Merge pull request #2432 from Nippey/patch-2

ImgurCom.py Galleries downloaded incompletely

1 files changed, 93 insertions, 4 deletions

diff --git a/module/plugins/crypter/ImgurCom.py b/module/plugins/crypter/ImgurCom.py
index 710f556c9..8b224cdaa 100644
--- a/module/plugins/crypter/ImgurCom.py
+++ b/module/plugins/crypter/ImgurCom.py
@@ -1,4 +1,5 @@
 import re
+import json
 
 from module.plugins.internal.SimpleCrypter import SimpleCrypter
 from module.plugins.internal.misc import uniqify
@@ -7,7 +8,7 @@ from module.plugins.internal.misc import uniqify
 class ImgurCom(SimpleCrypter):
     __name__    = "ImgurCom"
     __type__    = "crypter"
-    __version__ = "0.55"
+    __version__ = "0.56"
     __status__  = "testing"
 
     __pattern__ = r'https?://(?:www\.|m\.)?imgur\.com/(a|gallery|)/?\w{5,7}'
@@ -18,13 +19,101 @@ class ImgurCom(SimpleCrypter):
 
     __description__ = """Imgur.com decrypter plugin"""
     __license__     = "GPLv3"
-    __authors__     = [("nath_schwarz", "nathan.notwhite@gmail.com")]
+    __authors__     = [("nath_schwarz", "nathan.notwhite@gmail.com"),
+                       ("nippey",       "matthias.nippert@gmail.com")]
 
 
-    NAME_PATTERN = r'(?P<N>.+?) - Imgur'
+    NAME_PATTERN = r'(?P<N>.+?) - .*?Imgur'
     LINK_PATTERN = r'i\.imgur\.com/\w{7}s?\.(?:jpeg|jpg|png|gif|apng)'
+    
+    """ Imgur may only show the first 10 images of a gallery. The remaining bits may be found here: """
+    GALLERY_JSON = "http://imgur.com/ajaxalbums/getimages/{0}/hit.json?all=true"
+
+    def sanitize(self,name):
+        """ Turn Imgur Gallery title into a safe Package and Folder name """
+        keepcharacters = (' ','\t','.','_')
+        replacecharacters = (' ','\t')
+        return "".join(c if c not in replacecharacters else '_'   for c in name.strip()   if c.isalnum() or c in keepcharacters).strip('_')
+
+    def get_ids_from_json(self):
+        """ Check the embedded JSON and if needed the external JSON for more images """
+        
+        #Greedy re should match the closing bracket of json assuming JSON data is placed on a single line
+        m = re.search(r"\simage\s+:\s+({.*})", self.data)
+        if m:
+            json_all_imgs = json.loads(m.group(1))
+            
+            #Extract some metadata (ID, Title, NumImages)
+            self.gallery_id =      json_all_imgs['hash']
+            self.gallery_name =    self.sanitize( "{0}_{1}".format( json_all_imgs['hash'], json_all_imgs['title'] ) )
+            self.gallery_numImgs = int( json_all_imgs['num_images'] )
+
+            #Extract images
+            all_imgs = {e['hash']: e['ext'] for e in json_all_imgs['album_images']['images']}
+            self.log_debug('Found {0} of {1} expected links in embedded JSON'.format(len(all_imgs), self.gallery_numImgs) )
+            
+            #Depeding on the gallery, the embedded JSON may not contain all image information, then we also try the external JSON
+            #If this doesn't help either (which is possible),... TODO: Find out what to do
+            if self.gallery_numImgs > len(all_imgs):
+                json_data = self.load(self.GALLERY_JSON.format(self.gallery_id))
+                json_all_imgs = json.loads(json_data)
+                try:
+                    all_imgs = {e['hash']: e['ext'] for e in json_all_imgs['data']['images']}
+                    self.log_debug('Found {0} of {1} expected links in external JSON'.format(len(all_imgs), self.gallery_numImgs) )
+                except (KeyError, TypeError):
+                    self.log_debug('Could not extract links from external JSON')
+                    #It is possible that the returned JSON contains an empty 'data' section. We ignore it then.
+                    pass
+
+            return all_imgs
+        self.log_debug('Could not find embedded JSON')
+        return {}
+
+
+    def get_indirect_links(self, links_direct):
+        """ Try to find a list of all images and add those we didn't find already """
+        
+        #Extract IDs of known direct links
+        ids_direct = set([l for link in links_direct for l in re.findall(r'(\w{7})', link)])
+        
+        #Get filename extensions for new IDs
+        ids_json = self.get_ids_from_json()
+        ids_indirect = [id for id in ids_json.keys() if id not in ids_direct]
+
+        #No additional images found
+        if 0 == len(ids_indirect):
+            return []
+        
+        #Translate new IDs to Direct-URLs
+        return map(lambda id: "http://i.imgur.com/{0}{1}".format( id, ids_json[id]), ids_indirect)
 
 
     def get_links(self):
+        """ Extract embedded links from HTML // then check if there are further images which will be lazy-loaded """
+
         f = lambda url: "http://" + re.sub(r'(\w{7})s\.', r'\1.', url)
-        return uniqify(map(f, re.findall(self.LINK_PATTERN, self.data)))
+        direct_links = uniqify(map(f, re.findall(self.LINK_PATTERN, self.data)))
+        self.gallery_foundImgs = len(direct_links)
+        
+        #Imgur Galleryies may contain more images than initially shown. Find the rest now!
+        self.gallery_name = None
+        self.gallery_numImgs = 0
+        try:
+            indirect_links = self.get_indirect_links(direct_links)
+            self.log_debug('Found {0} additional links'.format(len(indirect_links)) )
+        except (TypeError, KeyError, ValueError) as e:
+            #Fail gracefull as we already had some success
+            self.log_error('Processing of additional links unsuccessful - {0}: {1}'.format(type(e).__name__, str(e)))
+            indirect_links = []
+        self.gallery_foundImgs += len(indirect_links)
+        
+        #Check if all images were found and inform the user
+        if self.gallery_numImgs > self.gallery_foundImgs:
+            self.log_error('Could not save all images of this gallery: {0}/{1}'.format( self.gallery_foundImgs, self.gallery_numImgs) )
+            
+        #If we could extract a name, use this to create a specific package
+        if self.gallery_name:
+            self.packages.append( (self.gallery_name, direct_links + indirect_links, self.gallery_name) )
+            return []
+
+        return direct_links + indirect_links