root/link-harvest/trunk/link-harvest.php

<?php

/*
Plugin Name: Link Harvest
Plugin URI: http://alexking.org/projects/wordpress
Description: This will harvest links from your WordPress database, creating a links list sorted by popularity. Once you have activated the plugin, you can configure the <a href="options-general.php?page=link-harvest.php">Settings</a> and see your <a href="index.php?page=link-harvest.php">list of links</a>. Also see <a href="options-general.php?page=link-harvest.php#aklh_template_tags">how to show the list of links</a> in your blog. Questions on configuration, etc.? Make sure to read the README.
Version: 1.1
Author: Alex King
Author URI: http://alexking.org
*/ 

// Copyright (c) 2006-2007 Alex King. All rights reserved.
// http://alexking.org/projects/wordpress
//
// Released under the GPL license
// http://www.opensource.org/licenses/gpl-license.php
//
// This is an add-on for WordPress
// http://wordpress.org/
//
// **********************************************************************
// This program is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. 
// **********************************************************************

@define('AK_WPROOT', '../../');

@define('AKLH_DEBUG', false);

if (AKLH_DEBUG) {
//    ini_set('display_errors', '1');
//    ini_set('error_reporting', E_ALL);
    $logfile = dirname(__FILE__).'/aklh_log.txt';
    if (!is_file($logfile) || !is_writeable($logfile)) {
        die('Debugging is enabled, but there is no wp-content/plugins/aklh_log.txt file or the file is not writeable.');
    }
}

if (!isset($wpdb)) {
    require(AK_WPROOT.'wp-blog-header.php');
}

load_plugin_textdomain('link-harvest');

if (!function_exists('is_admin_page')) {
    function is_admin_page() {
        if (function_exists('is_admin')) {
            return is_admin();
        }
        if (function_exists('check_admin_referer')) {
            return true;
        }
        else {
            return false;
        }
    }
}

if (!function_exists('ak_prototype')) {
    function ak_prototype() {
        if (!function_exists('wp_enqueue_script')) {
            global $ak_prototype;
            if (!isset($ak_prototype) || !$ak_prototype) {
                print('
        <script type="text/javascript" src="'.get_bloginfo('wpurl').'/wp-includes/js/prototype.js"></script>
                ');
            }
            $ak_prototype = true;
        }
    }
}


$wpdb->ak_domains = $wpdb->prefix.'ak_domains';
$wpdb->ak_linkharvest = $wpdb->prefix.'ak_linkharvest';

$aklh = new ak_link_harvest;

// CHECK FOR LINK HARVEST TABLES

if (isset($_GET['activate']) && $_GET['activate'] == 'true') {
    $tables = $wpdb->get_col("
        SHOW TABLES
    ");
    if (!in_array($wpdb->ak_linkharvest, $tables) && !in_array($wpdb->ak_domains, $tables)) {
        $aklh->install();
    }
}
$aklh->get_settings();

class ak_link_harvest {
    var $exclude;
    var $table_length;
    var $harvest_enabled;
    var $options;
    var $default_limit;
    var $timeout;
    var $token;
    
    function ak_link_harvest() {
        $this->exclude = array();
        $this->table_length = 100;
        $this->harvest_enabled = 1;
        $this->default_limit = 5;
        $this->timeout = 2;
        $this->token = 1;
        $this->options = array(
            'exclude' => 'explode'
            , 'table_length' => 'int'
            , 'harvest_enabled' => 'int'
            , 'token' => 'int'
        );
        $this->excluded_file_extensions = array(
            '.mov'
            , '.jpg'
            , '.gif'
            , '.png'
            , '.pdf'
            , '.mpg'
            , '.mp3'
            , '.mpeg'
            , '.avi'
            , '.swf'
            , '.doc'
            , '.xls'
            , '.wmv'
            , '.wmf'
            , '.wma'
            , '.txt'
            , '.m4p'
        );
    }
    
    function install() {
        global $wpdb;
        $result = $wpdb->query("
            CREATE TABLE `$wpdb->ak_domains` (
            `id` int(11) NOT NULL auto_increment,
            `domain` varchar(255) NOT NULL,
            `title` varchar(255) NULL,
            `count` int(11) NOT NULL default '0',
            PRIMARY KEY  (`id`),
            UNIQUE KEY `domain` (`domain`)
            )
        ");
        $result = $wpdb->query("
            CREATE TABLE `$wpdb->ak_linkharvest` (
            `id` int(11) NOT NULL auto_increment,
            `post_id` int(11) NOT NULL,
            `url` text NOT NULL,
            `title` varchar(255) NULL,
            `domain_id` varchar(255) NOT NULL,
            `modified` datetime NOT NULL,
            PRIMARY KEY  (`id`),
            KEY `post_id` (`post_id`),
            KEY `domain_id` (`domain_id`)
            )
        ");
        add_option('aklh_exclude', '', 'Ignore links to these domains.');
        add_option('aklh_table_length', '100', 'Number of domains to show in the links table.');
        add_option('aklh_harvest_enabled', '1', 'Can we run a link harvest?');
        add_option('aklh_token', '1', 'Use ###linkharvest### to show the links list?');
    }
    
    function clear_data() {
        global $wpdb;
        $wpdb->query("
            TRUNCATE TABLE $wpdb->ak_domains
        ");
        $wpdb->query("
            TRUNCATE TABLE $wpdb->ak_linkharvest
        ");
    }
    
    function get_settings() {
        foreach ($this->options as $option => $type) {
            $this->$option = get_option('aklh_'.$option);
            switch ($type) {
                case 'explode':
                    $this->$option = explode(' ', $this->$option);
                    break;
                case 'int':
                    $this->$option = intval($this->$option);
                    break;
            }
        }
    }
    
    function update_settings() {
        foreach ($this->options as $option => $type) {
            if (isset($_POST[$option])) {
                switch ($type) {
                    case 'explode':
                        $value = stripslashes($_POST[$option]);
                        break;
                    case 'int':
                        $value = intval($_POST[$option]);
                        break;
                    default:
                        $value = stripslashes($_POST[$option]);
                }
                update_option('aklh_'.$option, $value);
            }
        }

        header('Location: '.get_bloginfo('wpurl').'/wp-admin/options-general.php?page=link-harvest.php&updated=true');
        die();
    }
    
    function excluded_file_type($link) {
        foreach ($this->excluded_file_extensions as $ext) {
            if (substr($link, strlen($ext) * -1) == $ext) {
                return true;
            }
        }
        return false;
    }
    
    function get_links($body) {
// From: http://www.onaje.com/php/article.php4/46
        $href_regex ="<";            // 1 start of the tag
        $href_regex .="\s*";         // 2 zero or more whitespace
        $href_regex .="a";           // 3 the a of the <a> tag itself
        $href_regex .="\s+";         // 4 one or more whitespace
        $href_regex .="[^>]*";       // 5 zero or more of any character that is _not_ the end of the tag
        $href_regex .="href";        // 6 the href bit of the tag
        $href_regex .="\s*";         // 7 zero or more whitespace
        $href_regex .="=";           // 8 the = of the tag
        $href_regex .="\s*";         // 9 zero or more whitespace
        $href_regex .="[\"']?";      // 10 none or one of " or '
        $href_regex .="(";           // 11 opening parenthesis, start of the bit we want to capture
        $href_regex .="[^\"' >]+";   // 12 one or more of any character _except_ our closing characters
        $href_regex .=")";           // 13 closing parenthesis, end of the bit we want to capture
        $href_regex .="[\"' >]";     // 14 closing chartacters of the bit we want to capture
        
        $regex  = "/";         // regex start delimiter
        $regex .= $href_regex; //
        $regex .= "/";         // regex end delimiter
        $regex .= "i";         // Pattern Modifier - makes regex case insensative
        $regex .= "s";         // Pattern Modifier - makes a dot metacharater in the pattern 
                    // match all characters, including newlines
        $regex .= "U";         // Pattern Modifier - makes the regex ungready

        $urls = array();

        if (preg_match_all($regex, $body, $links)) {
            foreach ($links[1] as $link) {
                if (in_array(substr($link, 0, 7), array('http://', 'https:/')) && !$this->excluded_file_type($link)) {
                    $urls[] = trim($link);
                }
            }
        }
        
        return $urls;
    }
    
    function get_domain_id($domain) {
        global $wpdb;
        $domain_id = $wpdb->get_var("
            SELECT id
            FROM $wpdb->ak_domains
            WHERE domain = '$domain'
        ");
        if ($domain_id && !is_null($omain_id)) {
            return $domain_id;
        }
        else {
            $id = $this->add_domain($domain);
            if ($id != false) {
                return $id;
            }
        }
        return false;
    }

    function get_domain_ids($domains) {
        global $wpdb;
        if (!is_array($domains) || count($domains) == 0) {
            return false;
        }
        $domain_ids = array();
        $results = $wpdb->get_results("
            SELECT id, domain
            FROM $wpdb->ak_domains
            WHERE domain IN ('".implode("', '", $domains)."')
        ");
        if ($results && count($results) > 0) {
            foreach ($results as $data) {
                $domain_ids[$data->domain] = $data->id;
            }
        }
        $missing = array();
        foreach ($domains as $domain) {
            if (!isset($domain_ids[$domain])) {
                $domain_ids[$domain] = $this->add_domain($domain);
            }
        }
        return $domain_ids;
    }
    
    function process_content($content, $post_id) {
        $links = $this->get_links($content);
        if ($links == false) {
            return;
        }
        
        aklh_log('Found '.count($links).' links in post id: '.$post_id);

        $domains = array();
        $domain_count = array();
        $harvest = array();
        
        foreach ($links as $link) {
// FeedBurner hack, working around this:
// http://alexking.org/blog/2006/12/01/why-i-dont-use-feedburner
            if (!empty($link)) {
                if (strstr($link, '/feeds.feedburner.com/') || strstr($link, '/~r/')) {
                    require_once(ABSPATH.WPINC.'/class-snoopy.php');
                    $snoop = new Snoopy;
                    $snoop->maxlength = 2000;
                    $snoop->read_timeout = $this->timeout;
                    $snoop->fetch($link);
                    if (!empty($snoop->lastredirectaddr)) {
                        $link = $snoop->lastredirectaddr;
                    }
                }
                $domain = $this->get_domain($link);
                if (!empty($domain)) {
                    foreach ($this->exclude as $exclude) {
                        if (strstr($domain, $exclude)) {
                            return;
                        }
                    }
                    if (isset($domain_count[$domain])) {
                        $domain_count[$domain]++;
                    }
                    else {
                        $domains[] = $domain;
                        $domain_count[$domain] = 1;
                    }
                    $harvest[] = array($link, $domain);
                }

                aklh_log('Processed link: '.$link);

            }
            else {

                aklh_log('Skipped link: '.$link);

            }
        }
        
        if (count($domains) > 0) {
            $domain_ids = $this->get_domain_ids($domains);

            foreach ($harvest as $data) {
                $link = $data[0];
                $domain = $data[1];
                $this->add_link($link, $domain_ids[$domain], $post_id);
            }
            
            foreach ($domain_ids as $domain => $domain_id) {
                $this->set_domain_counter(null, $domain_id, $domain_count[$domain], '+');
            }
        }

    }
    
    function add_link($url, $domain_id, $post_id) {

        aklh_log('About to add link for post id: '.$post_id.' - '.$url);

        global $wpdb;
        $title = stripslashes($this->get_page_title($url));
        $result = $wpdb->query("
            INSERT INTO $wpdb->ak_linkharvest
            ( post_id
            , url
            , title
            , domain_id
            , modified
            )
            VALUES
            ( '".mysql_real_escape_string($post_id)."'
            , '".mysql_real_escape_string($url)."'
            , '".mysql_real_escape_string($title)."'
            , '".mysql_real_escape_string($domain_id)."'
            , '".current_time('mysql',1)."'
            )
        ");
        if (!$result) {

            aklh_log('Failed to add link for post id: '.$post_id.' - '.$url);

            return false;
        }
        else {

            aklh_log('Added link for for post id: '.$post_id.' - '.$url);

            return true;
        }
    }
    
    function add_domain($domain) {
        global $wpdb;
        $title = stripslashes($this->get_page_title('http://'.$domain));
        $result = $wpdb->query("
            INSERT INTO $wpdb->ak_domains
            ( domain
            , title
            , count
            )
            VALUES
            ( '".mysql_real_escape_string($domain)."'
            , '".mysql_real_escape_string($title)."'
            , '0'
            )
        ");
        if (!$result) {

            aklh_log('Failed to add domain: '.$domain);

            return false;
        }
        else {

            aklh_log('Added domain: '.$domain);

            return mysql_insert_id();
        }
    }
    
    function set_domain_counter($domain = null, $domain_id = null, $count, $mod = '+') {
        global $wpdb;
        if (!is_null($domain)) {
            $where = " WHERE domain = '".mysql_real_escape_string($domain)."' ";
        }
        else if (!is_null($domain_id)) {
            $where = " WHERE id = '".intval($domain_id)."' ";
        }
        if (isset($where)) {
            $result = $wpdb->query("
                UPDATE $wpdb->ak_domains
                SET count = count $mod ".intval($count)."
                $where
            ");
            if ($result != false) {
                return true;
            }
        }
        return false;
    }
    
    function harvest_posts($post_ids = array(), $start = 0, $limit = null, $delete_old = false) {
        global $wpdb;
        if (is_array($post_ids) && count($post_ids) > 0) {
            if ($delete_old) {
                foreach ($post_ids as $post_id) {
                    $links = $wpdb->get_results("
                        SELECT *
                        FROM $wpdb->ak_linkharvest
                        WHERE post_id = $post_id
                    ");
                    if (count($links) > 0) {
                        $domains = array();
                        foreach ($links as $link) {
                            if (!isset($domains[$link->domain_id])) {
                                $domains[$link->domain_id] = 1;
                            }
                            else {
                                $domains[$link->domain_id]++;
                            }
                        }
                        if (count($domains) > 0) {
                            foreach ($domains as $id => $count) {
                                $update = $wpdb->query("
                                    UPDATE $wpdb->ak_domains
                                    SET count = count - $count
                                    WHERE id = $id
                                ");
                            }
                        }
                        $delete = $wpdb->query("
                            DELETE
                            FROM $wpdb->ak_linkharvest
                            WHERE post_id = $post_id
                        ");
                    }
                }
            }
            $posts = $wpdb->get_results("
                SELECT *
                FROM $wpdb->posts
                WHERE (
                    post_status = 'publish'
                    OR post_status = 'static'
                )
                AND ID IN (".implode(',', $post_ids).")
            ");
        }
        else {
            if ($start == 0) {
                $this->clear_data();
            }
            if (is_null($limit)) {
                $limit = $this->default_limit;
            }
            $posts = $wpdb->get_results("
                SELECT *
                FROM $wpdb->posts
                WHERE (
                    post_status = 'publish'
                    OR post_status = 'static'
                )
                AND post_content LIKE '%http://%'
                ORDER BY ID
                LIMIT $start, $limit
            ");
        }
        foreach ($posts as $post) {

            aklh_log('== Start processing post id: '.$post->ID);

            $this->process_content($post->post_content, $post->ID);
            $external = get_post_meta($post->ID, 'external_link', true);
            if (!empty($external