From 9b9d90214a2d9e1745b5a0202068b2180423c051 Mon Sep 17 00:00:00 2001
From: Yen Chi Hsuan <yan12125@gmail.com>
Date: Thu, 14 Jan 2016 14:11:21 +0800
Subject: [PATCH] [WIP] Use YoutubeIE._formats as fallback

---
 youtube_dl/extractor/common.py  |  4 ++--
 youtube_dl/extractor/youtube.py | 28 ++++++++++++++++++----------
 youtube_dl/utils.py             |  2 ++
 3 files changed, 22 insertions(+), 12 deletions(-)

diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py
index b05b22a94..b9f38cda0 100644
--- a/youtube_dl/extractor/common.py
+++ b/youtube_dl/extractor/common.py
@@ -1021,9 +1021,9 @@ class InfoExtractor(object):
                     # TODO: looks like video codec is not always necessarily goes first
                     va_codecs = codecs.split(',')
                     if va_codecs[0]:
-                        f['vcodec'] = va_codecs[0].partition('.')[0]
+                        f['vcodec'] = va_codecs[0]
                     if len(va_codecs) > 1 and va_codecs[1]:
-                        f['acodec'] = va_codecs[1].partition('.')[0]
+                        f['acodec'] = va_codecs[1]
                 resolution = last_info.get('RESOLUTION')
                 if resolution:
                     width_str, height_str = resolution.split('x')
diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index e4f227f19..b3bd91132 100644
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -32,6 +32,7 @@ from ..utils import (
     get_element_by_attribute,
     get_element_by_id,
     int_or_none,
+    mimetype2ext,
     orderedSet,
     parse_duration,
     remove_quotes,
@@ -1083,9 +1084,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                         full_info.update(f)
                         codecs = r.attrib.get('codecs')
                         if codecs:
-                            if full_info.get('acodec') == 'none' and 'vcodec' not in full_info:
+                            if full_info.get('acodec') == 'none':
                                 full_info['vcodec'] = codecs
-                            elif full_info.get('vcodec') == 'none' and 'acodec' not in full_info:
+                            elif full_info.get('vcodec') == 'none':
                                 full_info['acodec'] = codecs
                         formats.append(full_info)
                     else:
@@ -1454,15 +1455,21 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 if 'ratebypass' not in url:
                     url += '&ratebypass=yes'
 
+                dct = {
+                    'format_id': format_id,
+                    'url': url,
+                    'player_url': player_url,
+                }
+                if format_id in self._formats:
+                    dct.update(self._formats[format_id])
+
                 # Some itags are not included in DASH manifest thus corresponding formats will
                 # lack metadata (see https://github.com/rg3/youtube-dl/pull/5993).
                 # Trying to extract metadata from url_encoded_fmt_stream_map entry.
                 mobj = re.search(r'^(?P<width>\d+)[xX](?P<height>\d+)$', url_data.get('size', [''])[0])
                 width, height = (int(mobj.group('width')), int(mobj.group('height'))) if mobj else (None, None)
-                dct = {
-                    'format_id': format_id,
-                    'url': url,
-                    'player_url': player_url,
+
+                more_fields = {
                     'filesize': int_or_none(url_data.get('clen', [None])[0]),
                     'tbr': float_or_none(url_data.get('bitrate', [None])[0], 1000),
                     'width': width,
@@ -1470,13 +1477,16 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     'fps': int_or_none(url_data.get('fps', [None])[0]),
                     'format_note': url_data.get('quality_label', [None])[0] or url_data.get('quality', [None])[0],
                 }
+                for key, value in more_fields.items():
+                    if value:
+                        dct[key] = value
                 type_ = url_data.get('type', [None])[0]
                 if type_:
                     type_split = type_.split(';')
                     kind_ext = type_split[0].split('/')
                     if len(kind_ext) == 2:
-                        kind, ext = kind_ext
-                        dct['ext'] = ext
+                        kind, _ = kind_ext
+                        dct['ext'] = mimetype2ext(type_split[0])
                         if kind in ('audio', 'video'):
                             codecs = None
                             for mobj in re.finditer(
@@ -1494,8 +1504,6 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                                     'acodec': acodec,
                                     'vcodec': vcodec,
                                 })
-                if format_id in self._formats:
-                    dct.update(self._formats[format_id])
                 formats.append(dct)
         elif video_info.get('hlsvp'):
             manifest_url = video_info['hlsvp'][0]
diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index 9c1c0e0bd..178d1dcb3 100644
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -1831,6 +1831,8 @@ def mimetype2ext(mt):
         'x-ms-wmv': 'wmv',
         'x-mp4-fragmented': 'mp4',
         'ttml+xml': 'ttml',
+        '3gpp': '3gp',
+        'x-flv': 'flv',
     }.get(res, res)