2016-01-04 10:45:54 +01:00
|
|
|
<?php
|
|
|
|
|
2018-12-03 01:10:39 +01:00
|
|
|
use Shaarli\Bookmark\LinkDB;
|
|
|
|
|
2016-01-04 10:45:54 +01:00
|
|
|
/**
|
2017-09-30 11:04:13 +02:00
|
|
|
* Get cURL callback function for CURLOPT_WRITEFUNCTION
|
2016-01-04 10:45:54 +01:00
|
|
|
*
|
2017-09-30 11:04:13 +02:00
|
|
|
* @param string $charset to extract from the downloaded page (reference)
|
|
|
|
* @param string $title to extract from the downloaded page (reference)
|
2019-06-08 13:59:19 +02:00
|
|
|
* @param string $description to extract from the downloaded page (reference)
|
|
|
|
* @param string $keywords to extract from the downloaded page (reference)
|
|
|
|
* @param bool $retrieveDescription Automatically tries to retrieve description and keywords from HTML content
|
2018-12-03 01:35:14 +01:00
|
|
|
* @param string $curlGetInfo Optionally overrides curl_getinfo function
|
2016-01-04 10:45:54 +01:00
|
|
|
*
|
2017-09-30 11:04:13 +02:00
|
|
|
* @return Closure
|
2016-01-04 10:45:54 +01:00
|
|
|
*/
|
2019-06-08 13:59:19 +02:00
|
|
|
function get_curl_download_callback(
|
|
|
|
&$charset,
|
|
|
|
&$title,
|
|
|
|
&$description,
|
|
|
|
&$keywords,
|
|
|
|
$retrieveDescription,
|
|
|
|
$curlGetInfo = 'curl_getinfo'
|
|
|
|
) {
|
2018-05-01 16:44:51 +02:00
|
|
|
$isRedirected = false;
|
2019-06-08 13:59:19 +02:00
|
|
|
$currentChunk = 0;
|
|
|
|
$foundChunk = null;
|
|
|
|
|
2017-09-30 11:04:13 +02:00
|
|
|
/**
|
|
|
|
* cURL callback function for CURLOPT_WRITEFUNCTION (called during the download).
|
|
|
|
*
|
|
|
|
* While downloading the remote page, we check that the HTTP code is 200 and content type is 'html/text'
|
|
|
|
* Then we extract the title and the charset and stop the download when it's done.
|
|
|
|
*
|
|
|
|
* @param resource $ch cURL resource
|
|
|
|
* @param string $data chunk of data being downloaded
|
|
|
|
*
|
|
|
|
* @return int|bool length of $data or false if we need to stop the download
|
|
|
|
*/
|
2019-06-08 13:59:19 +02:00
|
|
|
return function (&$ch, $data) use (
|
|
|
|
$retrieveDescription,
|
|
|
|
$curlGetInfo,
|
|
|
|
&$charset,
|
|
|
|
&$title,
|
|
|
|
&$description,
|
|
|
|
&$keywords,
|
|
|
|
&$isRedirected,
|
|
|
|
&$currentChunk,
|
|
|
|
&$foundChunk
|
|
|
|
) {
|
|
|
|
$currentChunk++;
|
2017-09-30 11:04:13 +02:00
|
|
|
$responseCode = $curlGetInfo($ch, CURLINFO_RESPONSE_CODE);
|
2018-05-01 16:44:51 +02:00
|
|
|
if (!empty($responseCode) && in_array($responseCode, [301, 302])) {
|
|
|
|
$isRedirected = true;
|
|
|
|
return strlen($data);
|
|
|
|
}
|
|
|
|
if (!empty($responseCode) && $responseCode !== 200) {
|
2017-09-30 11:04:13 +02:00
|
|
|
return false;
|
|
|
|
}
|
2018-05-01 16:44:51 +02:00
|
|
|
// After a redirection, the content type will keep the previous request value
|
|
|
|
// until it finds the next content-type header.
|
|
|
|
if (! $isRedirected || strpos(strtolower($data), 'content-type') !== false) {
|
|
|
|
$contentType = $curlGetInfo($ch, CURLINFO_CONTENT_TYPE);
|
|
|
|
}
|
2017-09-30 11:04:13 +02:00
|
|
|
if (!empty($contentType) && strpos($contentType, 'text/html') === false) {
|
|
|
|
return false;
|
|
|
|
}
|
2018-05-01 16:44:51 +02:00
|
|
|
if (!empty($contentType) && empty($charset)) {
|
2017-09-30 11:04:13 +02:00
|
|
|
$charset = header_extract_charset($contentType);
|
|
|
|
}
|
|
|
|
if (empty($charset)) {
|
|
|
|
$charset = html_extract_charset($data);
|
|
|
|
}
|
|
|
|
if (empty($title)) {
|
|
|
|
$title = html_extract_title($data);
|
2019-06-08 13:59:19 +02:00
|
|
|
$foundChunk = ! empty($title) ? $currentChunk : $foundChunk;
|
|
|
|
}
|
|
|
|
if ($retrieveDescription && empty($description)) {
|
|
|
|
$description = html_extract_tag('description', $data);
|
|
|
|
$foundChunk = ! empty($description) ? $currentChunk : $foundChunk;
|
2017-09-30 11:04:13 +02:00
|
|
|
}
|
2019-06-08 13:59:19 +02:00
|
|
|
if ($retrieveDescription && empty($keywords)) {
|
|
|
|
$keywords = html_extract_tag('keywords', $data);
|
|
|
|
if (! empty($keywords)) {
|
|
|
|
$foundChunk = $currentChunk;
|
|
|
|
// Keywords use the format tag1, tag2 multiple words, tag
|
|
|
|
// So we format them to match Shaarli's separator and glue multiple words with '-'
|
|
|
|
$keywords = implode(' ', array_map(function($keyword) {
|
|
|
|
return implode('-', preg_split('/\s+/', trim($keyword)));
|
|
|
|
}, explode(',', $keywords)));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2017-09-30 11:04:13 +02:00
|
|
|
// We got everything we want, stop the download.
|
2019-06-08 13:59:19 +02:00
|
|
|
// If we already found either the title, description or keywords,
|
|
|
|
// it's highly unlikely that we'll found the other metas further than
|
|
|
|
// in the same chunk of data or the next one. So we also stop the download after that.
|
|
|
|
if ((!empty($responseCode) && !empty($contentType) && !empty($charset)) && $foundChunk !== null
|
|
|
|
&& (! $retrieveDescription
|
|
|
|
|| $foundChunk < $currentChunk
|
|
|
|
|| (!empty($title) && !empty($description) && !empty($keywords))
|
|
|
|
)
|
|
|
|
) {
|
2017-09-30 11:04:13 +02:00
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
return strlen($data);
|
|
|
|
};
|
2016-01-04 10:45:54 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2017-09-30 11:04:13 +02:00
|
|
|
* Extract title from an HTML document.
|
2016-01-04 10:45:54 +01:00
|
|
|
*
|
2017-09-30 11:04:13 +02:00
|
|
|
* @param string $html HTML content where to look for a title.
|
2016-01-04 10:45:54 +01:00
|
|
|
*
|
2017-09-30 11:04:13 +02:00
|
|
|
* @return bool|string Extracted title if found, false otherwise.
|
2016-01-04 10:45:54 +01:00
|
|
|
*/
|
2017-09-30 11:04:13 +02:00
|
|
|
function html_extract_title($html)
|
2016-01-04 10:45:54 +01:00
|
|
|
{
|
2017-09-30 11:04:13 +02:00
|
|
|
if (preg_match('!<title.*?>(.*?)</title>!is', $html, $matches)) {
|
|
|
|
return trim(str_replace("\n", '', $matches[1]));
|
2016-01-04 10:45:54 +01:00
|
|
|
}
|
2017-09-30 11:04:13 +02:00
|
|
|
return false;
|
2016-01-04 10:45:54 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2017-09-30 11:04:13 +02:00
|
|
|
* Extract charset from HTTP header if it's defined.
|
2016-01-04 10:45:54 +01:00
|
|
|
*
|
2017-09-30 11:04:13 +02:00
|
|
|
* @param string $header HTTP header Content-Type line.
|
2016-01-04 10:45:54 +01:00
|
|
|
*
|
|
|
|
* @return bool|string Charset string if found (lowercase), false otherwise.
|
|
|
|
*/
|
2017-09-30 11:04:13 +02:00
|
|
|
function header_extract_charset($header)
|
2016-01-04 10:45:54 +01:00
|
|
|
{
|
2017-09-30 11:04:13 +02:00
|
|
|
preg_match('/charset="?([^; ]+)/i', $header, $match);
|
|
|
|
if (! empty($match[1])) {
|
|
|
|
return strtolower(trim($match[1]));
|
2016-01-04 10:45:54 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Extract charset HTML content (tag <meta charset>).
|
|
|
|
*
|
|
|
|
* @param string $html HTML content where to look for charset.
|
|
|
|
*
|
|
|
|
* @return bool|string Charset string if found, false otherwise.
|
|
|
|
*/
|
|
|
|
function html_extract_charset($html)
|
|
|
|
{
|
|
|
|
// Get encoding specified in HTML header.
|
2016-04-06 22:00:52 +02:00
|
|
|
preg_match('#<meta .*charset=["\']?([^";\'>/]+)["\']? */?>#Usi', $html, $enc);
|
2016-01-04 10:45:54 +01:00
|
|
|
if (!empty($enc[1])) {
|
|
|
|
return strtolower($enc[1]);
|
|
|
|
}
|
|
|
|
|
|
|
|
return false;
|
|
|
|
}
|
2016-05-11 00:05:22 +02:00
|
|
|
|
2019-06-08 13:59:19 +02:00
|
|
|
/**
|
|
|
|
* Extract meta tag from HTML content in either:
|
|
|
|
* - OpenGraph: <meta property="og:[tag]" ...>
|
|
|
|
* - Meta tag: <meta name="[tag]" ...>
|
|
|
|
*
|
|
|
|
* @param string $tag Name of the tag to retrieve.
|
|
|
|
* @param string $html HTML content where to look for charset.
|
|
|
|
*
|
|
|
|
* @return bool|string Charset string if found, false otherwise.
|
|
|
|
*/
|
|
|
|
function html_extract_tag($tag, $html)
|
|
|
|
{
|
|
|
|
$propertiesKey = ['property', 'name', 'itemprop'];
|
|
|
|
$properties = implode('|', $propertiesKey);
|
|
|
|
// Try to retrieve OpenGraph image.
|
|
|
|
$ogRegex = '#<meta[^>]+(?:'. $properties .')=["\']?(?:og:)?'. $tag .'["\'\s][^>]*content=["\']?(.*?)["\'/>]#';
|
|
|
|
// If the attributes are not in the order property => content (e.g. Github)
|
|
|
|
// New regex to keep this readable... more or less.
|
|
|
|
$ogRegexReverse = '#<meta[^>]+content=["\']([^"\']+)[^>]+(?:'. $properties .')=["\']?(?:og)?:'. $tag .'["\'\s/>]#';
|
|
|
|
|
|
|
|
if (preg_match($ogRegex, $html, $matches) > 0
|
|
|
|
|| preg_match($ogRegexReverse, $html, $matches) > 0
|
|
|
|
) {
|
|
|
|
return $matches[1];
|
|
|
|
}
|
|
|
|
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2016-05-11 00:05:22 +02:00
|
|
|
/**
|
|
|
|
* Count private links in given linklist.
|
|
|
|
*
|
2016-10-20 11:31:52 +02:00
|
|
|
* @param array|Countable $links Linklist.
|
2016-05-11 00:05:22 +02:00
|
|
|
*
|
|
|
|
* @return int Number of private links.
|
|
|
|
*/
|
|
|
|
function count_private($links)
|
|
|
|
{
|
|
|
|
$cpt = 0;
|
|
|
|
foreach ($links as $link) {
|
2017-01-06 18:54:29 +01:00
|
|
|
if ($link['private']) {
|
|
|
|
$cpt += 1;
|
|
|
|
}
|
2016-05-11 00:05:22 +02:00
|
|
|
}
|
2016-05-10 23:18:04 +02:00
|
|
|
|
2016-05-11 00:05:22 +02:00
|
|
|
return $cpt;
|
|
|
|
}
|
2016-05-10 23:18:04 +02:00
|
|
|
|
|
|
|
/**
|
|
|
|
* In a string, converts URLs to clickable links.
|
|
|
|
*
|
|
|
|
* @param string $text input string.
|
|
|
|
*
|
|
|
|
* @return string returns $text with all links converted to HTML links.
|
|
|
|
*
|
|
|
|
* @see Function inspired from http://www.php.net/manual/en/function.preg-replace.php#85722
|
|
|
|
*/
|
2019-02-09 13:52:12 +01:00
|
|
|
function text2clickable($text)
|
2016-05-10 23:18:04 +02:00
|
|
|
{
|
2017-09-29 18:52:38 +02:00
|
|
|
$regex = '!(((?:https?|ftp|file)://|apt:|magnet:)\S+[a-z0-9\(\)]/?)!si';
|
2019-02-09 13:52:12 +01:00
|
|
|
return preg_replace($regex, '<a href="$1">$1</a>', $text);
|
2016-05-10 23:18:04 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Auto-link hashtags.
|
|
|
|
*
|
|
|
|
* @param string $description Given description.
|
|
|
|
* @param string $indexUrl Root URL.
|
|
|
|
*
|
|
|
|
* @return string Description with auto-linked hashtags.
|
|
|
|
*/
|
|
|
|
function hashtag_autolink($description, $indexUrl = '')
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* To support unicode: http://stackoverflow.com/a/35498078/1484919
|
|
|
|
* \p{Pc} - to match underscore
|
|
|
|
* \p{N} - numeric character in any script
|
|
|
|
* \p{L} - letter from any language
|
|
|
|
* \p{Mn} - any non marking space (accents, umlauts, etc)
|
|
|
|
*/
|
|
|
|
$regex = '/(^|\s)#([\p{Pc}\p{N}\p{L}\p{Mn}]+)/mui';
|
|
|
|
$replacement = '$1<a href="'. $indexUrl .'?addtag=$2" title="Hashtag $2">#$2</a>';
|
|
|
|
return preg_replace($regex, $replacement, $description);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* This function inserts where relevant so that multiple spaces are properly displayed in HTML
|
|
|
|
* even in the absence of <pre> (This is used in description to keep text formatting).
|
|
|
|
*
|
|
|
|
* @param string $text input text.
|
|
|
|
*
|
|
|
|
* @return string formatted text.
|
|
|
|
*/
|
|
|
|
function space2nbsp($text)
|
|
|
|
{
|
|
|
|
return preg_replace('/(^| ) /m', '$1 ', $text);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* Format Shaarli's description
|
|
|
|
*
|
|
|
|
* @param string $description shaare's description.
|
2016-10-20 11:31:52 +02:00
|
|
|
* @param string $indexUrl URL to Shaarli's index.
|
2017-11-07 20:23:58 +01:00
|
|
|
|
2016-05-10 23:18:04 +02:00
|
|
|
* @return string formatted description.
|
|
|
|
*/
|
2019-02-09 13:52:12 +01:00
|
|
|
function format_description($description, $indexUrl = '')
|
2018-10-13 00:19:03 +02:00
|
|
|
{
|
2019-02-09 13:52:12 +01:00
|
|
|
return nl2br(space2nbsp(hashtag_autolink(text2clickable($description), $indexUrl)));
|
2016-05-10 23:18:04 +02:00
|
|
|
}
|
2016-11-28 18:24:15 +01:00
|
|
|
|
|
|
|
/**
|
|
|
|
* Generate a small hash for a link.
|
|
|
|
*
|
|
|
|
* @param DateTime $date Link creation date.
|
|
|
|
* @param int $id Link ID.
|
|
|
|
*
|
|
|
|
* @return string the small hash generated from link data.
|
|
|
|
*/
|
|
|
|
function link_small_hash($date, $id)
|
|
|
|
{
|
|
|
|
return smallHash($date->format(LinkDB::LINK_DATE_FORMAT) . $id);
|
|
|
|
}
|
2019-02-09 14:13:08 +01:00
|
|
|
|
|
|
|
/**
|
|
|
|
* Returns whether or not the link is an internal note.
|
|
|
|
* Its URL starts by `?` because it's actually a permalink.
|
|
|
|
*
|
|
|
|
* @param string $linkUrl
|
|
|
|
*
|
|
|
|
* @return bool true if internal note, false otherwise.
|
|
|
|
*/
|
|
|
|
function is_note($linkUrl)
|
|
|
|
{
|
|
|
|
return isset($linkUrl[0]) && $linkUrl[0] === '?';
|
|
|
|
}
|