Utilizing and building the Semantic Web with PHP
Christian Weiske
Semantic data?
- Web pages are made for humans:
Semantic data?
- Information extraction is hard for machines
- We need some way to make it better
RDF - Resource Description Framework
Statements
- subject predicate object
- This presentation has been created by Christian Weiske.
Resources
- This presentation: http://cweiske.de/download/talks/semantic-web-with-php.html
- created: http://purl.org/dc/elements/1.1/creator
- Christian Weiske: http://cweiske.de/
RDF #2
- example: cweiske.rdf
- Separate from webpage
- Linked via meta tag:
<link rel="meta" type="application/rdf+xml" title="FOAF" href="cweiske.rdf" />
RDF #3
- Vocabularies for predicates: Ontologies / taxonomies
- Many ontologies available:
- Dublin Core
- FOAF - Friend of a friend
- DOAP - Description Of A Project
- SIOC - Semantically Interlinked Online Communities
- CC - Creative Commons licenses
- Baetle - Issues and bugs
- Beer ontology
- Exhaustive list at schemaweb.info
RDF #4
The good
- Well defined standard
- Several serialization formats (N3, XML)
- Tool support, databases
The bad
- Separate from the data
- Hard to write
RDFa - Embedding RDF
- Embedding RDF directly into HTML
- Some new HTML properties:
- about - the subject
- typeof - the subject's type
- property - the predicate
- resource - the object
- datatype - the object's type
RDFa #2
<p xmlns:cal="http://www.w3.org/2002/12/cal/ical#"> I will be giving the talk <span property="cal:summary"> Utilizing and building the Semantic Web with PHP </span> on <span property="cal:dtstart" content="2010-08-19T14:00:00+02:00" datatype="xsd:dateTime" > Saturday 21st at 14:00 </span>. </p>
RDFa #3
The good
- Nearly as powerful as RDF
- Data are where the content is
The bad
- Still hard to write
Microformats
- Making it really easy to mark up data semantically
- Formats for common use cases:
- hCard
- hCalendar
- vote links
- XFN - XHTML Friends Network
- rel="tag"
- rel="license"
Microformat usage
- Define profile to use
<head profile='http://microformats.org/profile/vote-links'/>
<link rel="profile" href="http://microformats.org/profile/vote-links"/>This content uses
<a rel="profile" href="http://microformats.org/profile/vote-links" > vote links </a>.
Microformat usage #2
- Add classes
<a rev="vote-against" href="http://www.whitehouse.gov/president/gwbbio.html" > George W. Bush </a>
- Test it
GRDDL
Gleaning Resource Descriptions from Dialects of Languages
- Not everything is HTML
- Trillions of custom XML dialects in the wild
- Still want to express those data semantically
GRDDL #2
- Write XSL to transform custom XML dialect to RDF/XML
- Link transformation XSL in XML header
<xml version="1.0" encoding="utf-8"?> <superduperdata xmlns:grddl='http://www.w3.org/2003/g/data-view#' grddl:transformation="http://example.org/my/trans.xsl" >
Use cases
- Standardized ontologies - data exchange between applications
- Eases implementation of special search engines
- Search engines are using microformats:
- hCard for finding addresses
- rel="license" for usage right search
- hProduct + hReview for product search (Google and Yahoo! SearchMonkey)
- xFolk to import bookmarks from i.e. delicious directly into the browser or another bookmarking service.
- Browser extensions like Operator to make data available to users
Data storage
Generating semantic data
- Bringing data into RDF format is the programmer's task
- Tool support for RDF serialization is good
- RDF needs programmers
- Microformats can be embedded easily in HTML templates
Consuming semantic data
- Almost no one-stop solutions that parse everything
- Many specialized libraries with different levels of standards support
Libraries
Redland - librdf
- Written in C, bindings for PHP, Perl, Python and Ruby
- One-stop solution to data parsing: Raptor
- Data storage, querying and serialization
- librdf.org
RDF API 4 PHP
- Written in PHP
- Mature project (6 years)
- No activity since 2 years
- loads RDF, RSS and GRDDL
- Serializes RDF
- Storage in relational databases, supports SPARQL
- Simple web interface included
- rdfapi-php.sf.net
ARC
- Lightweight SPARQL API
- PHP4 and PHP5
- loads RDF, RDFa, Microformats
- Only for MySQL
- arc.semsol.org
Virtuoso Universal Server
- Database storage specialized on semantic data
- Access from PHP via ODBC: SQL and SPARQL
- Fastest triple storage
- Can import data via pluggable cartridges, available for many services (Flickr, IMDb, ...)
Erfurt
- Base of Ontowiki
- Partly based on RDFAPI for PHP
- Under active development
- Parses and stores RDF
- Storage + querying to relational databases, redland or Virtuoso
- ontowiki.net
PEAR RDF packages
- Attempt to add RDF API for PHP to PEAR
- Dead
PEAR's XML_GRDDL
- Extracts RDF from GRDDL resources
- pear.php.net/XML_GRDDL
hkit
- Microformat parser framework
- Supports hCalendar and hCard
- code.google.com/p/hkit/
Tools
Triplify
- Seeks to make exposing semantic data really easy
- Exports RDF from existing PHP applications
- Profiles: mapping from relational database to RDF
- Existing profiles for many popular applications (i.e. phpBB)
- triplify.org
OntoWiki
- Wiki application for triple stores
- Listing, searching, editing and display of RDF data
- Customizable renderers: Address book, calendar, mapping
- ontowiki.net
LESS
- Web service to render semantic data via templates
- Gathers data from SPARQL or RDFa
- Renders them with Smarty
- Makes it easy to display data from triple stores or SPARQL end points like dbpedia.org
- sourceforge.net/projects/less-st
- less.aksw.org
LDAP2Sparql
- Written in PHP
- Transforms LDAP into SPARQL queries
- Lets you use your triple store data from normal address books
- sourceforge.net/projects/ldap2sparql/
That's it
Questions?