WWW::Blog::Metadata - Extract common metadata from weblogs


    use WWW::Blog::Metadata;
    use Data::Dumper;
    my $uri;
    my $meta = WWW::Blog::Metadata->extract_from_uri($uri)
        or die WWW::Blog::Metadata->errstr;
    print Dumper $meta;


WWW::Blog::Metadata extracts common metadata from weblogs: syndication feed URIs, FOAF URIs, locative information, etc. Some benefits of using WWW::Blog::Metadata:

  • The extraction makes only one parsing pass over the HTML, rather than one for each type of metadata. It also attempts to be intelligent about only parsing as much of the HTML document as is required to give you the metadata that you need.

  • Many of the types of metadata that WWW::Blog::Metadata extracts can be found in multiple places in an HTML document. This module does the work for you, and abstracts it all behind an API.



Given a URI $uri pointing to a weblog, fetches the page contents, and attempts to extract common metadata from that weblog.

On error, returns undef, and the error message can be obtained by calling WWW::Blog::Metadata->errstr.

On success, returns a WWW::Blog::Metadata object.

WWW::Blog::Metadata->extract_from_html($html [, $base_uri ])

Uses the same extraction mechanism as extract_from_uri, but assumes that you've already fetched the HTML document and will provide it in $html, which should be a reference to a scalar containing the HTML.

If you know the base URI of the document, you should provide it in $base_uri. WWW::Blog::Metadata will attempt to find the base URI of the document if it's specified in the HTML itself, but you can give it a head start by passing in $base_uri.

This method has the same return value as extract_from_uri.


A reference to a hash of syndication feed URIs.

(Note: these are currently extracted using Feed::Find, which requires a separate parsing step, and sort of renders the above benefit #1 somewhat of a lie. This is done for maximum correctness, but it's possible it could change at some point.)


The URI for a FOAF file, specified in the standard manner used for FOAF auto-discovery.



The latitude and longitude specified for the weblog, from either icbm or geo.position <meta /> tags.


The tool that generated the weblog, from a generator <meta /> tag.


There are endless amounts of metadata that you might want to extract from a weblog, and the methods above are only what are provided by default. If you'd like to extract more information, you can use WWW::Blog::Metadata's plugin architecture to build access to the metadata that you want, while while making only one parsing pass over the HTML document.

The plugin architecture uses Module::Pluggable::Ordered, and it provides 2 pluggable events:

  • on_got_html

    This event is fired before the HTML document is parsed, so you should use it for extracting metadata after the page has been fetched (if you're using extract_from_uri), but before it's been parsed.

    Your method will receive 4 parameters: the class name; the WWW::Blog::Metadata object; a reference to a string containing the HTML document; and the base URI of the document.

    You could use this event to run heuristics on either the HTML or the URI, or both. The following example uses WWW::Blog::Identify to attempt to identify the true generator of the weblog:

        package WWW::Blog::Metadata::Flavor;
        use strict;
        use WWW::Blog::Identify qw( identify );
        use WWW::Blog::Metadata;
        WWW::Blog::Metadata->mk_accessors(qw( flavor ));
        sub on_got_html {
            my $class = shift;
            my($meta, $html, $base_uri) = @_;
            $meta->flavor( identify($base_uri, $$html) );
        sub on_got_html_order { 99 }

    This automatically adds a new accessor to the $meta object that is returned from the extract_from_* methods, so you can call

        my $meta = WWW::Blog::Metadata->extract_from_uri($uri);
        print $meta->flavor;

    to retrieve the name of the identified weblogging tool.

  • on_got_tag

    This event is fired for each HTML tag found in the document during the parsing phase.

    Your method will receive 5 parameters: the class name; the WWW::Blog::Metadata object; the tag name; a reference to a hash containing the tag attributes; and the base URI.

    The following example looks for the specific tag identifying the URI for the RSD (Really Simple Discoverability) file identifying the editing APIs that the weblog supports.

        package WWW::Blog::Metadata::RSD;
        use strict;
        use WWW::Blog::Metadata;
        WWW::Blog::Metadata->mk_accessors(qw( rsd_uri ));
        sub on_got_tag {
            my $class = shift;
            my($meta, $tag, $attr, $base_uri) = @_;
            if ($tag eq 'link' && $attr->{rel} =~ /\bEditURI\b/i &&
                $attr->{type} eq 'application/rsd+xml') {
                $meta->rsd_uri(URI->new_abs($attr->{href}, $base_uri)->as_string);
        sub on_got_tag_order { 99 }

    This automatically adds a new accessor to the $meta object that is returned from the extract_from_* methods, so you can call

        my $meta = WWW::Blog::Metadata->extract_from_uri($uri);
        print $meta->rsd_uri;

    to retrieve the URI for the RSD file.

  • on_finished

    This event is fired at the end of the extraction process.

    Your method will receive 2 parameters: the class name, and the WWW::Blog::Metadata object.


WWW::Blog::Metadata is free software; you may redistribute it and/or modify it under the same terms as Perl itself.


Except where otherwise noted, WWW::Blog::Metadata is Copyright 2005 Benjamin Trott, All rights reserved.