Details
-
New Feature
-
Status: Closed
-
Major
-
Resolution: Fixed
-
None
-
None
Description
This will allow to access the contents of a ContentItem by LDPath [1] and support
- Access to the metadata
- Access to the Content stored in Blobs via the fn:content(String mediaType)
- Optional access to additional metadata stored in ContentParts with a type compatible to Clerezza TripleCollection
- support for "shortcuts" to easily access "semantic concpets" defined by the enhancement structure including
-
-
- Used Namespaces:
-
@prefix fise : <http://fise.iks-project.eu/ontology/> ;
@prefix ci : <http://stanbol.apache.org/ontology/enhancer/content-item-ldpath-wrapper#> ;
The "fise" namepsace represents the currently used Enhnacement Structure. The namespace "ci" does not correspond to any real ontology. It is only used internally by this LDPath wrapper for supported "shortcuts" as described below.
-
- Supported Functions and Shortcuts
- Content support
content = fn:content(
{content-resource},{media-type},{media-type2},.., {media-typeN})This returns the value of the first Blob matching any of the parsed media types. The value of {content-resource}
must be the URI of the ContentItem OR the URI of a ContentPart compatible to the Blob interface.
- extreacted TextAnnotations
ci:textAnnotation = ^fise:extractedFrom[rdf:type is fise:TextAnnotation]
- EntityAnnotations
ci:entityAnnotation = ^fise:extractedFrom[rdf:type is fise:EntityAnnotation]
- EntitySuggestions for TextAnnotations
{text-annotation}/ci:suggestion = {text-annotation}
/^dc:relation[rdf:type is fise:EntityAnnotation]/fn:sort(fise:confidence,"DESC")
- Suggested Entities for a given
{text-annotation}
/ci:suggestedEntity = {{text-annotation}/^dc:relation[rdf:type is fise:EntityAnnotation]}/fise:entity-reference
- All extracted Entities
ci:extractedEntities = ^fise:extractedFrom[rdf:type is fise:EntityAnnotation]/fise:entity-reference
or if defined via the above 'ci:entityAnnotation' shorcut
ci:extractedEntities = ci:entityAnnotation/fise:entity-reference
Example of an envisioned LDPath program used to extract knowledge of a ContentItem as of LDPath Statements
text = fn:content(.,"text/plain");
personName = (ci:textAnnoation[dc:type is dbpedia:Person]/fise:selectedText) |
(ci:textAnnoation[dc:type is dbpedia:Person]/ci:sugestion/fise:entity-label) :: xsd:string
orgName = (ci:textAnnoation[dc:type is dbpedia:Organisation]/fise:selectedText) |
(ci:textAnnoation[dc:type is dbpedia:Organisation]/ci:sugestion/fise:entity-label) :: xsd:string
placeName = (ci:textAnnoation[dc:type is dbpedia:Place]/fise:selectedText) |
(ci:textAnnoation[dc:type is dbpedia:Place]/ci:sugestion/fise:entity-label) :: xsd:string
personRef = ci:entityAnntotation[fise:entity-type] is dbpedia:Person]/fise:entity-reference :: xsd:anyURI
orgRef = ci:entityAnntotation[fise:entity-type] is dbpedia:Organisation]/fise:entity-reference :: xsd:anyURI
placeRef = ci:entityAnntotation[fise:entity-type] is dbpedia:Place]/fise:entity-reference :: xsd:anyURI
conceptRef = ci:entityAnntotation[fise:entity-type] is skos:Concpet]/fise:entity-reference :: xsd:anyURI
NOTES:
- The "shortcuts" as defined above could be also implemented as function such as
fn:textAnnotation(
{content-item},[
{dc-type},
{dc-type2},...
{dc-typeN}])
fn:suggestion(
,[
{entity-type},
{entity-type2},...
{entity-typeN}])
this would allow usages such as
personName = (fn:textAnnoation(.,dbpedia:Person)/fise:selectedText) |
(fn:textAnnoation[.,dbpedia:Person)/fn:suggestion(.)/fise:entity-label) :: xsd:string
Attachments
Issue Links
- is related to
-
STANBOL-499 Contenthub: Semantic Indexes
- Open