Extraction: Fix url prefixing

3 files changed, 19 insertions, 8 deletions

diff --git a/youtube/channel.py b/youtube/channel.py
index de75eaa..79b7c9b 100644
--- a/youtube/channel.py
+++ b/youtube/channel.py
@@ -219,8 +219,7 @@ def extract_info(polymer_json, tab):
             else:
                 items = contents    # for search
 
-        # TODO: Fix this URL prefixing shit
-        additional_info = {'author': info['channel_name'], 'author_url': '/channel/' + channel_id}
+        additional_info = {'author': info['channel_name'], 'author_url': 'https://www.youtube.com/channel/' + channel_id}
         info['items'] = [yt_data_extract.renderer_info(renderer, additional_info) for renderer in items]
 
     elif tab == 'about':
@@ -258,8 +257,8 @@ def extract_info(polymer_json, tab):
     return info
 
 def post_process_channel_info(info):
-    info['avatar'] = '/' + info['avatar']
-    info['channel_url'] = '/' + info['channel_url']
+    info['avatar'] = util.prefix_url(info['avatar'])
+    info['channel_url'] = util.prefix_url(info['channel_url'])
     for item in info['items']:
         yt_data_extract.prefix_urls(item)
         yt_data_extract.add_extra_html_info(item)
diff --git a/youtube/util.py b/youtube/util.py
index 2205645..a81ae83 100644
--- a/youtube/util.py
+++ b/youtube/util.py
@@ -317,3 +317,7 @@ def uppercase_escape(s):
      return re.sub(
          r'\\U([0-9a-fA-F]{8})',
          lambda m: chr(int(m.group(1), base=16)), s)
+
+def prefix_url(url):
+    url = url.lstrip('/')     # some urls have // before them, which has a special meaning
+    return '/' + url
diff --git a/youtube/yt_data_extract.py b/youtube/yt_data_extract.py
index 5419084..663edc4 100644
--- a/youtube/yt_data_extract.py
+++ b/youtube/yt_data_extract.py
@@ -2,6 +2,7 @@ from youtube import util
 
 import html
 import json
+import re
 
 # videos (all of type str):
 
@@ -152,15 +153,22 @@ def ajax_info(item_json):
         raise
 
 
+youtube_url_re = re.compile(r'^(?:(?:(?:https?:)?//)?(?:www\.)?youtube\.com)?(/.*)$')
+def normalize_url(url):
+    match = youtube_url_re.fullmatch(url)
+    if match is None:
+        raise Exception()
+
+    return 'https://www.youtube.com' + match.group(1)
 
 def prefix_urls(item):
     try:
-        item['thumbnail'] = '/' + item['thumbnail'].lstrip('/')
+        item['thumbnail'] = util.prefix_url(item['thumbnail'])
     except KeyError:
         pass
 
     try:
-        item['author_url'] = util.URL_ORIGIN + item['author_url']
+        item['author_url'] = util.prefix_url(item['author_url'])
     except KeyError:
         pass
 
@@ -219,7 +227,7 @@ def renderer_info(renderer, additional_info={}):
 
         if 'ownerText' in renderer:
             info['author'] = renderer['ownerText']['runs'][0]['text']
-            info['author_url'] = renderer['ownerText']['runs'][0]['navigationEndpoint']['commandMetadata']['webCommandMetadata']['url']
+            info['author_url'] = normalize_url(renderer['ownerText']['runs'][0]['navigationEndpoint']['commandMetadata']['webCommandMetadata']['url'])
         try:
             overlays = renderer['thumbnailOverlays']
         except KeyError:
@@ -241,7 +249,7 @@ def renderer_info(renderer, additional_info={}):
             if key in ('longBylineText', 'shortBylineText'):
                 info['author'] = get_text(node)
                 try:
-                    info['author_url'] = get_url(node)
+                    info['author_url'] = normalize_url(get_url(node))
                 except KeyError:
                     pass