4get/scraper/crowdview.php

146 lines
2.8 KiB
PHP

<?php
class crowdview{
public function __construct(){
include "lib/backend.php";
$this->backend = new backend("crowdview");
include "lib/fuckhtml.php";
$this->fuckhtml = new fuckhtml();
}
public function getfilters($page){
return [];
}
private function get($proxy, $url, $get = []){
$curlproc = curl_init();
if($get !== []){
$get = http_build_query($get);
$url .= "?" . $get;
}
curl_setopt($curlproc, CURLOPT_URL, $url);
curl_setopt($curlproc, CURLOPT_ENCODING, ""); // default encoding
curl_setopt($curlproc, CURLOPT_HTTPHEADER,
["User-Agent: " . config::USER_AGENT,
"Accept: text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,*/*;q=0.8",
"Accept-Language: en-US,en;q=0.5",
"Accept-Encoding: gzip",
"DNT: 1",
"Connection: keep-alive",
"Upgrade-Insecure-Requests: 1",
"Sec-Fetch-Dest: document",
"Sec-Fetch-Mode: navigate",
"Sec-Fetch-Site: none",
"Sec-Fetch-User: ?1"]
);
curl_setopt($curlproc, CURLOPT_RETURNTRANSFER, true);
curl_setopt($curlproc, CURLOPT_SSL_VERIFYHOST, 2);
curl_setopt($curlproc, CURLOPT_SSL_VERIFYPEER, true);
curl_setopt($curlproc, CURLOPT_CONNECTTIMEOUT, 30);
curl_setopt($curlproc, CURLOPT_TIMEOUT, 30);
$this->backend->assign_proxy($curlproc, $proxy);
$data = curl_exec($curlproc);
if(curl_errno($curlproc)){
throw new Exception(curl_error($curlproc));
}
curl_close($curlproc);
return $data;
}
public function web($get){
$search = $get["s"];
if(strlen($search) === 0){
throw new Exception("Search term is empty!");
}
$proxy = $this->backend->get_ip();
try{
$json = $this->get(
$proxy,
"https://crowdview-next-js.onrender.com/api/search-v3",
[
"query" => $search
]
);
}catch(Exception $error){
throw new Exception("Failed to fetch JSON");
}
$out = [
"status" => "ok",
"spelling" => [
"type" => "no_correction",
"using" => null,
"correction" => null
],
"npt" => null,
"answer" => [],
"web" => [],
"image" => [],
"video" => [],
"news" => [],
"related" => []
];
$json = json_decode($json, true);
if($json === NULL){
throw new Exception("Failed to decode JSON");
}
foreach($json["results"] as $item){
$description = explode("<b>", $item["snippet"], 2);
$out["web"][] = [
"title" => $this->sanitize($item["title"]),
"description" => $this->sanitize($description[1]),
"url" => $item["link"],
"date" => strtotime($description[0]),
"type" => "web",
"thumb" => [
"url" => null,
"ratio" => null
],
"sublink" => [],
"table" => []
];
}
return $out;
}
private function sanitize($html){
return
trim(
$this->fuckhtml
->getTextContent(
html_entity_decode(
$html
)
),
". "
);
}
}