Top

16.12. Virtuoso Faceted Web Service

The Virtuoso Faceted web service is a general purpose RDF query facility for facet based browsing. It takes an XML description of the view desired and generates the reply as an XML tree containing the requested data. The user agent or a local web page can use XSLT for rendering this for the end user. The selection of facets and values is represented as an XML tree. The rationale for this is the fact that such a representation is easier to process in an application than the SPARQL source text or a parse tree of SPARQL and more compactly captures the specific subset of SPARQL needed for faceted browsing. The web service returns the SPARQL source text also, thus this can serve as a basis for and-crafted queries.

The top element of the tree is <query>, it must be in namespace "http://openlinksw.com/services/facets/1.0/".

This has the following attributes:

  • graph="graph_iri" - default is search in all graphs but system defaults may override this

  • timeout="no_of_msec" - default is no timeout, but system defaults may override this

  • inference="name" where name is a name of an inference context declared with rdfs_rule_set.

  • same-as="boolean" - If "boolean" is "yes", then owl:sameAs links will be considered in the query evaluation.

The result is a tree of the form:

<facets xmlns="http://openlinksw.com/services/facets/1.0/">
<result><row><column datatype="..." shortform="..." xml:lang="..">...</column></row></result>
<time>msecs</time>
<complete>yes or no</complete>
<db-activity>resource use string</db-activity>
<sparql>sparql statement text</sparql>
</facets>

By convention, the first column is the subject selected by the view element, typically a URI, the second a label of the URI and the third, if present, is either a count or a search summary.

The first column's text child is the text form of the value. The column element has the following attributes qualifying this further:

  • datatype - The xsd type of the value. If this is a URI, the datatype is "uri"

  • shortform - If the value is a URI, this is an abbreviated form where known namespaces are replaced with their prefixes and very long URI's are truncated preserving start and end.

  • xml:lang - if the value is a language tagged string, this is the language

The query has the top level element <query>. The child elements of this represent conditions pertaining to a single subject. A join is expressed with the property or property-of element. This has in turn children which state conditions on a property of the first subject. property and property-of elements can be nested to an arbitrary depth and many can occur inside one containing element. In this way, tree-shaped structures of joins can be expressed.

Expressing more complex relationships, such as intermediate grouping, subqueries, arithmetic or such requires writing the query in SPARQL. The XML format is a shorthand for easy automatic composition of queries needed for showing facets, not a replacement for SPARQL.

A facet query contains a single view element. This specifies which subject of the joined subjects is shown. Its attributes specify the manner of viewing, e.g. list of distinct values, distinct values with occurrence counts, properties or classes of the selected subjects etc.

The top query element or any property or property-of element can have the following types of children:

<text property="iri">text pattern</text>

The subject has an O that matches the text pattern. If property is given, the text pattern must occur in a value of this property. If not specified, any property will do. The value "none" for property is the same as not specifying a property. This is restricted to occurring directly under the top level query element.

<class iri="iri" inference="ctx_name" />

The S must be an instance of this class. If inference is specified then option (input:inference "ctx_name" is added and applies to this pattern alone.

<property iri="iri" same_as="yes" inference="ctx_name">

The child elements of this are conditions that apply to the value of this property of the S that is in scope in the enclosing <query> or <property> element. If same_as is present, then option (input:same-as "yes") is added to the triple pattern which specifies this property. If inference is present, then option (input:inference "ctx_name") is added to the triple pattern for the property.

<property-of iri="iri" same_as="yes" inference="ctx_name" >

The child elements of this are conditions that apply to an S which has property "iri" whose object is the S in scope in the enclosing <query> or <property> element. The options are otherwise the same as with property.

<value datatype="type" xml:lang="lng" op="= | < | > | >= | <=">value </value>

When this occurs inside <property> or <property-of> this means that the property in scope has the specified relation to the value. type and language can be used for XML typed or language tagged literals. The "uri" type means that the value is a qualified name of a URI. If this occurs directly under the <query> element, this means that the query starts with a fixed subject. If this is so, then there must be property or propertyof elements or the view element must specify properties or classes, list is not allowed as a view type. This is so because the query must have at least one triple pattern.

<view type="view" limit="n" offset="n" >

This may occur once inside a <query> element but may occur either at top level or inside property or property-of elements. This specifies what which subject is presented in the result set.

The type can be:

  • "properties" 

    SPARQL
SELECT ?p count (*) { ?this_s ?p ?any_o ...}
GROUP BY ?p
ORDER BY DESC 2
LIMIT l OFFSET 0

  • "properties-in" 

    SPARQL
SELECT ?p count (*) { ?any_s ?p ?this_s ... }
GROUP BY ?p
ORDER BY DESC 2
LIMIT L OFFSET 0

  • "classes" 

    SPARQL
SELECT ?c count (*)
WHERE { ?xx a ?c ... }
GROUP BY ?c
ORDER BY DESC 2
LIMIT l OFFSET 0

  • "text" 

    SPARQL
SELECT DISTINCT ?s (bif:search_excerpt (sql:search_terms (""pattern"), ?o)) ...
LIMIT l OFFSET 0

  • "list" 

    SPARQL
SELECT DISTINCT ?s long::sql:fct_label (?s) ...
LIMIT l OFFSET 0

  • "list-count" 

    SPARQL
SELECT ?s COUNT (*) ....
GROUP BY ?s
ORDER BY DESC 2

  • "alphabet" 

    SPARQL
SELECT (sql:subseq (?s, 0, 1)) count (*) ...
GROUP BY (sql:subseq (?s, 0, 1))
ORDER BY 1

  • "geo" 

    SPARQL
SELECT DISTINCT ?lat ?long ?s
WHERE ?s geo:lat ?lat . ?s geo:long ?long . ... }

  • "years" 

    SPARQL
SELECT sql::year (?s) count (*) ...
GROUP BY (bif:year (?s))
ORDER BY 1
OFFSET 0 LIMIT l

  • "months" 

    SPARQL
SELECT sql::round_month (?s) count (*) ...
GROUP BY (sql:round_month (?s))
ORDER BY 1 OFFSET 0 LIMIT l

  • "weeks" 

    SPARQL
SELECT sql::round_week (?s) COUNT (*) ...
GROUP BY (sql:round_week (?s))
ORDER BY 1 OFFSET 0 LIMIT l

  • "describe" 

    SPARQL describe ?s ... OFFSET 0 LIMIT l

16.12.1. Customizing

The following types of customization will be generally useful:

  • Resource accounting and limitations, managing access and login

  • Localization, choice of labels shown with class/property/instance URI's

  • Adding types of views, for example timelines, map or business graphics

  • Controlling navigation, for example choosing what type of view is initially presented when opening a given property.

  • Page layout, captions, help texts, etc.

The source code is divided in two SQL files and a number of XSLT sheets. The file facet.sql has the code for the web service. The facet_view.sql file contains the procedures for the sample HTML interface.

Prev  Up  Next
16.11.7. Examples  Home  16.12.2. Examples