Merge pull request #13933 from annando/fix2

Accidentally merged changes are reverted
This commit is contained in:
Michael Vogel 2024-02-24 12:40:19 +01:00 committed by GitHub
commit 1c5681c199
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
7 changed files with 334 additions and 50 deletions

View file

@ -22,6 +22,8 @@
namespace Friendica\Content;
use DOMDocument;
use DOMNode;
use DOMText;
use DOMXPath;
use Exception;
use Friendica\Core\Cache\Enum\Duration;
@ -30,10 +32,10 @@ use Friendica\Core\Renderer;
use Friendica\Database\Database;
use Friendica\Database\DBA;
use Friendica\DI;
use Friendica\Model\Post;
use Friendica\Network\HTTPClient\Client\HttpClientAccept;
use Friendica\Util\DateTimeFormat;
use Friendica\Util\Network;
use Friendica\Util\ParseUrl;
use Friendica\Util\Proxy;
use Friendica\Util\Strings;
@ -47,15 +49,32 @@ use Friendica\Util\Strings;
*/
class OEmbed
{
/**
* Callback for fetching URL, checking allowance and returning formatted HTML
*
* @param array $matches
* @return string Formatted HTML
*/
public static function replaceCallback(array $matches): string
{
$embedurl = $matches[1];
$j = self::fetchURL($embedurl, !self::isAllowedURL($embedurl));
$s = self::formatObject($j);
return $s;
}
/**
* Get data from an URL to embed its content.
*
* @param string $embedurl The URL from which the data should be fetched.
* @param bool $no_rich_type If set to true rich type content won't be fetched.
* @param bool $use_parseurl Use the "ParseUrl" functionality to add additional data
*
* @return \Friendica\Object\OEmbed
* @throws \Friendica\Network\HTTPException\InternalServerErrorException
*/
public static function fetchURL(string $embedurl): \Friendica\Object\OEmbed
public static function fetchURL(string $embedurl, bool $no_rich_type = false, bool $use_parseurl = true): \Friendica\Object\OEmbed
{
$embedurl = trim($embedurl, '\'"');
@ -100,7 +119,7 @@ class OEmbed
$href = str_replace(['http://www.youtube.com/', 'http://player.vimeo.com/'],
['https://www.youtube.com/', 'https://player.vimeo.com/'], $href);
$result = DI::httpClient()->fetchFull($href . '&maxwidth=' . $a->getThemeInfoValue('videowidth'));
if ($result->isSuccess()) {
if ($result->getReturnCode() === 200) {
$json_string = $result->getBodyString();
break;
}
@ -137,6 +156,60 @@ class OEmbed
$oembed->html = str_replace(['http://www.youtube.com/', 'http://player.vimeo.com/'], ['https://www.youtube.com/', 'https://player.vimeo.com/'], $oembed->html);
}
// Improve the OEmbed data with data from OpenGraph, Twitter cards and other sources
if ($use_parseurl) {
$data = ParseUrl::getSiteinfoCached($embedurl, false);
if (($oembed->type == 'error') && empty($data['title']) && empty($data['text'])) {
return $oembed;
}
if ($no_rich_type || ($oembed->type == 'error')) {
$oembed->html = '';
$oembed->type = $data['type'];
if ($oembed->type == 'photo') {
if (!empty($data['images'])) {
$oembed->url = $data['images'][0]['src'];
$oembed->width = $data['images'][0]['width'];
$oembed->height = $data['images'][0]['height'];
} else {
$oembed->type = 'link';
}
}
}
if (!empty($data['title'])) {
$oembed->title = $data['title'];
}
if (!empty($data['text'])) {
$oembed->description = $data['text'];
}
if (!empty($data['publisher_name'])) {
$oembed->provider_name = $data['publisher_name'];
}
if (!empty($data['publisher_url'])) {
$oembed->provider_url = $data['publisher_url'];
}
if (!empty($data['author_name'])) {
$oembed->author_name = $data['author_name'];
}
if (!empty($data['author_url'])) {
$oembed->author_url = $data['author_url'];
}
if (!empty($data['images']) && ($oembed->type != 'photo')) {
$oembed->thumbnail_url = $data['images'][0]['src'];
$oembed->thumbnail_width = $data['images'][0]['width'];
$oembed->thumbnail_height = $data['images'][0]['height'];
}
}
Hook::callAll('oembed_fetch_url', $embedurl, $oembed);
return $oembed;
@ -146,20 +219,12 @@ class OEmbed
* Returns a formatted string from OEmbed object
*
* @param \Friendica\Object\OEmbed $oembed
* @param int $uriid
* @return string
*/
private static function formatObject(\Friendica\Object\OEmbed $oembed, int $uriid): string
private static function formatObject(\Friendica\Object\OEmbed $oembed): string
{
$ret = '<div class="oembed ' . $oembed->type . '">';
if (!self::isAllowedURL($oembed->url)) {
$oembed->html = '';
if ($oembed->type == 'rich') {
$oembed->type = 'link';
}
}
switch ($oembed->type) {
case 'video':
if ($oembed->thumbnail_url) {
@ -176,15 +241,15 @@ class OEmbed
'$escapedhtml' => base64_encode($oembed->html),
'$tw' => $tw,
'$th' => $th,
'$turl' => Post\Link::getByLink($uriid, $oembed->thumbnail_url, Proxy::SIZE_MEDIUM),
'$turl' => $oembed->thumbnail_url,
]);
} else {
$ret .= Proxy::proxifyHtml($oembed->html);
$ret = $oembed->html;
}
break;
case 'photo':
$ret .= '<img width="' . $oembed->width . '" src="' . Post\Link::getByLink($uriid, $oembed->url, Proxy::SIZE_MEDIUM) . '">';
$ret .= '<img width="' . $oembed->width . '" src="' . Proxy::proxifyUrl($oembed->url) . '">';
break;
case 'link':
@ -229,14 +294,6 @@ class OEmbed
$ret .= '<a href="' . $oembed->embed_url . '" rel="oembed">' . $oembed->embed_url . '</a>';
}
$ret .= "</h4>";
if ($oembed->type == 'link') {
if (!empty($oembed->thumbnail_url)) {
$ret .= '<img width="' . $oembed->width . '" src="' . Post\Link::getByLink($uriid, $oembed->thumbnail_url, Proxy::SIZE_MEDIUM) . '">';
}
if (!empty($oembed->description)) {
$ret .= '<p>' . $oembed->description . '</p>';
}
}
} elseif (!strpos($oembed->html, $oembed->embed_url)) {
// add <a> for html2bbcode conversion
$ret .= '<a href="' . $oembed->embed_url . '" rel="oembed">' . $oembed->title . '</a>';
@ -251,20 +308,51 @@ class OEmbed
* Converts BBCode to HTML code
*
* @param string $text
* @param int $uriid
* @return string
*/
public static function BBCode2HTML(string $text, int $uriid): string
public static function BBCode2HTML(string $text): string
{
if (!preg_match_all("/\[embed\](.+?)\[\/embed\]/is", $text, $matches, PREG_SET_ORDER)) {
if (DI::config()->get('system', 'no_oembed')) {
return preg_replace("/\[embed\](.+?)\[\/embed\]/is", "<!-- oembed $1 --><i>" . DI::l10n()->t('Embedding disabled') . " : $1</i><!-- /oembed $1 -->", $text);
}
return preg_replace_callback("/\[embed\](.+?)\[\/embed\]/is", [self::class, 'replaceCallback'], $text);
}
/**
* Find <span class='oembed'>..<a href='url' rel='oembed'>..</a></span>
* and replace it with [embed]url[/embed]
*
* @param string $text
* @return string
*/
public static function HTML2BBCode(string $text): string
{
// start parser only if 'oembed' is in text
if (strpos($text, 'oembed')) {
// convert non ascii chars to html entities
$html_text = mb_convert_encoding($text, 'HTML-ENTITIES', mb_detect_encoding($text));
// If it doesn't parse at all, just return the text.
$dom = new DOMDocument();
if (!@$dom->loadHTML($html_text)) {
return $text;
}
$xpath = new DOMXPath($dom);
$xattr = self::buildXPath('class', 'oembed');
$entries = $xpath->query("//div[$xattr]");
$xattr = "@rel='oembed'"; //oe_build_xpath("rel","oembed");
foreach ($entries as $e) {
$href = $xpath->evaluate("a[$xattr]/@href", $e)->item(0)->nodeValue;
if (!is_null($href)) {
$e->parentNode->replaceChild(new DOMText('[embed]' . $href . '[/embed]'), $e);
}
}
return self::getInnerHTML($dom->getElementsByTagName('body')->item(0));
} else {
return $text;
}
foreach ($matches as $match) {
$data = self::fetchURL($match[1]);
$text = str_replace($match[0], self::formatObject($data, $uriid), $text);
}
return $text;
}
/**
@ -274,7 +362,7 @@ class OEmbed
* @return boolean
* @throws \Friendica\Network\HTTPException\InternalServerErrorException
*/
private static function isAllowedURL(string $url): bool
public static function isAllowedURL(string $url): bool
{
if (!DI::config()->get('system', 'no_oembed_rich_content')) {
return true;
@ -299,13 +387,12 @@ class OEmbed
* Returns a formatted HTML code from given URL and sets optional title
*
* @param string $url URL to fetch
* @param string $title title (default: what comes from OEmbed object)
* @param int $uriid
* @param string $title Optional title (default: what comes from OEmbed object)
* @return string Formatted HTML
*/
public static function getHTML(string $url, string $title, int $uriid): string
public static function getHTML(string $url, string $title = ''): string
{
$o = self::fetchURL($url);
$o = self::fetchURL($url, !self::isAllowedURL($url));
if (!is_object($o) || property_exists($o, 'type') && $o->type == 'error') {
throw new Exception('OEmbed failed for URL: ' . $url);
@ -315,8 +402,74 @@ class OEmbed
$o->title = $title;
}
$html = self::formatObject($o, $uriid);
$html = self::formatObject($o);
return $html;
}
/**
* Generates the iframe HTML for an oembed attachment.
*
* Width and height are given by the remote, and are regularly too small for
* the generated iframe.
*
* The width is entirely discarded for the actual width of the post, while fixed
* height is used as a starting point before the inevitable resizing.
*
* Since the iframe is automatically resized on load, there are no need for ugly
* and impractical scrollbars.
*
* @todo This function is currently unused until someone™ adds support for a separate OEmbed domain
*
* @param string $src Original remote URL to embed
* @param string $width
* @param string $height
* @return string Formatted HTML
*
* @throws \Friendica\Network\HTTPException\InternalServerErrorException
* @see oembed_format_object()
*/
private static function iframe(string $src, string $width, string $height): string
{
if (!$height || strstr($height, '%')) {
$height = '200';
}
$width = '100%';
$src = DI::baseUrl() . '/oembed/' . Strings::base64UrlEncode($src);
return '<iframe onload="resizeIframe(this);" class="embed_rich" height="' . $height . '" width="' . $width . '" src="' . $src . '" allowfullscreen scrolling="no" frameborder="no">' . DI::l10n()->t('Embedded content') . '</iframe>';
}
/**
* Generates attribute search XPath string
*
* Generates an XPath query to select elements whose provided attribute contains
* the provided value in a space-separated list.
*
* @param string $attr Name of the attribute to search
* @param string $value Value to search in a space-separated list
* @return string
*/
private static function buildXPath(string $attr, $value): string
{
// https://www.westhoffswelt.de/blog/2009/6/9/select-html-elements-with-more-than-one-css-class-using-xpath
return "contains(normalize-space(@$attr), ' $value ') or substring(normalize-space(@$attr), 1, string-length('$value') + 1) = '$value ' or substring(normalize-space(@$attr), string-length(@$attr) - string-length('$value')) = ' $value' or @$attr = '$value'";
}
/**
* Returns the inner XML string of a provided DOMNode
*
* @param DOMNode $node
* @return string
*/
private static function getInnerHTML(DOMNode $node): string
{
$innerHTML = '';
$children = $node->childNodes;
foreach ($children as $child) {
$innerHTML .= $child->ownerDocument->saveXML($child);
}
return $innerHTML;
}
}

View file

@ -452,7 +452,11 @@ class BBCode
$return = '';
try {
$return = OEmbed::getHTML($data['url'], $data['title'], $uriid);
if ($tryoembed && OEmbed::isAllowedURL($data['url'])) {
$return = OEmbed::getHTML($data['url'], $data['title']);
} else {
throw new Exception('OEmbed is disabled for this attachment.');
}
} catch (Exception $e) {
$data['title'] = ($data['title'] ?? '') ?: $data['url'];
@ -1354,12 +1358,12 @@ class BBCode
* $match[1] = $url
* $match[2] = $title or absent
*/
$try_oembed_callback = function (array $match) use ($uriid) {
$try_oembed_callback = function (array $match) {
$url = $match[1];
$title = $match[2] ?? '';
try {
$return = OEmbed::getHTML($url, $title, $uriid);
$return = OEmbed::getHTML($url, $title);
} catch (Exception $ex) {
$return = $match[0];
}
@ -1377,7 +1381,6 @@ class BBCode
$text = preg_replace("#\[(\w*)](\n*)#ism", '$2[$1]', $text);
$text = preg_replace("#(\n*)\[/(\w*)]#ism", '[/$2]$1', $text);
// oembed
// Extract the private images which use data urls since preg has issues with
// large data sizes. Stash them away while we do bbcode conversion, and then put them back
// in after we've done all the regex matching. We cannot use any preg functions to do this.
@ -1807,7 +1810,7 @@ class BBCode
}
// oembed tag
$text = OEmbed::BBCode2HTML($text, $uriid);
$text = OEmbed::BBCode2HTML($text);
// Avoid triple linefeeds through oembed
$text = str_replace("<br style='clear:left'></span><br><br>", "<br style='clear:left'></span><br>", $text);
@ -2055,6 +2058,9 @@ class BBCode
// Default iframe allowed domains/path
$allowedIframeDomains = [
DI::baseUrl()->getHost()
. (DI::baseUrl()->getPath() ? '/' . DI::baseUrl()->getPath() : '')
. '/oembed/', # The path part has to change with the source in Content\Oembed::iframe
'www.youtube.com/embed/',
'player.vimeo.com/video/',
];

View file

@ -365,7 +365,7 @@ class Media
*/
private static function addPage(array $media): array
{
$data = ParseUrl::getSiteinfoCached($media['url']);
$data = ParseUrl::getSiteinfoCached($media['url'], false);
$media['preview'] = $data['images'][0]['src'] ?? null;
$media['preview-height'] = $data['images'][0]['height'] ?? null;
$media['preview-width'] = $data['images'][0]['width'] ?? null;

74
src/Module/Oembed.php Normal file
View file

@ -0,0 +1,74 @@
<?php
/**
* @copyright Copyright (C) 2010-2024, the Friendica project
*
* @license GNU AGPL version 3 or any later version
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License as
* published by the Free Software Foundation, either version 3 of the
* License, or (at your option) any later version.
*
* 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. See the
* GNU Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see <https://www.gnu.org/licenses/>.
*
*/
namespace Friendica\Module;
use Friendica\BaseModule;
use Friendica\Content;
use Friendica\Core\System;
use Friendica\DI;
use Friendica\Util\Strings;
/**
* Oembed module
*
* Displays stored embed content based on a base64 hash of a remote URL
*
* Example: /oembed/aHR0cHM6Ly9...
*
* @author Hypolite Petovan <hypolite@mrpetovan.com>
*/
class Oembed extends BaseModule
{
protected function content(array $request = []): string
{
// Unused form: /oembed/b2h?url=...
if (DI::args()->getArgv()[1] == 'b2h') {
$url = ["", trim(hex2bin($_REQUEST['url']))];
echo Content\OEmbed::replaceCallback($url);
System::exit();
}
// Unused form: /oembed/h2b?text=...
if (DI::args()->getArgv()[1] == 'h2b') {
$text = trim(hex2bin($_REQUEST['text']));
echo Content\OEmbed::HTML2BBCode($text);
System::exit();
}
// @TODO: Replace with parameter from router
if (DI::args()->getArgc() == 2) {
echo '<html><body>';
$url = Strings::base64UrlDecode(DI::args()->getArgv()[1]);
$j = Content\OEmbed::fetchURL($url);
// workaround for media.ccc.de (and any other endpoint that return size 0)
if (substr($j->html, 0, 7) == "<iframe" && strstr($j->html, 'width="0"')) {
$j->html = '<style>html,body{margin:0;padding:0;} iframe{width:100%;height:100%;}</style>' . $j->html;
$j->html = str_replace('width="0"', '', $j->html);
$j->html = str_replace('height="0"', '', $j->html);
}
echo $j->html;
echo '</body></html>';
}
System::exit();
}
}

View file

@ -23,6 +23,7 @@ namespace Friendica\Util;
use DOMDocument;
use DOMXPath;
use Friendica\Content\OEmbed;
use Friendica\Content\Text\HTML;
use Friendica\Protocol\HTTP\MediaType;
use Friendica\Core\Hook;
@ -98,6 +99,8 @@ class ParseUrl
* Search for cached embeddable data of an url otherwise fetch it
*
* @param string $url The url of the page which should be scraped
* @param bool $do_oembed The false option is used by the function fetch_oembed()
* to avoid endless loops
*
* @return array which contains needed data for embedding
* string 'url' => The url of the parsed page
@ -112,7 +115,7 @@ class ParseUrl
* @see ParseUrl::getSiteinfo() for more information about scraping
* embeddable content
*/
public static function getSiteinfoCached(string $url): array
public static function getSiteinfoCached(string $url, bool $do_oembed = true): array
{
if (empty($url)) {
return [
@ -123,13 +126,15 @@ class ParseUrl
$urlHash = hash('sha256', $url);
$parsed_url = DBA::selectFirst('parsed_url', ['content'], ['url_hash' => $urlHash]);
$parsed_url = DBA::selectFirst('parsed_url', ['content'],
['url_hash' => $urlHash, 'oembed' => $do_oembed]
);
if (!empty($parsed_url['content'])) {
$data = unserialize($parsed_url['content']);
return $data;
}
$data = self::getSiteinfo($url);
$data = self::getSiteinfo($url, $do_oembed);
$expires = $data['expires'];
@ -139,7 +144,7 @@ class ParseUrl
'parsed_url',
[
'url_hash' => $urlHash,
'oembed' => false,
'oembed' => $do_oembed,
'url' => $url,
'content' => serialize($data),
'created' => DateTimeFormat::utcNow(),
@ -162,6 +167,8 @@ class ParseUrl
* \<meta name="description" content="An awesome description"\>
*
* @param string $url The url of the page which should be scraped
* @param bool $do_oembed The false option is used by the function fetch_oembed()
* to avoid endless loops
* @param int $count Internal counter to avoid endless loops
*
* @return array which contains needed data for embedding
@ -187,7 +194,7 @@ class ParseUrl
* </body>
* @endverbatim
*/
public static function getSiteinfo(string $url, int $count = 1): array
public static function getSiteinfo(string $url, bool $do_oembed = true, int $count = 1): array
{
if (empty($url)) {
return [
@ -247,6 +254,41 @@ class ParseUrl
$body = $curlResult->getBodyString();
if ($do_oembed) {
$oembed_data = OEmbed::fetchURL($url, false, false);
if (!empty($oembed_data->type)) {
if (!in_array($oembed_data->type, ['error', 'rich', 'image', 'video', 'audio', ''])) {
$siteinfo['type'] = $oembed_data->type;
}
// See https://github.com/friendica/friendica/pull/5763#discussion_r217913178
if ($siteinfo['type'] != 'photo') {
if (!empty($oembed_data->title)) {
$siteinfo['title'] = trim($oembed_data->title);
}
if (!empty($oembed_data->description)) {
$siteinfo['text'] = trim($oembed_data->description);
}
if (!empty($oembed_data->author_name)) {
$siteinfo['author_name'] = trim($oembed_data->author_name);
}
if (!empty($oembed_data->author_url)) {
$siteinfo['author_url'] = Network::sanitizeUrl($oembed_data->author_url);
}
if (!empty($oembed_data->provider_name)) {
$siteinfo['publisher_name'] = trim($oembed_data->provider_name);
}
if (!empty($oembed_data->provider_url)) {
$siteinfo['publisher_url'] = Network::sanitizeUrl($oembed_data->provider_url);
}
if (!empty($oembed_data->thumbnail_url)) {
$siteinfo['image'] = $oembed_data->thumbnail_url;
}
}
}
}
$charset = '';
try {
// Look for a charset, first in headers
@ -309,7 +351,7 @@ class ParseUrl
}
}
if ($content != '') {
$siteinfo = self::getSiteinfo($content, ++$count);
$siteinfo = self::getSiteinfo($content, $do_oembed, ++$count);
return $siteinfo;
}
}

View file

@ -429,6 +429,10 @@ return [
// Don't do count calculations (currently only when showing photo albums).
'no_count' => false,
// no_oembed (Boolean)
// Don't use OEmbed to fetch more information about a link.
'no_oembed' => false,
// no_redirect_list (Array)
// List of domains where HTTP redirects should be ignored.
'no_redirect_list' => [],

View file

@ -557,6 +557,11 @@ return [
'/objects/{guid}[/{activity}]' => [Module\ActivityPub\Objects::class, [R::GET]],
'/oembed' => [
'/b2h' => [Module\Oembed::class, [R::GET]],
'/h2b' => [Module\Oembed::class, [R::GET]],
'/{hash}' => [Module\Oembed::class, [R::GET]],
],
'/outbox/{nickname}' => [Module\ActivityPub\Outbox::class, [R::GET, R::POST]],
'/owa' => [Module\Owa::class, [R::GET]],
'/openid' => [Module\Security\OpenID::class, [R::GET]],