[youtube] Extract data from multiple clients (#536)

* `player_client` accepts multiple clients
* default `player_client` = `android,web`
* music clients can be specifically requested
* Add IOS `player_client`
* Hide live dash since they can't be downloaded

Closes #501

Authored-by: pukkandan, colethedj

README.md

@@ -1337,12 +1337,12 @@ Some extractors accept additional arguments which can be passed using `--extract
 The following extractors use this feature:
 * **youtube**
     * `skip`: `hls` or `dash` (or both) to skip download of the respective manifests
-    * `player_client`: `web` (default) or `android` (force use the android client fallbacks for video extraction)
-    * `player_skip`: `configs` - skip requests if applicable for client configs and use defaults
+    * `player_client`: Clients to extract video data from - one or more of `web`, `android`, `ios`, `web_music`, `android_music`, `ios_music`. By default, `android,web` is used. If the URL is from `music.youtube.com`, `android,web,android_music,web_music` is used
+    * `player_skip`: `configs` - skip any requests for client configs and use defaults
     * `comment_sort`: `top` or `new` (default) - choose comment sorting mode (on YouTube's side).
     * `max_comments`: maximum amount of comments to download (default all).
     * `max_comment_depth`: maximum depth for nested comments. YouTube supports depths 1 or 2 (default). 
-    
+
 * **funimation**
     * `language`: Languages to extract. Eg: `funimation:language=english,japanese`
     * `version`: The video version to extract - `uncut` or `simulcast`

yt_dlp/downloader/youtube_live_chat.py

@@ -116,7 +116,7 @@ class YoutubeLiveChatFD(FragmentFD):
                     if not success:
                         return False, None, None, None
                     try:
-                        data = ie._extract_yt_initial_data(video_id, raw_fragment.decode('utf-8', 'replace'))
+                        data = ie.extract_yt_initial_data(video_id, raw_fragment.decode('utf-8', 'replace'))
                     except RegexNotFoundError:
                         data = None
                     if not data:
@@ -146,7 +146,7 @@ class YoutubeLiveChatFD(FragmentFD):
         if not success:
             return False
         try:
-            data = ie._extract_yt_initial_data(video_id, raw_fragment.decode('utf-8', 'replace'))
+            data = ie.extract_yt_initial_data(video_id, raw_fragment.decode('utf-8', 'replace'))
         except RegexNotFoundError:
             return False
         continuation_id = try_get(
@@ -155,7 +155,7 @@ class YoutubeLiveChatFD(FragmentFD):
         # no data yet but required to call _append_fragment
         self._append_fragment(ctx, b'')
 
-        ytcfg = ie._extract_ytcfg(video_id, raw_fragment.decode('utf-8', 'replace'))
+        ytcfg = ie.extract_ytcfg(video_id, raw_fragment.decode('utf-8', 'replace'))
 
         if not ytcfg:
             return False
@@ -183,7 +183,7 @@ class YoutubeLiveChatFD(FragmentFD):
                 request_data['currentPlayerState'] = {'playerOffsetMs': str(max(offset - 5000, 0))}
                 if click_tracking_params:
                     request_data['context']['clickTracking'] = {'clickTrackingParams': click_tracking_params}
-                headers = ie._generate_api_headers(ytcfg, visitor_data=visitor_data)
+                headers = ie.generate_api_headers(ytcfg, visitor_data=visitor_data)
                 headers.update({'content-type': 'application/json'})
                 fragment_request_data = json.dumps(request_data, ensure_ascii=False).encode('utf-8') + b'\n'
                 success, continuation_id, offset, click_tracking_params = download_and_parse_fragment(