merge #5094: [webtoons] fix extracting comic and episode name with co…

…mmas
mikf · Jan 20, 2024 · 5bf156f · 5bf156f
2 parents b1561a2 + df71888
commit 5bf156f
Show file tree

Hide file tree

Showing 2 changed files with 41 additions and 4 deletions.
diff --git a/gallery_dl/extractor/webtoons.py b/gallery_dl/extractor/webtoons.py
@@ -88,10 +88,20 @@ def _init(self):
 
     def metadata(self, page):
         extr = text.extract_from(page)
-        keywords = extr('<meta name="keywords" content="', '"').split(", ")
         title = extr('<meta property="og:title" content="', '"')
         descr = extr('<meta property="og:description" content="', '"')
 
+        if extr('<div class="subj_info"', '\n'):
+            comic_name = extr('>', '<')
+            episode_name = extr('<h1 class="subj_episode" title="', '"')
+        else:
+            comic_name = episode_name = ""
+
+        if extr('<span class="tx _btnOpenEpisodeList ', '"'):
+            episode = extr('>#', '<')
+        else:
+            episode = ""
+
         if extr('<div class="author_area"', '\n'):
             username = extr('/creator/', '"')
             author_name = extr('<span>', '</span>')
@@ -104,9 +114,9 @@ def metadata(self, page):
             "title_no"    : self.title_no,
             "episode_no"  : self.episode_no,
             "title"       : text.unescape(title),
-            "episode"     : keywords[1],
-            "comic_name"  : text.unescape(keywords[0]),
-            "episode_name": text.unescape(keywords[2]),
+            "episode"     : episode,
+            "comic_name"  : text.unescape(comic_name),
+            "episode_name": text.unescape(episode_name),
             "username"    : username,
             "author_name" : text.unescape(author_name),
             "description" : text.unescape(descr),

diff --git a/test/results/webtoons.py b/test/results/webtoons.py
@@ -61,10 +61,37 @@
 
     "comic_name"  : "I want to be a cute anime girl",
     "episode_name": "209 - The story's story",
+    "episode"     : "214",
     "username"    : "m9huj",
     "author_name" : "Azul Crescent",
 },
 
+{
+    "#url"     : "https://www.webtoons.com/en/canvas/i-want-to-be-a-cute-anime-girl/174-not-194-it-was-a-typo-later/viewer?title_no=349416&episode_no=179",
+    "#category": ("", "webtoons", "episode"),
+    "#class"   : webtoons.WebtoonsEpisodeExtractor,
+    "#count"   : 4,
+
+    "comic_name"  : "I want to be a cute anime girl",
+    "episode_name": "174 (not 194, it was a typo) - Later",
+    "episode"     : "179",
+    "username"    : "m9huj",
+    "author_name" : "Azul Crescent",
+},
+
+{
+    "#url"     : "https://www.webtoons.com/en/canvas/us-over-here/1-the-wheel/viewer?title_no=919536&episode_no=1",
+    "#category": ("", "webtoons", "episode"),
+    "#class"   : webtoons.WebtoonsEpisodeExtractor,
+    "#count"   : 59,
+
+    "comic_name"  : "Us, over here",
+    "episode_name": "1. The Wheel",
+    "episode"     : "1",
+    "username"    : "i94q8",
+    "author_name" : "spin.ani",
+},
+
 {
     "#url"     : "https://www.webtoons.com/en/comedy/live-with-yourself/list?title_no=919",
     "#comment" : "english",