Pivot Collections (Part 5)

Part 1: Introduction

Part 2: SparqlCxml

Part 3: SparqlCxml Deep Zoom Collections

Part 4: HtmlPivotViewer

Part 5: Importing CXML

Importing CXML

Part 6: Facet Pivot Bridge

Part 7: DETs : Persisting SPARQL Query Results to DAV

Part 8: Frequently Asked Questions (FAQs)

Part 9: Glossary

Importing CXML

As well as being able to generate CXML from RDF, Virtuoso provides the facility to do the reverse and generate RDF from imported CXML. The sparql_cxml VAD includes the procedure:

DB.DBA.CXML_TO_RDF_IMPORT_URL (in url varchar)

Parameters:

url : the URL of the CXML file to import

Usage examples:

DB.DBA.CXML_TO_RDF_IMPORT_URL ('http://content.getpivot.com/Collections/uspresidents/uspresidents.cxml');
DB.DBA.CXML_TO_RDF_IMPORT_URL ('http://www.bluesware.ch/parlament/collection/Parlament.cxml');
DB.DBA.CXML_TO_RDF_IMPORT_URL ('http://pivot.lobsterpot.com.au/Collections/WindowsPhone7Pivot.cxml');
DB.DBA.CXML_TO_RDF_IMPORT_URL ('http://pivot.lobsterpot.com.au/collections/PASSChapters.cxml');
DB.DBA.CXML_TO_RDF_IMPORT_URL ('http://www.wssdemo.com/livepivot/collections/sharepoint/sharepoint.cxml');

The given CXML file is imported into a graph of the same name and can be queried like any other. e.g.

DESCRIBE ?s FROM <http://content.getpivot.com/Collections/uspresidents/uspresidents.cxml> where { ?s ?p ?o }

SELECT * FROM  <http://content.getpivot.com/Collections/uspresidents/uspresidents.cxml> WHERE {?s ?p ?o . ?s virtcxml:FacetImage ?image . }

define input:inference "virtrdf-label"
DESCRIBE  ?s from <http://pivot.lobsterpot.com.au/collections/PASSChapters.cxml> where { ?s ?p ?o }

Deep Zoom images associated with the collection are also imported into the DZC_xxx tables used to serve DZIs when generating CXML (see VirtSparqlCxmlDzcs). Imported images are assigned to the virtcxml:FacetImage predicate. This can be used directly with the ?image reserved query variable in SELECT statements used for CXML generation. The second DESCRIBE example above exploits inferencing to use the imported images as images for a new dynamic collection. The virtrdf-label inference context includes an entry:

virtcxml:FacetImage owl:equivalentProperty foaf:depiction

Graph virtcxml: in turn includes an entry:

foaf:depiction
  a virtcxml:FacetCategoryPattern ;
  virtcxml:FacetType "Item/@virt:Image" ;
  virtcxml:FacetDescription "A facet for absolute URI of depiction of a collection item" .

with the result that imported images associated with virtcxml:FacetImage automatically have a virtcxml:FacetType of "Item/@virt:Image" and are seen by DESCRIBE as candidate images for the dynamic collection.

Generated Subject URIs

Each item in the collection receives its own subject URI. The subject URI takes the form: <{collection-URI}#{item-ID}>.

e.g. Item #2565 in collection http://www.bluesware.ch/parlament/collection/Parlament.cxml

<Collection Name="Das Schweizer Parlament 2010 (aktualisiert am 6. Dez.2010)" SchemaVersion="1.0">
    <FacetCategories>
        <FacetCategory Name="Funktion" Type="String" p:IsFilterVisible="true" p:IsMetaDataVisible="true" p:IsWordWheelVisible="true"/>
        . . .
    </FacetCategories>
    <Items ImgBase="p_deepzoom\p1.dzc">
        <Item Id="2565" Name="Abate Fabio" Img="#0" Href="http://www.parlament.ch/d/suche/Seiten/biografie.aspx?biografie_id=801">
        . . .

is assigned subject URI <http://www.bluesware.ch/parlament/collection/Parlament.cxml#2565>

Generated Predicates

Predicates for Collection and Item Attributes

Item attributes Href, Description etc are mapped to predicates named virtcxml:Facet{attribute-name}, where:

prefix virtcxml: <http://www.openlinksw.com/schemas/virtcxml#>

CXML Entity	Attribute	Mapped to predicate
Item	Href	virtcxml:FacetHref
Item	Description	virtcxml:FacetDescription
Item	Image	virtcxml:FacetImage
Item	Related	virtcxml:FacetRelated
Collection	Copyright	virtcxml:FacetCopyright

Rather than just holding an image number, predicate virtcxml:FacetImage holds the full path of the corresponding Deep Zoom image, for example http://www.bluesware.ch/parlament/collection/p_deepzoom/2023.dzi. The image number in the .cxml file is dereferenced through the collection's .dzc file, identified by the Items/@ImgBase attribute, to obtain the DZI path.

Predicates for Item Facets

A predicate is generated for each facet category. The predicate URI takes the form virtcxml:Custom-{facet-category-name} e.g. A collection containing the following facet categories:

<Collection Name="Das Schweizer Parlament 2010 (aktualisiert am 6. Dez.2010)" SchemaVersion="1.0">
<FacetCategories>
<FacetCategory Name="Alter" Type="Number" p:IsFilterVisible="true" p:IsMetaDataVisible="true" p:IsWordWheelVisible="false"/>
<FacetCategory Name="Amtsjahre" Type="Number" p:IsFilterVisible="true" p:IsMetaDataVisible="true" p:IsWordWheelVisible="false"/>
<FacetCategory Name="Funktion" Type="String" p:IsFilterVisible="true" p:IsMetaDataVisible="true" p:IsWordWheelVisible="true"/>
<FacetCategory Name="Geburtsdatum" Type="DateTime" Format="dd. MMM. yyyy" p:IsFilterVisible="true" p:IsMetaDataVisible="true" p:IsWordWheelVisible="false"/>
<FacetCategory Name="Geschlecht" Type="String" p:IsFilterVisible="true" p:IsMetaDataVisible="true" p:IsWordWheelVisible="true"/>
<FacetCategory Name="Kanton" Type="String" p:IsFilterVisible="true" p:IsMetaDataVisible="true" p:IsWordWheelVisible="true"/>
<FacetCategory Name="Kinder" Type="Number" p:IsFilterVisible="true" p:IsMetaDataVisible="true" p:IsWordWheelVisible="false"/>
<FacetCategory Name="Parlamentsmandate" Type="Number" p:IsFilterVisible="true" p:IsMetaDataVisible="true" p:IsWordWheelVisible="false"/>
<FacetCategory Name="Partei" Type="String" p:IsFilterVisible="true" p:IsMetaDataVisible="true" p:IsWordWheelVisible="true"/>
<FacetCategory Name="Vorname" Type="String" p:IsFilterVisible="true" p:IsMetaDataVisible="true" p:IsWordWheelVisible="true"/>
</FacetCategories>
...

when imported, generates the predicates:

http://www.openlinksw.com/schemas/virtcxml#Custom-Alter
http://www.openlinksw.com/schemas/virtcxml#Custom-Amtsjahre
http://www.openlinksw.com/schemas/virtcxml#Custom-Funktion
http://www.openlinksw.com/schemas/virtcxml#Custom-Geburtsdatum
http://www.openlinksw.com/schemas/virtcxml#Custom-Geschlecht
http://www.openlinksw.com/schemas/virtcxml#Custom-Kanton
http://www.openlinksw.com/schemas/virtcxml#Custom-Kinder
http://www.openlinksw.com/schemas/virtcxml#Custom-Parlamentsmandate
http://www.openlinksw.com/schemas/virtcxml#Custom-Partei
http://www.openlinksw.com/schemas/virtcxml#Custom-Vorname

Converting Generated Predicates to Common Predicates

Following a CXML import, some of the generated predicates can be converted to more popular predicates using the Virtuoso/PL procedure DB.DBA.CXML_FACETS_TO_COMMON_PREDS. The affected predicates are detailed below:

Generated Predicate	Converted To
virtcxml:FacetName	rdfs:label
virtcxml:FacetDescription	dc:description
virtcxml:FacetHref	sioc:links_to
virtcxml:FacetImage	foaf:depiction

Usage Example:

sparql clear graph <http://qq/1>;
sparql insert in <http://qq/1> { ?s ?p ?o } from <http://content.getpivot.com/Collections/uspresidents/uspresidents.cxml> where {?s ?p ?o};
DB.DBA.CXML_FACETS_TO_COMMON_PREDS ('http://qq/1');