From 02e6ad458c64716910cdf8aad74d16d78abd018c Mon Sep 17 00:00:00 2001
From: ROMANGOLIENKO <romangolienko@yandex.ru>
Date: Fri, 26 Jan 2024 16:08:54 +0300
Subject: [PATCH 1/2] actual version, device change

---
 .gitignore                                    |  1 +
 app/Services/PageScrapper.php                 |  2 +-
 composer.json                                 |  4 ++-
 resources/views/menu/course.blade.php         | 26 ++++++++++++++-----
 ...i-mezhdunarodnoe-sotrudnichestvo.blade.php |  3 ++-
 5 files changed, 27 insertions(+), 9 deletions(-)

diff --git a/.gitignore b/.gitignore
index 3649080..b01c8c9 100644
--- a/.gitignore
+++ b/.gitignore
@@ -21,3 +21,4 @@ yarn-error.log
 /.phpstorm.meta.php
 /_ide_helper.php
 /_ide_helper_models.php
+/public/img/icons/play-button.png
diff --git a/app/Services/PageScrapper.php b/app/Services/PageScrapper.php
index 8dd6658..dd6a884 100644
--- a/app/Services/PageScrapper.php
+++ b/app/Services/PageScrapper.php
@@ -22,7 +22,7 @@ class PageScrapper
         $arr = [];
         $rez = preg_match_all($strForPregMatch, $page, $arr);
 
-        return $content = $arr[1][0];
+        return $content = $arr[0][0];
     }
 
     public function normalizeURLFile($content)
diff --git a/composer.json b/composer.json
index 711b14f..4fa8bb7 100644
--- a/composer.json
+++ b/composer.json
@@ -8,8 +8,9 @@
     ],
     "license": "MIT",
     "require": {
-        "php": "^8.2",
+        "php": "^8.1|8.2",
         "guzzlehttp/guzzle": "^7.8.1",
+        "imangazaliev/didom": "^2.0",
         "laravel/framework": "^10.41.0",
         "laravel/sanctum": "^3.3.3",
         "laravel/tinker": "^2.9.0",
@@ -64,6 +65,7 @@
         }
     },
     "config": {
+        "platform-check": false,
         "optimize-autoloader": true,
         "preferred-install": "dist",
         "sort-packages": true,
diff --git a/resources/views/menu/course.blade.php b/resources/views/menu/course.blade.php
index df990a3..8779bd1 100644
--- a/resources/views/menu/course.blade.php
+++ b/resources/views/menu/course.blade.php
@@ -1,4 +1,5 @@
 @extends('layouts.applicant-layout')
+
 @section('content')
 
     <style>
@@ -35,12 +36,25 @@
 
         <div class="col-10">
             @php
-                use App\Services\PageScrapper;
-                $pageScrapper = new PageScrapper("https://mkgtu.ru/postuplenie/podat-dokumenty-onlayn/", '<div class=["\']content_info["\']>');
-                $row = $pageScrapper->getHTML();
-                $content = $pageScrapper->normalizeURLFile($row);
-                $content = $pageScrapper->cutHTML($content,'/<footer(.*)<\/footer>/isU');
-                echo $content;
+
+
+
+
+                use DiDom\Document;
+
+                    $document = new Document('https://mkgtu.ru/postuplenie/podat-dokumenty-onlayn/', true);
+
+                    $posts = $document->find('.content_info');
+
+                    echo $document->first('.content_info')->html();
+    /*
+                    use App\Services\PageScrapper;
+                    $pageScrapper = new PageScrapper("https://mkgtu.ru/postuplenie/podat-dokumenty-onlayn/", '<div class=["\']content_info["\']>');
+                    $row = $pageScrapper->getHTML();
+                    $content = $pageScrapper->normalizeURLFile($row);
+                    $content = $pageScrapper->cutHTML($content,'/<footer(.*)<\/footer>/isU');
+                    echo $content;
+    */
 
             @endphp
 
diff --git a/resources/views/menu/inostrannym-abiturientam/akademicheskaya-mobilnost-i-mezhdunarodnoe-sotrudnichestvo.blade.php b/resources/views/menu/inostrannym-abiturientam/akademicheskaya-mobilnost-i-mezhdunarodnoe-sotrudnichestvo.blade.php
index 2fd3435..423ec8e 100644
--- a/resources/views/menu/inostrannym-abiturientam/akademicheskaya-mobilnost-i-mezhdunarodnoe-sotrudnichestvo.blade.php
+++ b/resources/views/menu/inostrannym-abiturientam/akademicheskaya-mobilnost-i-mezhdunarodnoe-sotrudnichestvo.blade.php
@@ -28,13 +28,14 @@
                 $row3 = $pageScrapper3->cutHTML($row3,'/<footer(.*)<\/footer>/isU');
 
                 $content = $row . $row2 . $row3;
+
                 //dd($content);
 
 
                 //$content = $pageScrapper->cutHTML($row,'/<footer(.*)<\/footer>/isU');
                // $content = $pageScrapper->normalizeURLFile($content);
 
-                echo htmlspecialchars($content);
+                echo $content;
 
             @endphp
 
-- 
2.41.0


From d24f32623b0d0ab7c6292723e8e1ae2b9632149c Mon Sep 17 00:00:00 2001
From: ROMANGOLIENKO <romangolienko@yandex.ru>
Date: Fri, 26 Jan 2024 14:45:33 +0300
Subject: [PATCH 2/2] newScrapper

---
 .gitignore                            |  1 +
 app/Services/PageScrapper.php         | 39 +++++++++++++++++++++++++++
 resources/views/menu/course.blade.php | 36 ++++++++++++++-----------
 3 files changed, 61 insertions(+), 15 deletions(-)

diff --git a/.gitignore b/.gitignore
index b01c8c9..2749dec 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,3 +1,4 @@
+/public/img/icons/play-button.png
 /.phpunit.cache
 /node_modules
 /public/build
diff --git a/app/Services/PageScrapper.php b/app/Services/PageScrapper.php
index dd6a884..fbf9061 100644
--- a/app/Services/PageScrapper.php
+++ b/app/Services/PageScrapper.php
@@ -2,6 +2,9 @@
 
 namespace App\Services;
 
+use DiDom\Document;
+
+/*
 class PageScrapper
 {
     private string $url;
@@ -60,3 +63,39 @@ class PageScrapper
         return $content;
     }
 }
+*/
+
+class PageScrapper
+{
+    private string $url;
+    private string $contentMarker;
+
+    public function __construct($url, $contentMarker)
+    {
+        $this->url = $url;
+        $this->contentMarker = $contentMarker;
+        $this->document = new Document($this->url, true);
+    }
+    public function getFullHTML()
+    {
+        return $this->document;
+    }
+    public function printHTML()
+    {
+        $rez = $this->document;
+        $content = $rez->first($this->contentMarker)->html();
+        return $content;
+    }
+    public function normalizePath()
+    {
+        $links = (string) $this->document->find('a[href]');
+        dd($links);
+        foreach ($links as $link) {
+            if (!str_starts_with($link, 'https')) {
+                $this->document = str_replace($link, 'https://mkgtu.ru' . $link, $this->document);
+            }
+        }
+
+        return $this->document;
+    }
+}
diff --git a/resources/views/menu/course.blade.php b/resources/views/menu/course.blade.php
index 8779bd1..4dc52b1 100644
--- a/resources/views/menu/course.blade.php
+++ b/resources/views/menu/course.blade.php
@@ -36,25 +36,31 @@
 
         <div class="col-10">
             @php
+                use App\Services\PageScrapper;
+                $pageScrapper = new PageScrapper("https://mkgtu.ru/postuplenie/podat-dokumenty-onlayn/", '.content_info');
+                $content = $pageScrapper->printHTML();
+                $content = $pageScrapper->normalizePath();
+
+                echo $content;
 
 
 
 
-                use DiDom\Document;
-
-                    $document = new Document('https://mkgtu.ru/postuplenie/podat-dokumenty-onlayn/', true);
-
-                    $posts = $document->find('.content_info');
-
-                    echo $document->first('.content_info')->html();
-    /*
-                    use App\Services\PageScrapper;
-                    $pageScrapper = new PageScrapper("https://mkgtu.ru/postuplenie/podat-dokumenty-onlayn/", '<div class=["\']content_info["\']>');
-                    $row = $pageScrapper->getHTML();
-                    $content = $pageScrapper->normalizeURLFile($row);
-                    $content = $pageScrapper->cutHTML($content,'/<footer(.*)<\/footer>/isU');
-                    echo $content;
-    */
+                //                    use DiDom\Document;
+                //
+                //                    $document = new Document('https://mkgtu.ru/postuplenie/podat-dokumenty-onlayn/', true);
+                //
+                //                    $posts = $document->find('.content_info');
+                //
+                //                    echo $document->first('.content_info')->html();
+                    /*
+                                    use App\Services\PageScrapper;
+                                    $pageScrapper = new PageScrapper("https://mkgtu.ru/postuplenie/podat-dokumenty-onlayn/", '<div class=["\']content_info["\']>');
+                                    $row = $pageScrapper->getHTML();
+                                    $content = $pageScrapper->normalizeURLFile($row);
+                                    $content = $pageScrapper->cutHTML($content,'/<footer(.*)<\/footer>/isU');
+                                    echo $content;
+                    */
 
             @endphp
 
-- 
2.41.0