diff options
Diffstat (limited to 'youtube')
| -rw-r--r-- | youtube/templates/watch.html | 30 | ||||
| -rw-r--r-- | youtube/watch.py | 19 | ||||
| -rw-r--r-- | youtube/yt_data_extract/__init__.py | 2 | ||||
| -rw-r--r-- | youtube/yt_data_extract/watch_extraction.py | 62 |
4 files changed, 101 insertions, 12 deletions
diff --git a/youtube/templates/watch.html b/youtube/templates/watch.html index c722115..04f963d 100644 --- a/youtube/templates/watch.html +++ b/youtube/templates/watch.html @@ -27,6 +27,27 @@ transform: translate(-50%, -50%); } + .live-url-choices{ + height: 360px; + width: 640px; + grid-column: 2; + background-color: var(--video-background-color); + padding: 25px 0px 0px 25px; + } + .live-url-choices ol{ + list-style: none; + padding:0px; + margin:0px; + margin-top: 15px; + } + .live-url-choices input{ + width: 400px; + } + .url-choice-label{ + display: inline-block; + width: 150px; + } + {% if theater_mode %} video{ grid-column: 1 / span 5; @@ -296,6 +317,15 @@ Reload without invidious (for usage of new identity button).</a> {% endif %} </span> </div> + {% elif (video_sources.__len__() == 0 or live) and hls_formats.__len__() != 0 %} + <div class="live-url-choices"> + <span>Copy a url into your video player:</span> + <ol> + {% for fmt in hls_formats %} + <li class="url-choice"><div class="url-choice-label">{{ fmt['video_quality'] }}: </div><input class="url-choice-copy" value="{{ fmt['url'] }}" readonly onclick="this.select();"></li> + {% endfor %} + </ol> + </div> {% else %} <video controls autofocus class="video"> {% for video_source in video_sources %} diff --git a/youtube/watch.py b/youtube/watch.py index 929a1ac..b5636b8 100644 --- a/youtube/watch.py +++ b/youtube/watch.py @@ -242,6 +242,22 @@ def extract_info(video_id, use_invidious, playlist_id=None, index=None): decryption_error = 'Error decrypting url signatures: ' + decryption_error info['playability_error'] = decryption_error + # livestream urls + # sometimes only the livestream urls work soon after the livestream is over + if info['hls_manifest_url'] and (info['live'] or not info['formats']): + manifest = util.fetch_url(info['hls_manifest_url'], + debug_name='hls_manifest.m3u8', + report_text='Fetched hls manifest' + ).decode('utf-8') + + info['hls_formats'], err = yt_data_extract.extract_hls_formats(manifest) + if not err: + info['playability_error'] = None + for fmt in info['hls_formats']: + fmt['video_quality'] = video_quality_string(fmt) + else: + info['hls_formats'] = [] + # check for 403 info['invidious_used'] = False info['invidious_reload_button'] = False @@ -396,7 +412,7 @@ def get_watch_page(video_id=None): download_formats = [] - for format in info['formats']: + for format in (info['formats'] + info['hls_formats']): if format['acodec'] and format['vcodec']: codecs_string = format['acodec'] + ', ' + format['vcodec'] else: @@ -435,6 +451,7 @@ def get_watch_page(video_id=None): download_formats = download_formats, video_info = json.dumps(video_info), video_sources = video_sources, + hls_formats = info['hls_formats'], subtitle_sources = get_subtitle_sources(info), related = info['related_videos'], playlist = info['playlist'], diff --git a/youtube/yt_data_extract/__init__.py b/youtube/yt_data_extract/__init__.py index 3378b8d..8934f74 100644 --- a/youtube/yt_data_extract/__init__.py +++ b/youtube/yt_data_extract/__init__.py @@ -9,4 +9,4 @@ from .everything_else import (extract_channel_info, extract_search_info, from .watch_extraction import (extract_watch_info, get_caption_url, update_with_age_restricted_info, requires_decryption, extract_decryption_function, decrypt_signatures, _formats, - update_format_with_type_info) + update_format_with_type_info, extract_hls_formats) diff --git a/youtube/yt_data_extract/watch_extraction.py b/youtube/yt_data_extract/watch_extraction.py index 9dbb252..5aaa318 100644 --- a/youtube/yt_data_extract/watch_extraction.py +++ b/youtube/yt_data_extract/watch_extraction.py @@ -307,6 +307,18 @@ def _extract_watch_info_desktop(top_level): return info +def update_format_with_codec_info(fmt, codec): + if (codec.startswith('av') + or codec in ('vp9', 'vp8', 'vp8.0', 'h263', 'h264', 'mp4v')): + if codec == 'vp8.0': + codec = 'vp8' + conservative_update(fmt, 'vcodec', codec) + elif (codec.startswith('mp4a') + or codec in ('opus', 'mp3', 'aac', 'dtse', 'ec-3', 'vorbis')): + conservative_update(fmt, 'acodec', codec) + else: + print('Warning: unrecognized codec: ' + codec) + fmt_type_re = re.compile( r'(text|audio|video)/([\w0-9]+); codecs="([\w0-9\.]+(?:, [\w0-9\.]+)*)"') def update_format_with_type_info(fmt, yt_fmt): @@ -319,16 +331,7 @@ def update_format_with_type_info(fmt, yt_fmt): type, fmt['ext'], codecs = match.groups() codecs = codecs.split(', ') for codec in codecs: - if (codec.startswith('av') - or codec in ('vp9', 'vp8', 'vp8.0', 'h263', 'h264', 'mp4v')): - if codec == 'vp8.0': - codec = 'vp8' - conservative_update(fmt, 'vcodec', codec) - elif (codec.startswith('mp4a') - or codec in ('opus', 'mp3', 'aac', 'dtse', 'ec-3', 'vorbis')): - conservative_update(fmt, 'acodec', codec) - else: - print('Warning: unrecognized codec: ' + codec) + update_format_with_codec_info(fmt, codec) if type == 'audio': assert len(codecs) == 1 @@ -337,6 +340,8 @@ def _extract_formats(info, player_response): yt_formats = streaming_data.get('formats', []) + streaming_data.get('adaptiveFormats', []) info['formats'] = [] + info['hls_manifest_url'] = streaming_data.get('hlsManifestUrl') + info['dash_manifest_url'] = streaming_data.get('dashManifestUrl') for yt_fmt in yt_formats: fmt = {} @@ -371,6 +376,43 @@ def _extract_formats(info, player_response): else: info['ip_address'] = None +hls_regex = re.compile(r'[\w_-]+=(?:"[^"]+"|[^",]+),') +def extract_hls_formats(hls_manifest): + '''returns hls_formats, err''' + hls_formats = [] + try: + lines = hls_manifest.splitlines() + i = 0 + while i < len(lines): + if lines[i].startswith('#EXT-X-STREAM-INF'): + fmt = {'acodec': None, 'vcodec': None, 'height': None, + 'width': None, 'fps': None, 'audio_bitrate': None, + 'itag': None, 'file_size': None, + 'audio_sample_rate': None, 'url': None} + properties = lines[i].split(':')[1] + properties += ',' # make regex work for last key-value pair + + for pair in hls_regex.findall(properties): + key, value = pair.rstrip(',').split('=') + if key == 'CODECS': + for codec in value.strip('"').split(','): + update_format_with_codec_info(fmt, codec) + elif key == 'RESOLUTION': + fmt['width'], fmt['height'] = map(int, value.split('x')) + fmt['resolution'] = value + elif key == 'FRAME-RATE': + fmt['fps'] = int(value) + i += 1 + fmt['url'] = lines[i] + assert fmt['url'].startswith('http') + fmt['ext'] = 'm3u8' + hls_formats.append(fmt) + i += 1 + except Exception as e: + traceback.print_exc() + return [], str(e) + return hls_formats, None + def _extract_playability_error(info, player_response, error_prefix=''): if info['formats']: |