Metagen.php

Go to the documentation of this file.

<?php
/*
 You may not change or alter any portion of this comment or credits
 of supporting developers from this source code or any supporting source code
 which is considered copyrighted (c) material of the original comment or credit authors.

 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.
 */

namespace Xmf;

class Metagen
{

    const ELLIPSIS = "…"; // unicode horizontal ellipsis U+2026

    public static function assignTitle($title)
    {
        global $xoopsTpl, $xoTheme;

        $title = trim($title);
        $title = self::asPlainText($title);
        if (!empty($title)) {
            if (is_object($xoTheme)) {
                $xoTheme->addMeta('meta', 'title', $title);
            }
            $xoopsTpl->assign('xoops_pagetitle', $title);
        }
    }

    public static function assignKeywords($keywords)
    {
        global $xoopsTpl, $xoTheme;

        if (!empty($keywords) && is_array($keywords)) {
            $keyword_tag = implode(', ', $keywords);

            if (!empty($keyword_tag)) {
                if (is_object($xoTheme)) {
                    $xoTheme->addMeta('meta', 'keywords', $keyword_tag);
                } else {
                    $xoopsTpl->assign('xoops_meta_keywords', $keyword_tag);
                }
            }
        }
    }

    public static function assignDescription($description)
    {
        global $xoopsTpl, $xoTheme;

        $description = trim($description);
        if (!empty($description)) {
            if (is_object($xoTheme)) {
                $xoTheme->addMeta('meta', 'description', $description);
            } else {
                $xoopsTpl->assign('xoops_meta_description', $description);
            }
        }
    }

    public static function generateKeywords(
        $body,
        $count = 20,
        $minLength = 4,
        $forceKeys = null
    ) {
        $keywords = array();
        $keycnt = array();
        if (!is_array($forceKeys)) {
            $forceKeys = array();
        }

        $text = self::asPlainText($body);
        $text = mb_strtolower($text);

        $originalKeywords = preg_split(
            '/[^a-zA-Z\'"-]+/',
            $text,
            -1,
            PREG_SPLIT_NO_EMPTY
        );

        foreach ($originalKeywords as $originalKeyword) {
            if (self::checkStopWords($originalKeyword)) {
                $secondRoundKeywords = explode("'", $originalKeyword);
                foreach ($secondRoundKeywords as $secondRoundKeyword) {
                    if (self::checkStopWords($secondRoundKeyword)
                        && strlen($secondRoundKeyword) >= $minLength
                    ) {
                        $keycnt[$secondRoundKeyword] =
                            empty($keycnt[$secondRoundKeyword]) ? 1 : $keycnt[$secondRoundKeyword] + 1;
                    }
                }
            }
        }

        while (!empty($forceKeys)) {
            $tempkey = strtolower(array_pop($forceKeys));
            $keycnt[$tempkey] = 999999;
        }

        arsort($keycnt, SORT_NUMERIC);
        $key = array_keys($keycnt);
        $keywords = array_slice($key, 0, $count);

        return $keywords;
    }

    protected static function checkStopWords($key)
    {
        static $stopwords = null;

        if (!$stopwords) {
            if (!defined('_XMF_STOPWORDS')) {
                \Xmf\Language::load('stopwords', 'xmf');
            }
            if (defined('_XMF_STOPWORDS')) {
                $sw = explode(' ', _XMF_STOPWORDS);
                $stopwords = array_fill_keys($sw, true);
            } else {
                $stopwords = array('_'=> true);
            }
        }
        if ($stopwords) {
            return !isset($stopwords[mb_strtolower($key)]);
        }
        return true;
    }

    public static function generateDescription($body, $wordCount = 100)
    {
        $text = self::asPlainText($body);

        $words = explode(" ", $text);

        // Only keep $maxWords words
        $newWords = array();
        $i = 0;
        while ($i < $wordCount - 1 && $i < count($words)) {
            $newWords[] = $words[$i];
            ++$i;
        }
        $ret = implode(' ', $newWords);
        $len = mb_strlen($ret);
        $lastperiod = mb_strrpos($ret, '.');
        $ret .= ($lastperiod === false) ? self::ELLIPSIS : '';
        if ($len>100 && ($len-$lastperiod)<30) {
            $ret = mb_substr($ret, 0, $lastperiod+1);
        }

        return $ret;
    }

    public static function generateMetaTags(
        $title,
        $body,
        $count = 20,
        $minLength = 4,
        $wordCount = 100,
        $forceKeys = null
    ) {
        $title_keywords = self::generateKeywords($title, $count, 3, $forceKeys);
        $keywords = self::generateKeywords($body, $count, $minLength, $title_keywords);
        $description = self::generateDescription($body, $wordCount);
        self::assignTitle($title);
        self::assignKeywords($keywords);
        self::assignDescription($description);
    }

    protected static function nonEmptyString($var)
    {
        return (strlen($var) > 0);
    }

    public static function generateSeoTitle($title = '', $extension = '')
    {
        $title = preg_replace("/[^a-zA-Z0-9]/", "-", $title);
        $title = \Normalizer::normalize($title, \Normalizer::FORM_C);

        $tableau = explode("-", $title);
        $tableau = array_filter($tableau, 'self::nonEmptyString');
        $tableau = array_filter($tableau, 'self::checkStopWords');
        $title = implode("-", $tableau);

        $title = (empty($title)) ? '' : $title . $extension;
        return $title;
    }

    public static function getSearchSummary($haystack, $needles = null, $length = 120)
    {
        $encoding = 'UTF-8';

        $haystack = self::asPlainText($haystack);
        $pos = self::getNeedlePositions($haystack, $needles);

        $start = empty($pos) ? 0 : min($pos);

        $start = max($start - (int)($length/2), 0);

        $pre = ($start > 0); // need an ellipsis in front?
        if ($pre) {
            // we are not at the begining so find first blank
            $temp = mb_strpos($haystack, ' ', $start, $encoding);
            $start = ($temp === false) ? $start : $temp;
            $haystack = mb_substr($haystack, $start, null, $encoding);
        }

        $post = !(mb_strlen($haystack, $encoding) < $length);  // need an ellipsis in back?
        if ($post) {
            $haystack = mb_substr($haystack, 0, $length, $encoding);
            $end = mb_strrpos($haystack, ' ', 0, $encoding);
            if ($end) {
                $haystack = mb_substr($haystack, 0, $end, $encoding);
            }
        }

        $haystack = ($pre ? self::ELLIPSIS : '') . trim($haystack) . ($post ? self::ELLIPSIS : '');
        return $haystack;
    }

    protected static function asPlainText($rawText)
    {
        $text = $rawText;
        $utilities = new Utilities();
        $text = $utilities->html2text($text);
        $text = $utilities->purifyText($text);

        $text = str_replace(array("\n", "\r"), ' ', $text);
        $text = preg_replace('/[ ]* [ ]*/', ' ', $text);

        return trim($text);
    }

    private static function getNeedlePositions($haystack, $needles)
    {
        $pos=array();
        $needles = empty($needles) ? array() : (array) $needles;
        foreach ($needles as $needle) {
            $i = mb_stripos($haystack, $needle, 0, 'UTF-8');
            if ($i!==false) {
                $pos[] = $i; // only store matches
            }
        }
        return $pos;
    }
}