary page.

(Assigning to Josef to come up with an appropriate example of a schema.org block for one of our ontologies.)

jonquet commented 6 years ago

Within our work with @antool we have isolated the Schema.org properties that appears most relevant for ontologies. They are available here: https://github.com/agroportal/documentation/tree/master/metadata (search for schema).

They have been "aligned" to other metadata vocabularies properties and implemented in AgroPortal (it might not be the schema.org property that has been selected in the new AgroPortal metadata model: https://link.springer.com/article/10.1007%2Fs13740-018-0091-5 )

14 of them will find equivalent in the current NCBO BioPortal metadata model (i.e., you can express 14 properties of the current BP's model with schema.org properties).

Available to comment more.

johardi commented 6 years ago

This is a rough sketch of BioPortal to Schema.org label/property mapping:

{
  "@context" : "http://schema.org",
  "@type": "Dataset",
  "@id": Ontology.PURL,
  "identifier": Ontology.PURL,
  "name": Ontology.Name,
  "description": Ontology.Description,
  "alternateName": Ontology.Acronym,
  "license": Ontology.License_Information,
  "creator": {
    "@type": "Organization",
    "contactPoint": {
      "@type": "ContactPoint",
      "contactType": "Customer Service",
      "name": Contact.Name,
      "email": Contact.Email,
      "telephone": Contact.Telephone
    }
  },
  "sourceOrganization": {
    "@type": "Organization",
    "name": Contact.Groups
  },
  "distribution": {
    "@type": "DataDownload",
    "contentUrl": Submissions.Downloads.Link,
    "encodingFormat": Submissions.Downloads.Format,
    "version": Submissions.Version,
    "datePublished": Submissions.Released,
    "uploadDate": Submissions.Uploaded
  },
  "subjectOf": {
    "@type": "CreativeWork",
    "name": Projects_Using_This_Ontology.Name,
    "url": Projects_Using_This_Ontology.URL
  }
}
jonquet commented 6 years ago

Hello,
We have been doing similar things in our implementation in AgroPortal. This would be good our mappings to map !

---- from our JoDS paper : We have identified the schema:Dataset type as the closest one to describe ontologies. Schema.org is very rich to describe schema:Dataset (including properties inherited of schema:CreativeWork and schema:Thing), we have identified 41 relevant properties:

schema:distribution, schema:includedInDataCatalog, schema:spatial, schema:about, schema:alternativeHeadline, schema:associatedMedia, schema:audience, schema:author, schema:award, schema:comments, schema:contributor, schema:copyrightHolder, schema:creator, schema:dateCreated, schema:dateModified, schema:datePublished, schema:workExample, schema:fileFormat, schema:hasPart, schema:isPartOf, schema:inLanguage, schema:isBasedOn, schema:keywords, schema:license, schema:mainEntity, schema:publisher, schema:publishingPrinciples, schema:review, schema:schemaVersion, schema:sourceOrganization, schema:translator, schema:version, schema:alternateName, schema:description, schema:image, schema:mainEntityOfPage, schema:citation, schema:name, schema:url, schema:translationOfWork, schema:translation

For instance, the property schema:includedInDataCatalog may be used to store the fact that an ontology is hosted in different ontology libraries.


These are the mappings we have done with the BioPortal metadata model:

jonquet commented 6 years ago

omv:name > schema:name bpm:viewOf > schema:isPartOf bpm:reviews > schema:review bpm:notes > schema:comments bpm:ui > schema:url bpm:hasOntologyLanguage > schema:fileFormat bpm:homepage > schema:mainEntityOfPage bpm:publication > schema:citation bpm:naturalLanguage > schema:inLanguage bpm:version > schema:version bpm:creationDate > schema:datePublished bpm:description > schema:description bpm: released > schema:dateCreated bpm:download_format=rdf > schema:distribution

All the other ones selected in Schema.org did not found a match in BioPortal metadata model. We have in our new model either selected by default the schema.orf one or if another vocabulary was more relevant (DC, DCAT...) we choose another vocab.

johardi commented 6 years ago

Thanks, @jonquet! I'll take into account your mapping above and discuss it with @jvendetti