meilisearch/Classes/HtmlContentExtractor.php
2021-04-17 21:20:54 +02:00

258 lines
8.2 KiB
PHP
Raw Blame History

<?php
namespace WapplerSystems\Meilisearch;
/***************************************************************
* Copyright notice
*
* (c) 2011-2015 Ingo Renner <ingo@typo3.org>
* All rights reserved
*
* This script is part of the TYPO3 project. The TYPO3 project is
* free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 3 of the License, or
* (at your option) any later version.
*
* The GNU General Public License can be found at
* http://www.gnu.org/copyleft/gpl.html.
*
* This script 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. See the
* GNU General Public License for more details.
*
* This copyright notice MUST APPEAR in all copies of the script!
***************************************************************/
use WapplerSystems\Meilisearch\System\Configuration\TypoScriptConfiguration;
/**
* A content extractor to get clean, indexable content from HTML markup.
*
* @author Ingo Renner <ingo@typo3.org>
*/
class HtmlContentExtractor
{
/**
* Unicode ranges which should get stripped before sending a document to meilisearch.
* This is necessary if a document (PDF, etc.) contains unicode characters which
* are valid in the font being used in the document but are not available in the
* font being used for displaying results.
*
* This is often the case if PDFs are being indexed where special fonts are used
* for displaying bullets, etc. Usually those bullets reside in one of the unicode
* "Private Use Zones" or the "Private Use Area" (plane 15 + 16)
*
* @see http://en.wikipedia.org/wiki/Unicode_block
* @var array
*/
protected static $stripUnicodeRanges = [
['FFFD', 'FFFD'],
// Replacement Character (<28>) @see http://en.wikipedia.org/wiki/Specials_%28Unicode_block%29
['E000', 'F8FF'],
// Private Use Area (part of Plane 0)
['F0000', 'FFFFF'],
// Supplementary Private Use Area (Plane 15)
['100000', '10FFFF'],
// Supplementary Private Use Area (Plane 16)
];
/**
* The raw HTML markup content to extract clean content from.
*
* @var string
*/
protected $content;
/**
* Mapping of HTML tags to Meilisearch document fields.
*
* @var array
*/
protected $tagToFieldMapping = [
'h1' => 'tagsH1',
'h2' => 'tagsH2H3',
'h3' => 'tagsH2H3',
'h4' => 'tagsH4H5H6',
'h5' => 'tagsH4H5H6',
'h6' => 'tagsH4H5H6',
'u' => 'tagsInline',
'b' => 'tagsInline',
'strong' => 'tagsInline',
'i' => 'tagsInline',
'em' => 'tagsInline',
'a' => 'tagsA',
];
/**
* @var TypoScriptConfiguration
*/
private $configuration;
/**
* Constructor.
*
* @param string $content Content HTML markup
*/
public function __construct($content)
{
// @extensionScannerIgnoreLine
$this->content = $content;
}
/**
* @return TypoScriptConfiguration|array
*/
protected function getConfiguration()
{
if ($this->configuration == null) {
$this->configuration = Util::getMeilisearchConfiguration();
}
return $this->configuration;
}
/**
* @param TypoScriptConfiguration $configuration
*/
public function setConfiguration(TypoScriptConfiguration $configuration)
{
$this->configuration = $configuration;
}
/**
* Returns the cleaned indexable content from the page's HTML markup.
*
* The content is cleaned from HTML tags and control chars Meilisearch could
* stumble on.
*
* @return string Indexable, cleaned content ready for indexing.
*/
public function getIndexableContent()
{
// @extensionScannerIgnoreLine
$content = self::cleanContent($this->content);
$content = trim($content);
return $content;
}
/**
* Strips html tags, and tab, new-line, carriage-return, &nbsp; whitespace
* characters.
*
* @param string $content String to clean
* @return string String cleaned from tags and special whitespace characters
*/
public static function cleanContent($content)
{
$content = self::stripControlCharacters($content);
// remove Javascript
$content = preg_replace('@<script[^>]*>.*?<\/script>@msi', '', $content);
// remove internal CSS styles
$content = preg_replace('@<style[^>]*>.*?<\/style>@msi', '', $content);
// prevents concatenated words when stripping tags afterwards
$content = str_replace(['<', '>'], [' <', '> '], $content);
$content = str_replace(["\t", "\n", "\r", '&nbsp;'], ' ', $content);
$content = strip_tags($content);
$content = html_entity_decode($content, ENT_QUOTES, 'UTF-8');
$content = self::stripUnicodeRanges($content);
$content = preg_replace('/\s{2,}/u', ' ', $content);
$content = trim($content);
return $content;
}
/**
* Strips control characters that cause Jetty/Meilisearch to fail.
*
* @param string $content the content to sanitize
* @return string the sanitized content
* @see http://w3.org/International/questions/qa-forms-utf-8.html
*/
public static function stripControlCharacters($content)
{
// Printable utf-8 does not include any of these chars below x7F
return preg_replace('@[\x00-\x08\x0B\x0C\x0E-\x1F]@', ' ', $content);
}
/**
* Strips unusable unicode ranges
*
* @param string $content Content to sanitize
* @return string Sanitized content
*/
public static function stripUnicodeRanges($content)
{
foreach (self::$stripUnicodeRanges as $range) {
$content = self::stripUnicodeRange($content, $range[0], $range[1]);
}
return $content;
}
/**
* Strips a UTF-8 character range
*
* @param string $content Content to sanitize
* @param string $start Unicode range start character as uppercase hexadecimal string
* @param string $end Unicode range end character as uppercase hexadecimal string
* @return string Sanitized content
*/
public static function stripUnicodeRange($content, $start, $end)
{
return preg_replace('/[\x{' . $start . '}-\x{' . $end . '}]/u', '',
$content);
}
/**
* Shortcut method to retrieve the raw content marked for indexing.
*
* @return string Content marked for indexing.
*/
public function getContentMarkedForIndexing()
{
// @extensionScannerIgnoreLine
return $this->content;
}
/**
* Extracts HTML tag content from tags in the content marked for indexing.
*
* @return array A mapping of Meilisearch document field names to content found in defined tags.
*/
public function getTagContent()
{
$result = [];
$matches = [];
$content = $this->getContentMarkedForIndexing();
// strip all ignored tags
$content = strip_tags(
$content,
'<' . implode('><', array_keys($this->tagToFieldMapping)) . '>'
);
preg_match_all(
'@<(' . implode('|',
array_keys($this->tagToFieldMapping)) . ')[^>]*>(.*)</\1>@Ui',
$content,
$matches
);
foreach ($matches[1] as $key => $tag) {
// We don't want to index links auto-generated by the url filter.
$pattern = '@(?:http://|https://|ftp://|mailto:|smb://|afp://|file://|gopher://|news://|ssl://|sslv2://|sslv3://|tls://|tcp://|udp://|www\.)[a-zA-Z0-9]+@';
if ($tag != 'a' || !preg_match($pattern, $matches[2][$key])) {
$fieldName = $this->tagToFieldMapping[$tag];
$hasContentForFieldName = empty($result[$fieldName]);
$separator = ($hasContentForFieldName) ? '' : ' ';
$result[$fieldName] .= $separator . $matches[2][$key];
}
}
return $result;
}
}