[cleanup] Minor fixes

Changelog.md

@@ -785,7 +785,7 @@
 * [build] Improvements
     * Build standalone MacOS packages by [smplayer-dev](https://github.com/smplayer-dev)
     * Release windows exe built with `py2exe`
-    * Enable lazy-extractors in releases. 
+    * Enable lazy-extractors in releases
         * Set env var `YTDLP_NO_LAZY_EXTRACTORS` to forcefully disable this (experimental)
     * Clean up error reporting in update
     * Refactor `pyinst.py`, misc cleanup and improve docs
@@ -1393,7 +1393,7 @@
 * [youtube] Non-fatal alert reporting for unavailable videos page by [coletdjnz](https://github.com/coletdjnz)
 * [twitcasting] Websocket support by [nao20010128nao](https://github.com/nao20010128nao)
 * [mediasite] Extract slides by [fstirlitz](https://github.com/fstirlitz)
-* [funimation] Extract subtitles 
+* [funimation] Extract subtitles
 * [pornhub] Extract `cast`
 * [hotstar] Use server time for authentication instead of local time
 * [EmbedThumbnail] Fix for already downloaded thumbnail
@@ -1489,7 +1489,7 @@
 
 ### 2021.05.20
 
-* **Youtube improvements**: 
+* **Youtube improvements**:
     * Support youtube music `MP`, `VL` and `browse` pages
     * Extract more formats for youtube music by [craftingmod](https://github.com/craftingmod), [coletdjnz](https://github.com/coletdjnz) and [pukkandan](https://github.com/pukkandan)
     * Extract multiple subtitles in same language by [pukkandan](https://github.com/pukkandan) and [tpikonen](https://github.com/tpikonen)
@@ -2031,7 +2031,7 @@
 * **Format Sort:** Added `--format-sort` (`-S`), `--format-sort-force` (`--S-force`) - See [Sorting Formats](README.md#sorting-formats) for details
 * **Format Selection:** See [Format Selection](README.md#format-selection) for details
     * New format selectors: `best*`, `worst*`, `bestvideo*`, `bestaudio*`, `worstvideo*`, `worstaudio*`
-    * Changed video format sorting to show video only files and video+audio files together.
+    * Changed video format sorting to show video only files and video+audio files together
     * Added `--video-multistreams`, `--no-video-multistreams`, `--audio-multistreams`, `--no-audio-multistreams`
     * Added `b`,`w`,`v`,`a` as alias for `best`, `worst`, `video` and `audio` respectively
 * Shortcut Options: Added `--write-link`, `--write-url-link`, `--write-webloc-link`, `--write-desktop-link` by [h-h-h-h](https://github.com/h-h-h-h) - See [Internet Shortcut Options](README.md#internet-shortcut-options) for details

devscripts/make_lazy_extractors.py

@@ -1,6 +1,6 @@
 #!/usr/bin/env python3
-import os
 import optparse
+import os
 import sys
 from inspect import getsource
 

yt_dlp/YoutubeDL.py

@@ -1924,7 +1924,7 @@ class YoutubeDL:
             and download
             and (
                 not can_merge()
-                or info_dict.get('is_live', False)
+                or info_dict.get('is_live') and not self.params.get('live_from_start')
                 or self.outtmpl_dict['default'] == '-'))
         compat = (
             prefer_best

yt_dlp/__init__.py

@@ -869,6 +869,7 @@ def main(argv=None):
 
 
 from .extractor import gen_extractors, list_extractors
+
 __all__ = [
     'main',
     'YoutubeDL',

yt_dlp/extractor/common.py

@@ -1343,7 +1343,7 @@ class InfoExtractor:
         return self._og_search_property('url', html, **kargs)
 
     def _html_extract_title(self, html, name='title', *, fatal=False, **kwargs):
-        return self._html_search_regex(r'(?s)<title>([^<]+)</title>', html, name, fatal=fatal, **kwargs)
+        return self._html_search_regex(r'(?s)<title\b[^>]*>([^<]+)</title>', html, name, fatal=fatal, **kwargs)
 
     def _html_search_meta(self, name, html, display_name=None, fatal=False, **kwargs):
         name = variadic(name)
@@ -1509,8 +1509,9 @@ class InfoExtractor:
                 'url': url_or_none(e.get('contentUrl')),
                 'title': unescapeHTML(e.get('name')),
                 'description': unescapeHTML(e.get('description')),
-                'thumbnails': [{'url': url_or_none(url)}
+                'thumbnails': [{'url': url}
-                               for url in variadic(traverse_obj(e, 'thumbnailUrl', 'thumbnailURL'))],
+                               for url in variadic(traverse_obj(e, 'thumbnailUrl', 'thumbnailURL'))
+                               if url_or_none(url)],
                 'duration': parse_duration(e.get('duration')),
                 'timestamp': unified_timestamp(e.get('uploadDate')),
                 # author can be an instance of 'Organization' or 'Person' types.
@@ -2803,13 +2804,18 @@ class InfoExtractor:
                     mime_type = representation_attrib['mimeType']
                     content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
 
-                    codecs = parse_codecs(representation_attrib.get('codecs', ''))
+                    codec_str = representation_attrib.get('codecs', '')
+                    # Some kind of binary subtitle found in some youtube livestreams
+                    if mime_type == 'application/x-rawcc':
+                        codecs = {'scodec': codec_str}
+                    else:
+                        codecs = parse_codecs(codec_str)
                     if content_type not in ('video', 'audio', 'text'):
                         if mime_type == 'image/jpeg':
                             content_type = mime_type
-                        elif codecs['vcodec'] != 'none':
+                        elif codecs.get('vcodec', 'none') != 'none':
                             content_type = 'video'
-                        elif codecs['acodec'] != 'none':
+                        elif codecs.get('acodec', 'none') != 'none':
                             content_type = 'audio'
                         elif codecs.get('scodec', 'none') != 'none':
                             content_type = 'text'

yt_dlp/extractor/fc2.py

@@ -10,7 +10,6 @@ from ..utils import (
     WebSocketsWrapper,
     js_to_json,
     sanitized_Request,
-    std_headers,
     traverse_obj,
     update_url_query,
     urlencode_postdata,
@@ -207,7 +206,7 @@ class FC2LiveIE(InfoExtractor):
             'Cookie': str(self._get_cookies('https://live.fc2.com/'))[12:],
             'Origin': 'https://live.fc2.com',
             'Accept': '*/*',
-            'User-Agent': std_headers['User-Agent'],
+            'User-Agent': self.get_param('http_headers')['User-Agent'],
         })
 
         self.write_debug('[debug] Sending HLS server request')

yt_dlp/extractor/voicy.py

@@ -1,3 +1,5 @@
+import itertools
+
 from .common import InfoExtractor
 from ..compat import compat_str
 from ..utils import (
@@ -9,8 +11,6 @@ from ..utils import (
     unsmuggle_url,
 )
 
-import itertools
-
 
 class VoicyBaseIE(InfoExtractor):
     def _extract_from_playlist_data(self, value):
@@ -105,7 +105,7 @@ class VoicyChannelIE(VoicyBaseIE):
 
     @classmethod
     def suitable(cls, url):
-        return not VoicyIE.suitable(url) and super(VoicyChannelIE, cls).suitable(url)
+        return not VoicyIE.suitable(url) and super().suitable(url)
 
     def _entries(self, channel_id):
         pager = ''

yt_dlp/utils.py

@@ -714,7 +714,9 @@ def sanitize_path(s, force=False):
 def sanitize_url(url):
     # Prepend protocol-less URLs with `http:` scheme in order to mitigate
     # the number of unwanted failures due to missing protocol
-    if url.startswith('//'):
+    if url is None:
+        return
+    elif url.startswith('//'):
         return 'http:%s' % url
     # Fix some common typos seen so far
     COMMON_TYPOS = (