path: root/docs/raptor-parsers.xml

<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.3//EN" 
               "http://www.oasis-open.org/docbook/xml/4.3/docbookx.dtd">
<chapter id="raptor-parsers">
<title>Parsers in Raptor (syntax to triples)</title>

<section id="raptor-parsers-intro">
<title>Introduction</title>

<para>This section describes the parsers that can be compiled into
Raptor and their options.  The exact parsers supported may vary
by different builds of raptor and can be queried at run-time by
use of the 
<link linkend="raptor-world-get-parser-description"><function>raptor_world_get_parser_description</function></link>
function</para>

<para>The options that may be set on parsers can also
be queried at run-time with the 
<link linkend="raptor-world-get-option-description"><function>raptor_world_get_option_description</function></link>
function.</para>

</section>


<section id="parser-grddl">
<title>GRDDL parser (name <literal>grddl</literal>)</title>
<para>A parser for the
<ulink url="http://www.w3.org/TR/2007/PR-grddl-20070716/">Gleaning Resource Descriptions from Dialects of Languages (GRDDL)</ulink>,
W3C Proposed Recommendation of 2007-07-16 which allows reading XHTML
and XML as RDF triples by using profiles in the document that declare
XSLT transforms from the XHTML or XML content into RDF/XML or other
RDF syntax which can then be parsed.</para>

<para>The GRDDL parser is rather complex and different from the other
parsers in that it retrieves URIs, reads HTML documents (possibly
with errors), transforms the documents with XSLT and turns the result
into a single graph.  The default configuration of the GRDDL parser
also reads microformats (hcard, hcalendar) and follows &lt;link&gt;
tags that point to RDF/XML.  Parts of the GRDDL process can be
altered by configuration, which are describe below.
</para>

<para>The GRDDL parser defines 'base', 'Base' and 'url' XSLT parameters
with the value of the base URI to allow some XSLT sheets to work. These
set of parameters cannot be disabled.
</para>

<para>If the XSLT transform returns an empty string, no further
processing of the result is done, and a warning is generated.  The
xsl:output method is mapped to result document mime types as follows:
'text' to text/plain; 'xml' to application/xml and 'html' to text/html.
Any result that is of type 'application/xml' or unknown mime type
is assumed to be RDF/XML.
</para>

<para>The URIs that are processed during GRDDL operations can be checked
and skipped if required using a handler set with the
<link linkend="raptor-parser-set-uri-filter"><function>raptor_parser_set_uri_filter()</function></link>
function.  If the handler returns non-0, the URI is rejected.
This uses
<link linkend="raptor-www-set-uri-filter"><function>raptor_www_set_uri_filter()</function></link>
internally.
</para>

<para>If the value of option
<link linkend="RAPTOR-OPTION-WWW-TIMEOUT:CAPS"><literal>RAPTOR_OPTION_WWW_TIMEOUT</literal></link>
if set to a number &gt;0, it is used as the timeout in seconds
for retrieving of URIs during GRDDL processing.
This uses
<link linkend="raptor-www-set-connection-timeout"><function>raptor_www_set_connection_timeout()</function></link>
internally.
</para>

<para>The hardcoded support for hcard and hcalendar
microformats can be disabled by setting parser option
<link linkend="RAPTOR-OPTION-MICROFORMATS:CAPS"><literal>RAPTOR_OPTION_MICROFORMATS</literal></link>
to 0
or using
<link linkend="raptor-parser-set-option"><function>raptor_parser_set_option()</function></link>
with option
<link linkend="RAPTOR-OPTION-STRICT:CAPS"><literal>RAPTOR_OPTION_STRICT</literal></link>
and a boolean value of 1.
</para>

<para>The GRDDL parser by default will try an XML parser on the
content followed by a lax HTML parser.  This can be disabled by
setting parser option
<link linkend="RAPTOR-OPTION-HTML-TAG-SOUP:CAPS"><literal>RAPTOR_OPTION_HTML_TAG_SOUP</literal></link>
to 0
or using 
<link linkend="raptor-parser-set-option"><function>raptor_parser_set_option()</function></link>
with option
<link linkend="RAPTOR-OPTION-STRICT:CAPS"><literal>RAPTOR_OPTION_STRICT</literal></link>
and a boolean value of 1.
</para>

<para>The GRDDL parser by default will try to look for an HTML
&lt;link&gt; tag that points to RDF/XML.  This can be disabled by
setting parser option
<link linkend="RAPTOR-OPTION-HTML-LINK:CAPS"><literal>RAPTOR_OPTION_HTML_LINK</literal></link>
to 0
or using
<link linkend="raptor-parser-set-option"><function>raptor_parser_set_option()</function></link>
with option
<link linkend="RAPTOR-OPTION-STRICT:CAPS"><literal>RAPTOR_OPTION_STRICT</literal></link>
and a boolean value of 1.
</para>

</section>


<section id="parser-guess">
<title>Guess parser (name <literal>guess</literal>)</title>
<para>
This is a special parser that picks the actual parser to use based
on the content type, the content bytes or the content identifier.  The
content name can be either from a local file or from a URI.
</para>

<para>If the protocol that delivered the content (such as HTTP)
provided a <emphasis>Content Type</emphasis> (aka MIME Type) then
this will be the primary means for identifying th ecotnent.
</para>

<para>The secondary means to identify the content are the bytes of
the content (if available), otherwise the content identifier is used,
which is the least reliable.
</para>

</section>


<section id="parser-json">
<title>JSON parser (name <literal>json</literal>)</title>

<para>A parser for both the
resource-centric RDF/JSON syntax as defined by Talis at
<ulink url="http://n2.talis.com/wiki/RDF_JSON_Specification">RDF/JSON Specification</ulink>
and the triples-centric format based on the SPARQL results in JSON format.
</para>

</section>


<section id="parser-ntriples">
<title>N-Triples parser (name <literal>ntriples</literal>)</title>

<para>A parser for the
<ulink url="http://www.w3.org/TR/rdf-testcases/#ntriples">N-Triples</ulink>
syntax as used by the 
<ulink url="http://www.w3.org/2001/sw/RDFCore/">W3C RDF Core working group</ulink>
for the <ulink url="http://www.w3.org/TR/rdf-testcases/">RDF Test Cases</ulink>.
</para>

</section>


<section id="parser-rdfa">
<title>RDFa parser - (name <literal>rdfa</literal>)</title>
<para>
A parser for the
<ulink url="http://www.w3.org/TR/2008/CR-rdfa-syntax-20080620/">RDFa</ulink>
syntax, W3C Candidate Recommendation 20 June 2008 which allows reading XHTML
and XML as RDF triples by interpreting attributes on elements to
describe which ones have RDF semantics.   This is implemented via
<ulink url="http://rdfa.digitalbazaar.com/librdfa/">librdfa</ulink>
linked inside Raptor, written by Manu Sporny of Digital Bazaar,
and licensed with the same license as Raptor.
</para>

<para>
This parser is beta quality and passes all but 4 of the RDFa tests as
of Raptor 1.4.18.
</para>

</section>


<section id="parser-rdfxml">
<title>RDF/XML parser - default (name <literal>rdfxml</literal>)</title>
<para>
A parser for the standard
<ulink url="http://www.w3.org/TR/rdf-syntax-grammar/">RDF/XML syntax</ulink>
as revised by the
<ulink url="http://www.w3.org/2001/sw/RDFCore/">W3C RDF Core working group</ulink>.</para>

<para>This is the default parser in Raptor.</para>

<para>Features of this parser:</para>
<itemizedlist>
<listitem><para>Fully handles the <ulink url="http://www.w3.org/TR/rdf-syntax-grammar/">RDF/XML syntax updates</ulink> for <ulink url="http://www.w3.org/TR/xmlbase/">XML Base</ulink>, <literal>xml:lang</literal>, RDF datatyping and Collections.</para></listitem>

<listitem><para>Handles all RDF vocabularies such as <ulink url="http://www.foaf-project.org/">FOAF</ulink>, <ulink url="http://www.purl.org/rss/1.0/">RSS 1.0</ulink>, <ulink url="http://dublincore.org/">Dublin Core</ulink>, <ulink url="http://www.w3.org/TR/owl-features/">OWL</ulink>, <ulink url="http://usefulinc.com/doap">DOAP</ulink></para></listitem>

<listitem><para>Handles <literal>rdf:resource</literal> / <literal>resource</literal> attributes</para></listitem>

<listitem><para>Uses <ulink url="http://expat.sourceforge.net/">expat</ulink> and/or (GNOME) <ulink url="http://xmlsoft.org/">libxml</ulink> XML parsers as available or required</para></listitem>

</itemizedlist>

</section>


<section id="parser-rss-tag-soup">
<title>RSS Tag Soup parser (name <literal>rss-tag-soup</literal>)</title>

<para>A parser for the multiple XML RSS formats that use the elements
such as <literal>channel</literal>, <literal>item</literal>,
<literal>title</literal>, <literal>description</literal>
in different ways.
This includes support for the Atom 1.0 syndication format defined in IETF
<ulink url="http://www.ietf.org/rfc/rfc4287.txt">RFC 4287</ulink>
</para>

<para>The parser attempts to turn the input into
<ulink url="http://www.purl.org/rss/1.0/">RSS 1.0</ulink>
RDF triples in the RSS 1.0 model of a syndication feed.
This includes triples for RSS Enclosures.
</para>

<para>
True <ulink url="http://www.purl.org/rss/1.0/">RSS 1.0</ulink> when
wanted to be used as a full RDF vocabulary, is best parsed by the
RDF/XML parser (name <literal>rdfxml</literal>).
</para>

</section>


<section id="parser-trig">
<title>TRiG parser (name <literal>trig</literal>)</title>

<para>A parser for the
<ulink url="http://www.wiwiss.fu-berlin.de/suhl/bizer/TriG/Spec/">TriG - Turtle with Named Graphs</ulink>
syntax.
</para>

<para>The parser is alpha quality and may not support the entire TRiG
specification.</para>

</section>


<section id="parser-turtle">
<title>Turtle Terse RDF Triple Language parser (name <literal>turtle</literal>)</title>

<para>A parser for the
<ulink url="http://www.dajobe.org/2004/01/turtle/">Turtle Terse RDF Triple Language</ulink>
syntax, designed as a useful subset of
<ulink url="http://www.w3.org/DesignIssues/Notation3">Notation 3</ulink>.
</para>

</section>


</chapter>

<!--
Local variables:
mode: sgml
sgml-parent-document: ("raptor-docs.xml" "book" "part")
End:
-->