Merge pull request #1 from postaddictme/master

Update
postaddictme · Mar 7, 2017 · fec6ecd · fec6ecd
2 parents 5093c2e + 89fe67c
commit fec6ecd
Show file tree

Hide file tree

Showing 7 changed files with 168 additions and 106 deletions.
diff --git a/CHANGELOG b/CHANGELOG
@@ -1,5 +1,38 @@
                    Instagram PHP Scraper Change Log
                    ================================
+Version v0.4.5
+--------------
+- Enh: BIG thing. Get best resolutions for medias possible
+
+Version v0.4.4
+--------------
+- Bug: Image urls broken for some medias in getMediaByCode()
+
+Version v0.4.3
+--------------
+- Enh: Make methods getContentsFromUrl() and generateRandomString() static
+
+Version v0.4.2
+--------------
+- Bug: Fix 405 error with fetching accounts by id
+- Bug: Fix 405 error with fetching comments by id
+
+Version v0.4.1
+--------------
+- Enh: Url encode endpoints to support for example Japan language
+
+Version v0.4.0
+--------------
+- Enh: New methods to paginate medias getPaginateMedias() and getPaginateMediasByTag()
+
+Version v0.3.5
+--------------
+- Enh: Convenient media pagination getMediasByTag()
+
+Version v0.3.4
+--------------
+- Bug: Fix media urls
+
 Version v0.3.3
 --------------
 - Bug: Include class Location and Comment in InstagramScraper.php

diff --git a/README.md b/README.md
@@ -35,9 +35,9 @@ echo $account->username;
 
 ### Search users by username
 ```php
-$medias = Instagram::searchAccountsByUsername('durov');
+$users = Instagram::searchAccountsByUsername('durov');
 echo '<pre>';
-echo json_encode($medias);
+echo json_encode($users);
 echo '</pre><br/>';
 ```
 
@@ -56,9 +56,20 @@ Available properties:
     $imageStandardResolutionUrl;
     $imageHighResolutionUrl;
     $caption;
+    $captionIsEdited;
+    $isAd;
     $videoLowResolutionUrl;
     $videoStandardResolutionUrl;
     $videoLowBandwidthUrl;
+    $videoViews;
+    $code;
+    $owner;
+    $ownerId;
+    $likesCount;
+    $locationId;
+    $locationName;
+    $commentsCount;
+
 */
 echo $medias[0]->imageHighResolutionUrl;
 echo $medias[0]->caption;
@@ -163,4 +174,7 @@ $medias = Instagram::getLocationTopMediasById(1);
 ### Get location medias by location id
 ```php
 $medias = Instagram::getLocationMediasById(1);
-```
+```
+
+### Other
+Java library: https://github.com/postaddictme/instagram-java-scraper
diff --git a/index.php b/index.php
@@ -3,13 +3,15 @@
 
 require_once 'vendor/autoload.php';
 require_once 'src/InstagramScraper.php';
-use InstagramScraper\Exception\InstagramException;
+
 use InstagramScraper\Instagram;
 
-$instagram = new Instagram();
 try {
-    $medias = Instagram::getMedias('kevin', 1000);
-    echo $medias[998]->imageThumbnailUrl;
+//    $medias = Instagram::getMedias('kevin', 1497);
+//    echo json_encode($medias[1497]);
+    $medias = InstagramScraper\Instagram::getMediasByTag('paveldurov', 300);
+    echo sizeof($medias) . '\n';
+//    echo json_encode($medias);
 } catch (\Exception $ex) {
     print_r($ex);
 }

diff --git a/src/InstagramScraper/Endpoints.php b/src/InstagramScraper/Endpoints.php
@@ -13,11 +13,13 @@ class Endpoints
     const MEDIA_JSON_BY_LOCATION_ID = 'https://www.instagram.com/explore/locations/{{facebookLocationId}}/?__a=1&max_id={{maxId}}';
     const MEDIA_JSON_BY_TAG = 'https://www.instagram.com/explore/tags/{tag}/?__a=1&max_id={max_id}';
     const GENERAL_SEARCH = 'https://www.instagram.com/web/search/topsearch/?query={query}';
-    const ACCOUNT_JSON_INFO_BY_ID = 'https://www.instagram.com/query/?q=ig_user({userId}){id,username,external_url,full_name,profile_pic_url,biography,followed_by{count},follows{count},media{count},is_private,is_verified}';
-    const LAST_COMMENTS_BY_CODE = 'https://www.instagram.com/query/?q=ig_shortcode({{code}}){comments.last({{count}}){count,nodes{id,created_at,text,user{id,profile_pic_url,username,follows{count},followed_by{count},biography,full_name,media{count},is_private,external_url,is_verified}},page_info}}';
-    const COMMENTS_BEFORE_COMMENT_ID_BY_CODE = 'https://www.instagram.com/query/?q=ig_shortcode({{code}}){comments.before({{commentId}},{{count}}){count,nodes{id,created_at,text,user{id,profile_pic_url,username,follows{count},followed_by{count},biography,full_name,media{count},is_private,external_url,is_verified}},page_info}}';
-    const LAST_LIKES_BY_CODE = 'https://www.instagram.com/query/?q=ig_shortcode({{code}}){likes{nodes{id,user{id,profile_pic_url,username,follows{count},followed_by{count},biography,full_name,media{count},is_private,external_url,is_verified}},page_info}}';
+    const ACCOUNT_JSON_INFO_BY_ID = 'ig_user({userId}){id,username,external_url,full_name,profile_pic_url,biography,followed_by{count},follows{count},media{count},is_private,is_verified}';
+    const LAST_COMMENTS_BY_CODE = 'ig_shortcode({{code}}){comments.last({{count}}){count,nodes{id,created_at,text,user{id,profile_pic_url,username,follows{count},followed_by{count},biography,full_name,media{count},is_private,external_url,is_verified}},page_info}}';
+    const COMMENTS_BEFORE_COMMENT_ID_BY_CODE = 'ig_shortcode({{code}}){comments.before({{commentId}},{{count}}){count,nodes{id,created_at,text,user{id,profile_pic_url,username,follows{count},followed_by{count},biography,full_name,media{count},is_private,external_url,is_verified}},page_info}}';
+    const LAST_LIKES_BY_CODE = 'ig_shortcode({{code}}){likes{nodes{id,user{id,profile_pic_url,username,follows{count},followed_by{count},biography,full_name,media{count},is_private,external_url,is_verified}},page_info}}';
 
+    const INSTAGRAM_QUERY_URL = 'https://www.instagram.com/query/';
+    const INSTAGRAM_CDN_URL = 'https://scontent.cdninstagram.com/';
 
     public static function getAccountPageLink($username)
     {

diff --git a/src/InstagramScraper/Instagram.php b/src/InstagramScraper/Instagram.php
@@ -34,18 +34,11 @@ public static function getAccount($username)
 
     public static function getAccountById($id)
     {
-
         if (!is_numeric($id)) {
             throw new \InvalidArgumentException('User id must be integer or integer wrapped in string');
         }
-        $response = Request::get(Endpoints::getAccountJsonInfoLinkByAccountId($id));
-        if ($response->code === 404) {
-            throw new InstagramNotFoundException('Account with given username does not exist.');
-        }
-        if ($response->code !== 200) {
-            throw new InstagramException('Response code is ' . $response->code . '. Body: ' . $response->body . ' Something went wrong. Please report issue.');
-        }
-        $userArray = json_decode($response->raw_body, true);
+        $parameters = Endpoints::getAccountJsonInfoLinkByAccountId($id);
+        $userArray = json_decode(self::getContentsFromUrl($parameters), true);
         if ($userArray['status'] === 'fail') {
             throw new InstagramException($userArray['message']);
         }
@@ -55,6 +48,38 @@ public static function getAccountById($id)
         return Account::fromAccountPage($userArray);
     }
 
+    private static function getContentsFromUrl($parameters)
+    {
+        if (!function_exists('curl_init')) {
+            return false;
+        }
+        $random = self::generateRandomString();
+        $ch = curl_init();
+        curl_setopt($ch, CURLOPT_URL, Endpoints::INSTAGRAM_QUERY_URL);
+        curl_setopt($ch, CURLOPT_RETURNTRANSFER, TRUE);
+        curl_setopt($ch, CURLOPT_POST, 1);
+        curl_setopt($ch, CURLOPT_POSTFIELDS, 'q=' . $parameters);
+        $headers = array();
+        $headers[] = "Cookie:  csrftoken=$random;";
+        $headers[] = "X-Csrftoken: $random";
+        $headers[] = "Referer: https://www.instagram.com/";
+        curl_setopt($ch, CURLOPT_HTTPHEADER, $headers);
+        $output = curl_exec($ch);
+        curl_close($ch);
+        return $output;
+    }
+
+    private static function generateRandomString($length = 10)
+    {
+        $characters = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ';
+        $charactersLength = strlen($characters);
+        $randomString = '';
+        for ($i = 0; $i < $length; $i++) {
+            $randomString .= $characters[rand(0, $charactersLength - 1)];
+        }
+        return $randomString;
+    }
+
     public static function getMedias($username, $count = 20, $maxId = '')
     {
         $index = 0;
@@ -95,8 +120,8 @@ public static function getPaginateMedias($username, $maxId = '')
         $medias = [];
 
         $toReturn = [
-            'medias'      => $medias,
-            'maxId'       => $maxId,
+            'medias' => $medias,
+            'maxId' => $maxId,
             'hasNextPage' => $hasNextPage
         ];
 
@@ -124,8 +149,8 @@ public static function getPaginateMedias($username, $maxId = '')
         $hasNextPage = $arr['more_available'];
 
         $toReturn = [
-            'medias'      => $medias,
-            'maxId'       => $maxId,
+            'medias' => $medias,
+            'maxId' => $maxId,
             'hasNextPage' => $hasNextPage
         ];
 
@@ -160,6 +185,7 @@ public static function getMediasByTag($tag, $count = 12, $maxId = '')
     {
         $index = 0;
         $medias = [];
+        $mediaIds = [];
         $hasNextPage = true;
         while ($index < $count && $hasNextPage) {
             $response = Request::get(Endpoints::getMediasJsonByTagLink($tag, $maxId));
@@ -179,7 +205,12 @@ public static function getMediasByTag($tag, $count = 12, $maxId = '')
                 if ($index === $count) {
                     return $medias;
                 }
-                $medias[] = Media::fromTagPage($mediaArray);
+                $media = Media::fromTagPage($mediaArray);
+                if (in_array($media->id, $mediaIds)) {
+                    return $medias;
+                }
+                $mediaIds[] = $media->id;
+                $medias[] = $media;
                 $index++;
             }
             if (count($nodes) == 0) {
@@ -197,8 +228,8 @@ public static function getPaginateMediasByTag($tag, $maxId = '')
         $medias = [];
 
         $toReturn = [
-            'medias'      => $medias,
-            'maxId'       => $maxId,
+            'medias' => $medias,
+            'maxId' => $maxId,
             'hasNextPage' => $hasNextPage
         ];
 
@@ -230,11 +261,13 @@ public static function getPaginateMediasByTag($tag, $maxId = '')
 
         $maxId = $arr['tag']['media']['page_info']['end_cursor'];
         $hasNextPage = $arr['tag']['media']['page_info']['has_next_page'];
+        $count = $arr['tag']['media']['count'];
 
         $toReturn = [
-            'medias'      => $medias,
-            'maxId'       => $maxId,
-            'hasNextPage' => $hasNextPage
+            'medias' => $medias,
+            'count' => $count,
+            'maxId' => $maxId,
+            'hasNextPage' => $hasNextPage,
         ];
 
         return $toReturn;
@@ -336,17 +369,12 @@ public static function getMediaCommentsByCode($code, $count = 10, $maxId = null)
                 $remain = 0;
             }
             if (!isset($maxId)) {
-                $response = Request::get(Endpoints::getLastCommentsByCodeLink($code, $numberOfCommentsToRetreive));
+                $parameters = Endpoints::getLastCommentsByCodeLink($code, $numberOfCommentsToRetreive);
+
             } else {
-                $response = Request::get(Endpoints::getCommentsBeforeCommentIdByCode($code, $numberOfCommentsToRetreive, $maxId));
+                $parameters = Endpoints::getCommentsBeforeCommentIdByCode($code, $numberOfCommentsToRetreive, $maxId);
             }
-            if ($response->code === 404) {
-                throw new InstagramNotFoundException('Account with given username does not exist.');
-            }
-            if ($response->code !== 200) {
-                throw new InstagramException('Response code is ' . $response->code . '. Body: ' . $response->body . ' Something went wrong. Please report issue.');
-            }
-            $jsonResponse = json_decode($response->raw_body, true);
+            $jsonResponse = json_decode(self::getContentsFromUrl($parameters), true);
             $nodes = $jsonResponse['comments']['nodes'];
             foreach ($nodes as $commentArray) {
                 $comments[] = Comment::fromApi($commentArray);
@@ -422,21 +450,4 @@ public static function getLocationById($facebookLocationId)
         $jsonResponse = json_decode($response->raw_body, true);
         return Location::makeLocation($jsonResponse['location']);
     }
-
-    public static function getLastLikesByCode($code)
-    {
-        $response = Request::get(Endpoints::getLastLikesByCodeLink($code));
-        if ($response->code === 404) {
-            throw new InstagramNotFoundException('Media with this shortcode doesn\'t exist');
-        }
-        if ($response->code !== 200) {
-            throw new InstagramException('Response code is ' . $response->code . '. Body: ' . $response->body . ' Something went wrong. Please report issue.');
-        }
-        $jsonResponse = json_decode($response->raw_body, true);
-        $users = [];
-        foreach ($jsonResponse['likes']['nodes'] as $userArray) {
-            $users[] = Account::fromAccountPage($userArray['user']);
-        }
-        return $users;
-    }
 }