From da8c982608c9308765e0960e08fc28cccb74b215 Mon Sep 17 00:00:00 2001
From: Mort Yao <soi@mort.ninja>
Date: Wed, 8 Aug 2018 16:21:49 +0200
Subject: [PATCH] [twitter] fix #2609

---
 src/you_get/extractors/twitter.py | 50 +++++++++++--------------------
 1 file changed, 17 insertions(+), 33 deletions(-)

diff --git a/src/you_get/extractors/twitter.py b/src/you_get/extractors/twitter.py
index 9cc3c5c7..8ed400db 100644
--- a/src/you_get/extractors/twitter.py
+++ b/src/you_get/extractors/twitter.py
@@ -64,42 +64,26 @@ def twitter_download(url, output_dir='.', merge=True, info_only=False, **kwargs)
                               output_dir=output_dir)
 
     except: # extract video
-        # always use i/cards or videos url
-        if not re.match(r'https?://twitter.com/i/', url):
-            url = r1(r'<meta\s*property="og:video:url"\s*content="([^"]+)"', html)
-            if not url:
-                url = 'https://twitter.com/i/videos/%s' % item_id
-            try:
-                html = get_content(url)
-            except:
-                return
+        #i_url = 'https://twitter.com/i/videos/' + item_id
+        #i_content = get_content(i_url)
+        #js_url = r1(r'src="([^"]+)"', i_content)
+        #js_content = get_content(js_url)
+        #authorization = r1(r'"(Bearer [^"]+)"', js_content)
+        authorization = 'Bearer AAAAAAAAAAAAAAAAAAAAANRILgAAAAAAnNwIzUejRCOuH5E6I8xnZz4puTs%3D1Zv7ttfk8LF81IUq16cHjhLTvJu4FA33AGWWjCpTnA'
 
-        data_config = r1(r'data-config="([^"]*)"', html) or \
-            r1(r'data-player-config="([^"]*)"', html)
-        i = json.loads(unescape_html(data_config))
-        if 'video_url' in i:
-            source = i['video_url']
-            item_id = i['tweet_id']
-            page_title = "{} [{}]".format(screen_name, item_id)
-        elif 'playlist' in i:
-            source = i['playlist'][0]['source']
-            if not item_id: page_title = i['playlist'][0]['contentId']
-        elif 'vmap_url' in i:
-            vmap_url = i['vmap_url']
-            vmap = get_content(vmap_url)
-            source = r1(r'<MediaFile>\s*<!\[CDATA\[(.*)\]\]>', vmap)
-            item_id = i['tweet_id']
-            page_title = "{} [{}]".format(screen_name, item_id)
-        elif 'scribe_playlist_url' in i:
-            scribe_playlist_url = i['scribe_playlist_url']
-            return vine_download(scribe_playlist_url, output_dir, merge=merge, info_only=info_only)
+        ga_url = 'https://api.twitter.com/1.1/guest/activate.json'
+        ga_content = post_content(ga_url, headers={'authorization': authorization})
+        guest_token = json.loads(ga_content)['guest_token']
 
-        try:
-            urls = extract_m3u(source)
-        except:
-            urls = [source]
+        api_url = 'https://api.twitter.com/2/timeline/conversation/%s.json?tweet_mode=extended' % item_id
+        api_content = get_content(api_url, headers={'authorization': authorization, 'x-guest-token': guest_token})
+
+        info = json.loads(api_content)
+        variants = info['globalObjects']['tweets'][item_id]['extended_entities']['media'][0]['video_info']['variants']
+        variants = sorted(variants, key=lambda kv: kv.get('bitrate', 0))
+        urls = [ variants[-1]['url'] ]
         size = urls_size(urls)
-        mime, ext = 'video/mp4', 'mp4'
+        mime, ext = variants[-1]['content_type'], 'mp4'
 
         print_info(site_info, page_title, mime, size)
         if not info_only: