foaf/linkedin2rdf.pl



#!/usr/bin/perl -w

use HTML::Microformats;
use RDF::TrineShortcuts;
use File::Slurp;

my $markup = read_file( $ARGV[0] );
my $base_uri = $ARGV[1];

# hint: locate common abbreviations at http://prefix.cc/
# 
my $NAMESPACES= {
    rdfs => 'http://www.w3.org/2000/01/rdf-schema#',
    dc => 'http://purl.org/dc/terms/',
    foaf => 'http://xmlns.com/foaf/0.1/',
    xhtml => 'http://www.w3.org/1999/xhtml/vocab#',
    vcard => 'http://www.w3.org/2006/vcard/ns#',
    vcardx => 'http://buzzword.org.uk/rdf/vcardx#',
    hcterms => 'http://purl.org/uF/hCard/terms/',
    ical => 'http://www.w3.org/2002/12/cal/icaltzd#',
    cv => 'http://purl.org/captsolo/resume-rdf/0.2/cv#',
    hresume => 'http://ontologi.es/hresume#',
# TODO: somehow make RDF::Trine abbreviate labels
#   'xsd:dateTime' => 'http://www.w3.org/2001/XMLSchema#dateTime',
#   'geo:SpatialThing' => 'http://www.w3.org/2003/01/geo/wgs84_pos#SpatialThing',
};

my $doc = HTML::Microformats
    ->new_document($markup, $base_uri)
#   ->assume_profile(qw(hResume hCard hCalendar))
    ->assume_all_profiles
;
$doc->parse_microformats;
#print $doc->serialise_model(as => 'RDFXML');
my $model = $doc->model;
my $output = rdf_string($model,'RDFXML',
    namespaces => { %$NAMESPACES },
);

# TODO: somehow make RDF::Trine generate stable IDs
$output =~ s/(?<=rdf:nodeID=")B[a-f0-9]{32}(?=0[0-9]{3}")//g;

print $output;