forked from blallo/rss-bridge
6f24858124
This commit is a squash of all commits that fix coding styles for the new coding style policy. [ABCTabsBridge] Fix coding style [AcrimedBridge] Fix coding style [AllocineFRBridge] Fix coding style [AnimeUltimeBridge] Fix coding style [Arte7Bridge] Fix coding style [AskfmBridge] Fix coding style [BandcampBridge] Fix coding style [BastaBridge] Fix coding style [BlaguesDeMerdeBridge] Fix coding style [BooruprojectBridge] Fix coding style [CADBridge] Fix coding style [CNETBridge] Fix coding style [CastorusBridge] Fix coding style [CollegeDeFranceBridge] Fix coding style [CommonDreamsBridge] Fix coding style [CopieDoubleBridge] Fix coding style [CourrierInternationalBridge] Fix coding style [CpasbienBridge] Fix coding style [CryptomeBridge] Fix coding style [DailymotionBridge] Fix coding style [DanbooruBridge] Fix coding style [DansTonChatBridge] Fix coding style [DauphineLibereBridge] Fix coding style [DeveloppezDotComBridge] Fix coding style [DemoBridge] Fix coding style [DilbertBridge] Fix coding style [DuckDuckGoBridge] Fix coding style [DollbooruBridge] Fix coding style [EliteDangerousGalnetBridge] Fix coding style [ElsevierBridge] Fix coding style [EstCeQuonMetEnProdBridge] Fix coding style [EZTVBridge] Fix coding style [FacebookBridge] Fix coding style [FeedExpanderExampleBridge] Fix coding style [FB2Bridge] Fix coding style [FierPandaBridge] Fix coding style [FlickrBridge] Fix coding style [FootitoBridge] Fix coding style [FourchanBridge] Fix coding style [FuturaSciencesBridge] Fix coding style [GBAtempBridge] Fix coding style [GelbooruBridge] Fix coding style [GiphyBridge] Fix coding style [GithubIssueBridge] Fix coding style [GizmodoBridge] Fix coding style [GoComicsBridge] Fix coding style [GooglePlusPostBridge] Fix coding style [GoogleSearchBridge] Fix coding style [HDWallpapersBridge] Fix coding style [HentaiHavenBridge] Fix coding style [IdenticaBridge] Fix coding style [InstagramBridge] Fix coding style [IsoHuntBridge] Fix coding style [JapanExpoBridge] Fix coding style [KonachanBridge] Fix coding style [KoreusBridge] Fix coding style [KununuBridge] Fix coding style [LeBonCoinBridge] Fix coding style [LegifranceJOBBridge] Fix coding style [LeMondeInformatiqueBridge] Fix coding style [LesJoiesDuCodeBridge] Fix coding style [LichessBridge] Fix coding style [LinkedInCompanyBridge] Fix coding style [LolibooruBridge] Fix coding style [LWNprevBridge] Fix coding style [MangareaderBridge] Fix coding style [MilbooruBridge] Fix coding style [MixCloudBridge] Fix coding style [MoebooruBridge] Fix coding style [MondeDiploBridge] Fix coding style [MsnMondeBridge] Fix coding style [MspabooruBridge] Fix coding style [NasaApodBridge] Fix coding style [NeuviemeArtBridge] Fix coding style [NextgovBridge] Fix coding style [NextInpactBridge] Fix coding style [NiceMatinBridge] Fix coding style [NovelUpdatesBridge] Fix coding style [OpenClassroomsBridge] Fix coding style [ParuVenduImmoBridge] Fix coding style [PickyWallpapersBridge] Fix coding style [PinterestBridge] Fix coding style [PlanetLibreBridge] Fix coding style [ReadComicsBridge] Fix coding style [Releases3DSBridge] Fix coding style [ReporterreBridge] Fix coding style [RTBFBridge] Fix coding style [Rue89Bridge] Fix coding style [Rule34Bridge] Fix coding style [Rule34pahealBridge] Fix coding style [SafebooruBridge] Fix coding style [SakugabooruBridge] Fix coding style [ScmbBridge] Fix coding style [ScoopItBridge] Fix coding style [SensCritiqueBridge] Fix coding style [SexactuBridge] Fix coding style [ShanaprojectBridge] Fix coding style [Shimmie2Bridge] Fix coding style [SoundcloudBridge] Fix coding style [StripeAPIChangeLogBridge] Fix coding style [SuperbWallpapersBridge] Fix coding style [T411Bridge] Fix coding style [TagBoardBridge] Fix coding style [TbibBridge] Fix coding style [TheCodingLoveBridge] Fix coding style [TheHackerNewsBridge] Fix coding style [ThePirateBayBridge] Fix coding style [TheTVDBBridge] Fix coding style [Torrent9Bridge] Fix coding style [TwitterBridge] Fix coding style [UnsplashBridge] Fix coding style [ViadeoCompanyBridge] Fix coding style [VineBridge] Fix coding style [VkBridge] Fix coding style [WallpaperStopBridge] Fix coding style [WebfailBridge] Fix coding style [WeLiveSecurityBridge] Fix coding style [WhydBridge] Fix coding style [WikipediaBridge] Fix coding style [WordPressBridge] Fix coding style [WorldOfTanksBridge] Fix coding style [XbooruBridge] Fix coding style [YandereBridge] Fix coding style [YoutubeBridge] Fix coding style [ZDNetBridge] Fix coding style
249 lines
7 KiB
PHP
249 lines
7 KiB
PHP
<?php
|
|
class FB2Bridge extends BridgeAbstract {
|
|
|
|
const MAINTAINER = 'teromene';
|
|
const NAME = 'Facebook Alternate';
|
|
const URI = 'https://www.facebook.com/';
|
|
const CACHE_TIMEOUT = 1000;
|
|
const DESCRIPTION = 'Input a page title or a profile log. For a profile log,
|
|
please insert the parameter as follow : myExamplePage/132621766841117';
|
|
|
|
const PARAMETERS = array( array(
|
|
'u' => array(
|
|
'name' => 'Username',
|
|
'required' => true
|
|
)
|
|
));
|
|
|
|
public function collectData(){
|
|
|
|
function extractFromDelimiters($string, $start, $end){
|
|
if(strpos($string, $start) !== false){
|
|
$section_retrieved = substr($string, strpos($string, $start) + strlen($start));
|
|
$section_retrieved = substr($section_retrieved, 0, strpos($section_retrieved, $end));
|
|
return $section_retrieved;
|
|
}
|
|
|
|
return false;
|
|
}
|
|
|
|
//Utility function for cleaning a Facebook link
|
|
$unescape_fb_link = function($matches){
|
|
if(is_array($matches) && count($matches) > 1){
|
|
$link = $matches[1];
|
|
if(strpos($link, '/') === 0)
|
|
$link = self::URI . $link . '"';
|
|
if(strpos($link, 'facebook.com/l.php?u=') !== false)
|
|
$link = urldecode(extractFromDelimiters($link, 'facebook.com/l.php?u=', '&'));
|
|
return ' href="' . $link . '"';
|
|
}
|
|
};
|
|
|
|
//Utility function for converting facebook emoticons
|
|
$unescape_fb_emote = function($matches){
|
|
static $facebook_emoticons = array(
|
|
'smile' => ':)',
|
|
'frown' => ':(',
|
|
'tongue' => ':P',
|
|
'grin' => ':D',
|
|
'gasp' => ':O',
|
|
'wink' => ';)',
|
|
'pacman' => ':<',
|
|
'grumpy' => '>_<',
|
|
'unsure' => ':/',
|
|
'cry' => ':\'(',
|
|
'kiki' => '^_^',
|
|
'glasses' => '8-)',
|
|
'sunglasses' => 'B-)',
|
|
'heart' => '<3',
|
|
'devil' => ']:D',
|
|
'angel' => '0:)',
|
|
'squint' => '-_-',
|
|
'confused' => 'o_O',
|
|
'upset' => 'xD',
|
|
'colonthree' => ':3',
|
|
'like' => '👍');
|
|
$len = count($matches);
|
|
if ($len > 1)
|
|
for ($i = 1; $i < $len; $i++)
|
|
foreach ($facebook_emoticons as $name => $emote)
|
|
if ($matches[$i] === $name)
|
|
return $emote;
|
|
return $matches[0];
|
|
};
|
|
|
|
if($this->getInput('u') !== null){
|
|
$page = 'https://touch.facebook.com/' . $this->getInput('u');
|
|
$cookies = $this->getCookies($page);
|
|
$pageID = $this->getPageID($page, $cookies);
|
|
|
|
if($pageID === null){
|
|
echo <<<EOD
|
|
Unable to get the page id. You should consider getting the ID by hand, then importing it into FB2Bridge
|
|
EOD;
|
|
die();
|
|
}
|
|
}
|
|
|
|
//Build the string for the first request
|
|
$requestString = 'https://touch.facebook.com/pages_reaction_units/more/?page_id='
|
|
. $pageID
|
|
. '&cursor={"card_id"%3A"videos"%2C"has_next_page"%3Atrue}&surface=mobile_page_home&unit_count=8';
|
|
|
|
$fileContent = file_get_contents($requestString);
|
|
|
|
$articleIndex = 0;
|
|
$maxArticle = 3;
|
|
|
|
$html = $this->buildContent($fileContent);
|
|
$author = $this->getInput('u');
|
|
|
|
foreach($html->find("article") as $content){
|
|
|
|
$item = array();
|
|
$item['uri'] = "http://touch.facebook.com"
|
|
. $content->find("div._52jc", 0)->find("a", 0)->getAttribute("href");
|
|
|
|
$content->find("header", 0)->innertext = "";
|
|
$content->find("footer", 0)->innertext = "";
|
|
|
|
//Remove html nodes, keep only img, links, basic formatting
|
|
$content = strip_tags($content, '<a><img><i><u><br><p>');
|
|
|
|
//Adapt link hrefs: convert relative links into absolute links and bypass external link redirection
|
|
$content = preg_replace_callback('/ href=\"([^"]+)\"/i', $unescape_fb_link, $content);
|
|
|
|
//Clean useless html tag properties and fix link closing tags
|
|
foreach (array(
|
|
'onmouseover',
|
|
'onclick',
|
|
'target',
|
|
'ajaxify',
|
|
'tabindex',
|
|
'class',
|
|
'style',
|
|
'data-[^=]*',
|
|
'aria-[^=]*',
|
|
'role',
|
|
'rel',
|
|
'id') as $property_name)
|
|
$content = preg_replace('/ ' . $property_name . '=\"[^"]*\"/i', '', $content);
|
|
$content = preg_replace('/<\/a [^>]+>/i', '</a>', $content);
|
|
|
|
//Convert textual representation of emoticons eg
|
|
// "<i><u>smile emoticon</u></i>" back to ASCII emoticons eg ":)"
|
|
$content = preg_replace_callback('/<i><u>([^ <>]+) ([^<>]+)<\/u><\/i>/i', $unescape_fb_emote, $content);
|
|
|
|
$item['content'] = $content;
|
|
|
|
$title = $author;
|
|
if (strlen($title) > 24)
|
|
$title = substr($title, 0, strpos(wordwrap($title, 24), "\n")) . '...';
|
|
$title = $title . ' | ' . strip_tags($content);
|
|
if (strlen($title) > 64)
|
|
$title = substr($title, 0, strpos(wordwrap($title, 64), "\n")) . '...';
|
|
|
|
$item['title'] = $title;
|
|
$item['author'] = $author;
|
|
|
|
array_push($this->items, $item);
|
|
}
|
|
}
|
|
|
|
|
|
// Currently not used. Is used to get more than only 3 elements, as they appear on another page.
|
|
private function computeNextLink($string, $pageID){
|
|
|
|
$regex = "/timeline_unit\\\\\\\\u00253A1\\\\\\\\u00253A([0-9]*)\\\\\\\\u00253A([0-9]*)\\\\\\\\u00253A([0-9]*)\\\\\\\\u00253A([0-9]*)/";
|
|
preg_match($regex, $string, $result);
|
|
|
|
return "https://touch.facebook.com/pages_reaction_units/more/?page_id="
|
|
. $pageID
|
|
. "&cursor=%7B%22timeline_cursor%22%3A%22timeline_unit%3A1%3A"
|
|
. $result[1]
|
|
. "%3A"
|
|
. $result[2]
|
|
. "%3A"
|
|
. $result[3]
|
|
. "%3A"
|
|
. $result[4]
|
|
. "%22%2C%22timeline_section_cursor%22%3A%7B%7D%2C%22has_next_page%22%3Atrue%7D&surface=mobile_page_home&unit_count=3";
|
|
}
|
|
|
|
//Builds the HTML from the encoded JS that Facebook provides.
|
|
private function buildContent($pageContent){
|
|
|
|
$regex = "/\\\"html\\\":\\\"(.*?)\\\",\\\"replace/";
|
|
preg_match($regex, $pageContent, $result);
|
|
|
|
return str_get_html(html_entity_decode(json_decode('"' . $result[1] . '"')));
|
|
}
|
|
|
|
|
|
//Builds the cookie from the page, as Facebook sometimes refuses to give
|
|
//the page if no cookie is provided.
|
|
private function getCookies($pageURL){
|
|
|
|
$ctx = stream_context_create(array(
|
|
'http' => array(
|
|
'user_agent' => "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:46.0) Gecko/20100101 Firefox/46.0",
|
|
'Accept' => 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8'
|
|
)
|
|
)
|
|
);
|
|
$a = file_get_contents($pageURL, 0, $ctx);
|
|
|
|
//First request to get the cookie
|
|
$cookies = "";
|
|
foreach($http_response_header as $hdr){
|
|
if(strpos($hdr, "Set-Cookie") !== false){
|
|
$cLine = explode(":", $hdr)[1];
|
|
$cLine = explode(";", $cLine)[0];
|
|
$cookies .= ";" . $cLine;
|
|
}
|
|
}
|
|
|
|
return substr($cookies, 1);
|
|
}
|
|
|
|
//Get the page ID from the Facebook page.
|
|
private function getPageID($page, $cookies){
|
|
|
|
$context = stream_context_create(array(
|
|
'http' => array(
|
|
'user_agent' => "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:46.0) Gecko/20100101 Firefox/46.0",
|
|
'header' => 'Cookie: ' . $cookies
|
|
)
|
|
)
|
|
);
|
|
|
|
$pageContent = file_get_contents($page, 0, $context);
|
|
|
|
//Get the page ID if we don't have a captcha
|
|
$regex = "/page_id=([0-9]*)&/";
|
|
preg_match($regex, $pageContent, $matches);
|
|
|
|
if(count($matches) > 0){
|
|
return $matches[1];
|
|
}
|
|
|
|
//Get the page ID if we do have a captcha
|
|
$regex = "/\"pageID\":\"([0-9]*)\"/";
|
|
preg_match($regex, $pageContent, $matches);
|
|
|
|
return $matches[1];
|
|
|
|
}
|
|
|
|
public function getName(){
|
|
return (isset($this->name) ? $this->name . ' - ' : '') . 'Facebook Bridge';
|
|
}
|
|
|
|
public function getURI(){
|
|
return 'http://facebook.com';
|
|
}
|
|
|
|
public function getCacheDuration(){
|
|
return 60 * 60 * 3; // 5 minutes
|
|
}
|
|
}
|