Merge pull request #8092 from bpfoley/twitter-thumbnail

[youtube-dl] / youtube_dl / extractor / common.py

diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py
index d2443c93ceaaead83805a72bbefc7a3a70683c0d..770105a5b58013bbcf76e342280b76827724dae4 100644 (file)
--- a/youtube_dl/extractor/common.py
+++ b/youtube_dl/extractor/common.py
@@ -862,6 +862,7 @@ class InfoExtractor(object):
             proto_preference = 0 if determine_protocol(f) in ['http', 'https'] else -0.1
 
             if f.get('vcodec') == 'none':  # audio only
+                preference -= 50
                 if self._downloader.params.get('prefer_free_formats'):
                     ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus']
                 else:
@@ -872,6 +873,8 @@ class InfoExtractor(object):
                 except ValueError:
                     audio_ext_preference = -1
             else:
+                if f.get('acodec') == 'none':  # video only
+                    preference -= 40
                 if self._downloader.params.get('prefer_free_formats'):
                     ORDER = ['flv', 'mp4', 'webm']
                 else:
@@ -973,6 +976,13 @@ class InfoExtractor(object):
         if manifest is False:
             return []
 
+        return self._parse_f4m_formats(
+            manifest, manifest_url, video_id, preference=preference, f4m_id=f4m_id,
+            transform_source=transform_source, fatal=fatal)
+
+    def _parse_f4m_formats(self, manifest, manifest_url, video_id, preference=None, f4m_id=None,
+                           transform_source=lambda s: fix_xml_ampersands(s).strip(),
+                           fatal=True):
         formats = []
         manifest_version = '1.0'
         media_nodes = manifest.findall('{http://ns.adobe.com/f4m/1.0}media')
@@ -998,7 +1008,8 @@ class InfoExtractor(object):
                 # bitrate in f4m downloader
                 if determine_ext(manifest_url) == 'f4m':
                     formats.extend(self._extract_f4m_formats(
-                        manifest_url, video_id, preference, f4m_id, fatal=fatal))
+                        manifest_url, video_id, preference=preference, f4m_id=f4m_id,
+                        transform_source=transform_source, fatal=fatal))
                     continue
             tbr = int_or_none(media_el.attrib.get('bitrate'))
             formats.append({
@@ -1147,8 +1158,8 @@ class InfoExtractor(object):
                 out.append('{%s}%s' % (namespace, c))
         return '/'.join(out)
 
-    def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None):
-        smil = self._download_smil(smil_url, video_id, fatal=fatal)
+    def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None, transform_source=None):
+        smil = self._download_smil(smil_url, video_id, fatal=fatal, transform_source=transform_source)
 
         if smil is False:
             assert not fatal
@@ -1165,10 +1176,10 @@ class InfoExtractor(object):
             return {}
         return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params)
 
-    def _download_smil(self, smil_url, video_id, fatal=True):
+    def _download_smil(self, smil_url, video_id, fatal=True, transform_source=None):
         return self._download_xml(
             smil_url, video_id, 'Downloading SMIL file',
-            'Unable to download SMIL file', fatal=fatal)
+            'Unable to download SMIL file', fatal=fatal, transform_source=transform_source)
 
     def _parse_smil(self, smil, smil_url, video_id, f4m_params=None):
         namespace = self._parse_smil_namespace(smil)
@@ -1454,8 +1465,9 @@ class InfoExtractor(object):
                         continue
                     representation_attrib = adaptation_set.attrib.copy()
                     representation_attrib.update(representation.attrib)
-                    mime_type = representation_attrib.get('mimeType')
-                    content_type = mime_type.split('/')[0] if mime_type else representation_attrib.get('contentType')
+                    # According to page 41 of ISO/IEC 29001-1:2014, @mimeType is mandatory
+                    mime_type = representation_attrib['mimeType']
+                    content_type = mime_type.split('/')[0]
                     if content_type == 'text':
                         # TODO implement WebVTT downloading
                         pass
@@ -1478,6 +1490,7 @@ class InfoExtractor(object):
                         f = {
                             'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
                             'url': base_url,
+                            'ext': mimetype2ext(mime_type),
                             'width': int_or_none(representation_attrib.get('width')),
                             'height': int_or_none(representation_attrib.get('height')),
                             'tbr': int_or_none(representation_attrib.get('bandwidth'), 1000),