captcha and imgur, findthatmeme, yep imagesearch

1 files changed, 249 insertions, 0 deletions

diff --git a/scraper/imgur.php b/scraper/imgur.php
new file mode 100644
index 0000000..4a16de7
--- /dev/null
+++ b/scraper/imgur.php
@@ -0,0 +1,249 @@
+<?php
+
+class imgur{
+	
+	public function __construct(){
+		
+		include "lib/nextpage.php";
+		$this->nextpage = new nextpage("imgur");
+		
+		include "lib/fuckhtml.php";
+		$this->fuckhtml = new fuckhtml();
+	}
+	
+	public function getfilters($page){
+		
+		return [
+			"sort" => [ // /score/
+				"display" => "Sort by",
+				"option" => [
+					"score" => "Highest scoring",
+					"relevance" => "Most relevant",
+					"time" => "Newest first"
+				]
+			],
+			"time" => [ // /score/day/
+				"display" => "Time posted",
+				"option" => [
+					"all" => "All time",
+					"day" => "Today",
+					"week" => "This week",
+					"month" => "This month",
+					"year" => "This year"
+				]
+			],
+			"format" => [ // q_type
+				"display" => "Format",
+				"option" => [
+					"any" => "Any format",
+					"jpg" => "JPG",
+					"png" => "PNG",
+					"gif" => "GIF",
+					"anigif" => "Animated GIF",
+					"album" => "Albums"
+				]
+			],
+			"size" => [ // q_size_px
+				"display" => "Size",
+				"option" => [
+					"any" => "Any size",
+					"small" => "Small (500px or less)",
+					"med" => "Medium (500px to 2000px)",
+					"big" => "Big (2000px to 5000px)",
+					"lrg" => "Large (5000px to 10000px)",
+					"huge" => "Huge (10000px and above)"
+				]
+			]
+		];
+	}
+	
+	private function get($url, $get = []){
+		
+		$curlproc = curl_init();
+		
+		if($get !== []){
+			$get = http_build_query($get);
+			$url .= "?scrolled&" . $get;
+		}
+		
+		curl_setopt($curlproc, CURLOPT_URL, $url);
+		
+		curl_setopt($curlproc, CURLOPT_ENCODING, ""); // default encoding
+		curl_setopt($curlproc, CURLOPT_HTTPHEADER,
+			["User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:107.0) Gecko/20100101 Firefox/110.0",
+			"Accept: text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,*/*;q=0.8",
+			"Accept-Language: en-US,en;q=0.5",
+			"Accept-Encoding: gzip",
+			"DNT: 1",
+			"Referer: https://imgur.com/search/",
+			"Connection: keep-alive",
+			"Sec-Fetch-Dest: empty",
+			"Sec-Fetch-Mode: cors",
+			"Sec-Fetch-Site: same-origin",
+			"TE: trailers",
+			"X-Requested-With: XMLHttpRequest"]
+		);
+		
+		curl_setopt($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);
+		
+		$data = curl_exec($curlproc);
+		
+		if(curl_errno($curlproc)){
+			
+			throw new Exception(curl_error($curlproc));
+		}
+		
+		curl_close($curlproc);
+		return $data;
+	}
+	
+	public function image($get){
+		
+		if($get["npt"]){
+			
+			$filter =
+				json_decode(
+					$this->nextpage->get(
+						$get["npt"],
+						"images"
+					),
+					true
+				);
+			
+			$search = $filter["s"];
+			unset($filter["s"]);
+			
+			$sort = $filter["sort"];
+			unset($filter["sort"]);
+			
+			$time = $filter["time"];
+			unset($filter["time"]);
+			
+			$format = $filter["format"];
+			unset($filter["format"]);
+			
+			$size = $filter["size"];
+			unset($filter["size"]);
+			
+			$page = $filter["page"];
+			unset($filter["page"]);
+		}else{
+			
+			$search = $get["s"];
+			$sort = $get["sort"];
+			$time = $get["time"];
+			$format = $get["format"];
+			$size = $get["size"];
+			$page = 0;
+			
+			$filter = [
+				"q" => $search
+			];
+			
+			if($format != "any"){
+				
+				$filter["q_type"] = $format;
+			}
+			
+			if($size != "any"){
+				
+				$filter["q_size_px"] = $size;
+				$filter["q_size_is_mpx"] = "off";
+			}
+		}
+		
+		$out = [
+			"status" => "ok",
+			"npt" => null,
+			"image" => []
+		];
+		
+		try{
+			$html =
+				$this->get(
+					"https://imgur.com/search/$sort/$time/page/$page",
+					$filter
+				);
+			
+		}catch(Exception $error){
+			
+			throw new Exception("Failed to fetch HTML");
+		}
+		
+		$this->fuckhtml->load($html);
+		
+		$posts =
+			$this->fuckhtml
+			->getElementsByClassName(
+				"post",
+				"div"
+			);
+		
+		foreach($posts as $post){
+			
+			$this->fuckhtml->load($post);
+			
+			$image =
+				$this->fuckhtml
+				->getElementsByTagName("img")[0];
+			
+			$image_url = "https:" . substr($this->fuckhtml->getTextContent($image["attributes"]["src"]), 0, -5);
+			
+			$out["image"][] = [
+				"title" =>
+					$this->fuckhtml
+					->getTextContent(
+						$image["attributes"]["alt"]
+					),
+				"source" => [
+					[
+						"url" => $image_url . ".jpg",
+						"width" => null,
+						"height" => null
+					],
+					[
+						"url" => $image_url . "m.jpg",
+						"width" => null,
+						"height" => null
+					]
+				],
+				"url" =>
+					"https://imgur.com" .
+					$this->fuckhtml
+					->getTextContent(
+						$this->fuckhtml
+						->getElementsByClassName(
+							"image-list-link",
+							"a"
+						)
+						[0]
+						["attributes"]
+						["href"]
+					)
+			];
+		}
+		
+		if(isset($out["image"][0])){
+			
+			// store nextpage
+			$filter["s"] = $search;
+			$filter["sort"] = $sort;
+			$filter["time"] = $time;
+			$filter["format"] = $format;
+			$filter["size"] = $size;
+			$filter["page"] = $page + 1;
+			
+			$out["npt"] =
+				$this->nextpage->store(
+					json_encode($filter),
+					"images"
+				);
+		}
+		
+		return $out;
+	}
+}