Generate search index for saved pages in ProcessWire

init.php

<?php

/**
 * Generate search index for saved pages
 *
 */
$this->addHook('Pages::saveReady', function(HookEvent $event) {
    $page = $event->arguments[0];
    if ($page->id && $page->hasField('search_index')) {
        $page->search_index = indexPage($page, array(
            'title',
            'headline',
            'summary',
            'body',
            'footer',
            'publish_from',
            'meta_keywords',
            'meta_description',
        ));
    }
});

/**
 * Store page data in search index and return generated index as a string
 *
 * @param Page $page
 * @param array $index_fields
 * @return string
 */
function indexPage(Page $page, $index_fields = array()) {
    $search_index = "";
    if ($page->id && !empty($index_fields)) {
        foreach ($page->fields as $field) {
            if (in_array($field->name, $index_fields)) {
                $formatted_data = $page->getFormatted($field->name);
                $search_index .= strip_tags($formatted_data);
                $search_index .= getIndexUrls($formatted_data);
            } else if ($field->type == "FieldtypePageTable") {
                foreach ($page->get($field->name) as $child) {
                    $search_index .= indexPage($child, $index_fields);
                }
            }
        }
    }
    return $search_index;
}

/**
 * Grab URLs from field data, mash them together into a space separated string,
 * and return the resulting string
 *
 * @param string $data
 * @return string
 */
function getIndexUrls($data) {
    $out = "";
    if ($data && preg_match_all('/href=([\"\'])(.*?)\1/i', $data, $matches)) {
        $out = " link:" . implode(" link:", $matches[2]) . " ";
    }
    return $out;
}