POE::Component::IRC::Plugin::WWW::OhNoRobotCom::Search - search website from IRC

        use strict;
        use warnings;

        use POE qw(Component::IRC Component::IRC::Plugin::WWW::OhNoRobotCom::Search);

        my $irc = POE::Component::IRC->spawn(
            nick        => 'OhNoRobotComBot',
            server      => '',
            port        => 6667,
            ircname     => 'Bot for searching',

            package_states => [
                main => [ qw(_start irc_001) ],


        sub _start {
            $irc->yield( register => 'all' );

                'OhNoRobot' =>
                        comic_id    => 56 # XKCD comics

            $irc->yield( connect => {} );

        sub irc_001 {
            $_[KERNEL]->post( $_[SENDER] => join => '#zofbot' );

        <Zoffix> OhNoRobotComBot, comic foo
        <Zoffix> OhNoRobotComBot, comic bar

    This module is a POE::Component::IRC plugin which uses
    POE::Component::IRC::Plugin for its base. It provides interface to
    search <> website from IRC. It accepts input from
    public channel events, "/notice" messages as well as "/msg" (private
    messages); although that can be configured at will.

        # plain and simple
            'OhNoRobotCom' => POE::Component::IRC::Plugin::WWW::OhNoRobotCom::Search->new

        # juicy flavor
            'OhNoRobotCom' =>
                    auto             => 1,
                    response_event   => 'irc_ohnorobot_results',
                    banned           => [ qr/aol\.com$/i ],
                    addressed        => 1,
                    root             => [ qr/$/i ],
                    trigger          => qr/^comics?\s+(?=\S)/i,
                    obj_args         => { debug => 1 },
                    max_results      => 3,
                    include          => [ qw(all_text meta) ],
                    comic_id         => 56, # XKCD comics
                    lucky            => 1,
                    listen_for_input => [ qw(public notice privmsg) ],
                    eat              => 1,
                    debug            => 0,

    The "new()" method constructs and returns a new
    "POE::Component::IRC::Plugin::WWW::OhNoRobotCom::Search" object suitable
    to be fed to POE::Component::IRC's "plugin_add" method. The constructor
    takes a few arguments, but *all of them are optional*. Note: most of
    these arguments can be changed on the fly by changing them as keys in
    your plugin's object, i.e. "$plug_obj->{banned} = [qr/.*/ ];". The
    possible arguments/values are as follows:

        ->new( auto => 0 );

    Optional. Takes either true or false values, specifies whether or not
    the plugin should auto respond to requests. When the "auto" argument is
    set to a true value plugin will respond to the requesting person with
    the results automatically. When the "auto" argument is set to a false
    value plugin will not respond and you will have to listen to the events
    emited by the plugin to retrieve the results (see EMITED EVENTS section
    and "response_event" argument for details). Defaults to: 1.

        ->new( response_event => 'event_name_to_recieve_results' );

    Optional. Takes a scalar string specifying the name of the event to emit
    when the results of the request are ready. See EMITED EVENTS section for
    more information. Defaults to: "irc_ohnorobot_results"

        ->new( banned => [ qr/aol\.com$/i ] );

    Optional. Takes an arrayref of regexes as a value. If the usermask of
    the person (or thing) making the request matches any of the regexes
    listed in the "banned" arrayref, plugin will ignore the request.
    Defaults to: "[]" (no bans are set).

        ->new( root => [ qr/\\E$/i ] );

    Optional. As opposed to "banned" argument, the "root" argument allows
    access only to people whose usermasks match any of the regexen you
    specify in the arrayref the argument takes as a value. By default: it is
    not specified. Note: as opposed to "banned" specifying an empty arrayref
    to "root" argument will restrict access to everyone.

        ->new( trigger => qr/^comics?\s+(?=\S)/i );

    Optional. Takes a regex as an argument. Messages matching this regex
    will be considered as requests. See also addressed option below which is
    enabled by default. Note: the trigger will be removed from the message,
    therefore make sure your trigger doesn't match the actual data that
    needs to be processed. Defaults to: "qr/^comics?\s+(?=\S)/i"

        ->new( obj_args => { debug => 1 } );

    Optional. Takes a hashref as an argument which will be dereferenced
    directly into POE::Component::WWW::OhNoRobotCom::Search's constructor.
    See documentation for POE::Component::WWW::OhNoRobotCom::Search for
    possible arguments. Defaults to: "{}" ( default constructor )

        ->( comic_id => 56 )

    Optional. The "comic_id" argument takes a scalar as a value which should
    be a comic ID number or an empty string which indicates that search
    should be done on all comics. To obtain the comic ID number go to
    <>, "View Source" and
    search for the name of the comic, when you'll find an <option> the
    "value=""" attribute of that option will be the number you are looking
    for. Idealy, it would make sense to make the "comic_id" argument accepts
    names instead of those numbers, but there are just too many (500+)
    different comics sites and new are being added, blah blah. Defaults to:
    empty string, meaning search through all the comics.

        ->( include => [ qw(all_text meta) ] )

    Optional. Specifies what kind of "things" to include into consideration
    when performing the search. Takes an arrayref as an argument. Defaults
    to: all possible elements included which are as follows:

    all_text  Include *All comic text*.

    scene     Include *Scene descriptions*.

    sound     Include *Sound effects*.

    speakers  Include *Speakers' names*

    link      Include *Link text*.

    meta      Include *Meta information*

        ->new( max_results => 3 )

    Optional. The number of results displayed on <>
    is 10, the object will send out several requests if needed to obtain the
    number of results specified in the "max_results" argument. Don't use
    extremely large values here, as the amount of requests will NOT be
    "max_results / 10" because results are often repeating and the object
    will count only unique URIs on the results page. Note: values less than
    10 for this argument will yield to "max_results" results being presented
    on IRC (when "auto" is turned on) but the response event will likely to
    recieve more than that. Defaults to: 3 (this does not necessarily mean
    that the object will send only one request).

        ->new( lucky => 1 );

    ARE YOU FEELING LUCKY?!!? If so, the "lucky" argument, when set to a
    true value, will "press" the *Let the robot decide* button on
    <> and the "search" method/event will fetch a
    poiting to the comic which the *ahem* "robot" thinks is what you want.
    Note: when using the "lucky" argument search will error out (see "EMITED
    EVENTS" section below) if nothing was found. Defaults to: a false value
    (no feelin' lucky :( )

        ->new( addressed => 1 );

    Optional. Takes either true or false values. When set to a true value
    all the public messages must be *addressed to the bot*. In other words,
    if your bot's nickname is "Nick" and your trigger is "qr/^trig\s+/" you
    would make the request by saying "Nick, trig foo". When addressed mode
    is turned on, the bot's nickname, including any whitespace and common
    punctuation character will be removed before matching the "trigger" (see
    above). When "addressed" argument it set to a false value, public
    messages will only have to match "trigger" regex in order to make a
    request. Note: this argument has no effect on "/notice" and "/msg"
    requests. Defaults to: 1

        ->new( listen_for_input => [ qw(public  notice  privmsg) ] );

    Optional. Takes an arrayref as a value which can contain any of the
    three elements, namely "public", "notice" and "privmsg" which indicate
    which kind of input plugin should respond to. When the arrayref contains
    "public" element, plugin will respond to requests sent from messages in
    public channels (see "addressed" argument above for specifics). When the
    arrayref contains "notice" element plugin will respond to requests sent
    to it via "/notice" messages. When the arrayref contains "privmsg"
    element, the plugin will respond to requests sent to it via "/msg"
    (private messages). You can specify any of these. In other words,
    setting "( listen_for_input =" [ qr(notice privmsg) ] )> will enable
    functionality only via "/notice" and "/msg" messages. Defaults to: "[
    qw(public notice privmsg) ]"

        ->new( eat => 0 );

    Optional. If set to a false value plugin will return a "PCI_EAT_NONE"
    after responding. If eat is set to a true value, plugin will return a
    "PCI_EAT_ALL" after responding. See POE::Component::IRC::Plugin
    documentation for more information if you are interested. Defaults to: 1

        ->new( debug => 1 );

    Optional. Takes either a true or false value. When "debug" argument is
    set to a true value some debugging information will be printed out. When
    "debug" argument is set to a false value no debug info will be printed.
    Defaults to: 0.

        $VAR1 = {
            'who' => 'Zoffix!n=Zoffix@unaffiliated/zoffix',
            'type' => 'public',
            'channel' => '#zofbot',
            'term' => 'bar',
            'message' => 'OhNoRobotComBot, comic bar',
            'results' => {
                '' => 'Scary Go Round :: Monday-Friday Comic by John Allison',
                '' => 'Trashspotting',
                '' => 'The Suburban Jungle, Starring Tiffany Tiger - Archives',
                '' => '#99',
                '' => 'Goats comic strip from December / 31 / 1999: tastes like chicken candy',
                '' => 'College Roomies from Hell!!! for Thursday, March 6, 2008',
                '' => 'Webcomic Rogues of Clwyd-Rhan: The Rogues find they have an unexpected ally.',
                '' => 'Jesus and Mo',
                '' => 'General Protection Fault--The Comic Strip',
                '' => 'Almost Psychic'

    The event handler set up to handle the event, name of which you've
    specified in the "response_event" argument to the constructor (it
    defaults to "irc_ohnorobot_results") will recieve input every time
    request is completed. The input will come in a form of a hashref in
    $_[ARG0]. The possible keys/values of that hashref are as follows:

        { 'who' => 'Zoffix!n=Zoffix@unaffiliated/zoffix' }

    The usermask of the person who made the request.

        { 'term' => 'bar' }

    The search term which was used for searching. (this is "message" key
    with trigger stripped out)

        { 'type' => 'public' }

    The type of the request. This will be either "public", "notice" or

        { 'channel' => '#zofbot' }

    The channel where the message came from (this will only make sense when
    the request came from a public channel as opposed to /notice or /msg)

        { 'message' => 'OhNoRobotComBot, comic bar' }

    The full message that the user has sent.

        { 'error' => 'Network error: 500 read timeout' }

    If an error occured during the search the "error" key will be present
    and will contain explanation for the error.

        'results' => {
            '' => 'Scary Go Round :: Monday-Friday Comic by John Allison',
            '' => 'Trashspotting',
            '' => 'The Suburban Jungle, Starring Tiffany Tiger - Archives',


    The "results" key will contain a hashref with search results. The keys
    will be links poiting to comics found and values will be the titles of
    the comics. Note: if "max_results" argument to the constructor is set to
    a value less than 10 (it's 3 by default) then "results" key will likely
    to contain more results than "max_results".

    Zoffix Znet, "<zoffix at>" (<>,

    Please report any bugs or feature requests to
    "bug-poe-component-irc-plugin-www-ohnorobotcom-search at",
    or through the web interface at
    -WWW-OhNoRobotCom-Search>. I will be notified, and then you'll
    automatically be notified of progress on your bug as I make changes.

    You can find documentation for this module with the perldoc command.

        perldoc POE::Component::IRC::Plugin::WWW::OhNoRobotCom::Search

    You can also look for information at:

    *   RT: CPAN's request tracker


    *   AnnoCPAN: Annotated CPAN documentation


    *   CPAN Ratings


    *   Search CPAN


    Copyright 2008 Zoffix Znet, all rights reserved.

    This program is free software; you can redistribute it and/or modify it
    under the same terms as Perl itself.