Merge remote-tracking branch 'dstftw/generic-webpage-unescape'

[youtube-dl] / youtube_dl / YoutubeDL.py

diff --git a/youtube_dl/YoutubeDL.py b/youtube_dl/YoutubeDL.py
index 523eba3c1e1287c6813de63825613ab39d63d8a7..c5d08b0bbabb572c3711d1ae8119e7eeb7e40e71 100644 (file)
--- a/youtube_dl/YoutubeDL.py
+++ b/youtube_dl/YoutubeDL.py
@@ -4,6 +4,7 @@
 from __future__ import absolute_import, unicode_literals
 
 import collections
+import datetime
 import errno
 import io
 import json
@@ -147,6 +148,8 @@ class YoutubeDL(object):
                        again.
     cookiefile:        File name where cookies should be read from and dumped to.
     nocheckcertificate:Do not verify SSL certificates
+    prefer_insecure:   Use HTTP instead of HTTPS to retrieve information.
+                       At the moment, this is only supported by YouTube.
     proxy:             URL of the proxy server to use
     socket_timeout:    Time to wait for unresponsive hosts, in seconds
     bidi_workaround:   Work around buggy terminals without bidirectional text
@@ -532,7 +535,7 @@ class YoutubeDL(object):
                 else:
                     raise
         else:
-            self.report_error('no suitable InfoExtractor: %s' % url)
+            self.report_error('no suitable InfoExtractor for URL %s' % url)
 
     def process_ie_result(self, ie_result, download=True, extra_info={}):
         """
@@ -666,6 +669,18 @@ class YoutubeDL(object):
                 if f.get('vcodec') == 'none']
             if audio_formats:
                 return audio_formats[0]
+        elif format_spec == 'bestvideo':
+            video_formats = [
+                f for f in available_formats
+                if f.get('acodec') == 'none']
+            if video_formats:
+                return video_formats[-1]
+        elif format_spec == 'worstvideo':
+            video_formats = [
+                f for f in available_formats
+                if f.get('acodec') == 'none']
+            if video_formats:
+                return video_formats[0]
         else:
             extensions = ['mp4', 'flv', 'webm', '3gp']
             if format_spec in extensions:
@@ -688,6 +703,11 @@ class YoutubeDL(object):
         if 'display_id' not in info_dict and 'id' in info_dict:
             info_dict['display_id'] = info_dict['id']
 
+        if info_dict.get('upload_date') is None and info_dict.get('timestamp') is not None:
+            upload_date = datetime.datetime.utcfromtimestamp(
+                info_dict['timestamp'])
+            info_dict['upload_date'] = upload_date.strftime('%Y%m%d')
+
         # This extractors handle format selection themselves
         if info_dict['extractor'] in ['Youku']:
             if download:
@@ -701,8 +721,11 @@ class YoutubeDL(object):
         else:
             formats = info_dict['formats']
 
+        if not formats:
+            raise ExtractorError('No video formats found!')
+
         # We check that all the formats have the format and format_id fields
-        for (i, format) in enumerate(formats):
+        for i, format in enumerate(formats):
             if format.get('format_id') is None:
                 format['format_id'] = compat_str(i)
             if format.get('format') is None:
@@ -1167,7 +1190,7 @@ class YoutubeDL(object):
 
     def urlopen(self, req):
         """ Start an HTTP download """
-        return self._opener.open(req)
+        return self._opener.open(req, timeout=self._socket_timeout)
 
     def print_debug_header(self):
         if not self.params.get('verbose'):
@@ -1198,7 +1221,7 @@ class YoutubeDL(object):
 
     def _setup_opener(self):
         timeout_val = self.params.get('socket_timeout')
-        timeout = 600 if timeout_val is None else float(timeout_val)
+        self._socket_timeout = 600 if timeout_val is None else float(timeout_val)
 
         opts_cookiefile = self.params.get('cookiefile')
         opts_proxy = self.params.get('proxy')
@@ -1236,7 +1259,3 @@ class YoutubeDL(object):
         # (See https://github.com/rg3/youtube-dl/issues/1309 for details)
         opener.addheaders = []
         self._opener = opener
-
-        # TODO remove this global modification
-        compat_urllib_request.install_opener(opener)
-        socket.setdefaulttimeout(timeout)