From cb2520802d7b8efdc71d6d97aeca984b5f878716 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Sergey=20M=E2=80=A4?= Date: Sat, 22 Apr 2017 07:01:00 +0700 Subject: [PATCH] [extractor/common] Improve m3u8 extraction (closes #12211) * Extract m3u8 parsing to separate method * Improve rendition groups extraction * Build stream name according stream GROUP-ID * Ignore reference to AUDIO group without URI when stream has no CODECS + Add test coverage for parsing m3u8 from #11507, #11995, #12211 and twitch vod --- .gitignore | 1 - test/test_InfoExtractor.py | 308 ++++++++++++++++++- test/testdata/m3u8/pluzz_francetv_11507.m3u8 | 14 + test/testdata/m3u8/teamcoco_11995.m3u8 | 16 + test/testdata/m3u8/toggle_mobile_12211.m3u8 | 13 + test/testdata/m3u8/twitch_vod.m3u8 | 20 ++ youtube_dl/extractor/common.py | 154 ++++++---- 7 files changed, 466 insertions(+), 60 deletions(-) create mode 100644 test/testdata/m3u8/pluzz_francetv_11507.m3u8 create mode 100644 test/testdata/m3u8/teamcoco_11995.m3u8 create mode 100644 test/testdata/m3u8/toggle_mobile_12211.m3u8 create mode 100644 test/testdata/m3u8/twitch_vod.m3u8 diff --git a/.gitignore b/.gitignore index 9ce4b5e2d5..bef084baa4 100644 --- a/.gitignore +++ b/.gitignore @@ -36,7 +36,6 @@ updates_key.pem *.swf *.part *.swp -test/testdata test/local_parameters.json .tox youtube-dl.zsh diff --git a/test/test_InfoExtractor.py b/test/test_InfoExtractor.py index 881197afbe..57205f6ae9 100644 --- a/test/test_InfoExtractor.py +++ b/test/test_InfoExtractor.py @@ -3,12 +3,13 @@ from __future__ import unicode_literals # Allow direct execution +import io import os import sys import unittest sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__)))) -from test.helper import FakeYDL, expect_dict +from test.helper import FakeYDL, expect_dict, expect_value from youtube_dl.extractor.common import InfoExtractor from youtube_dl.extractor import YoutubeIE, get_info_extractor from youtube_dl.utils import encode_data_uri, strip_jsonp, ExtractorError, RegexNotFoundError @@ -175,6 +176,311 @@ jwplayer("mediaplayer").setup({"abouttext":"Visit Indie DB","aboutlink":"http:\/ }] }) + def test_parse_m3u8_formats(self): + _TEST_CASES = [ + ( + # https://github.com/rg3/youtube-dl/issues/11507 + # http://pluzz.francetv.fr/videos/le_ministere.html + 'pluzz_francetv_11507', + 'http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/master.m3u8?caption=2017%2F16%2F156589847-1492488987.m3u8%3Afra%3AFrancais&audiotrack=0%3Afra%3AFrancais', + [{ + 'url': 'http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/master.m3u8?caption=2017%2F16%2F156589847-1492488987.m3u8%3Afra%3AFrancais&audiotrack=0%3Afra%3AFrancais', + 'ext': 'mp4', + 'format_id': 'meta', + 'format_note': 'Quality selection URL', + 'protocol': 'm3u8', + 'preference': -100, + 'resolution': 'multiple' + }, { + 'url': 'http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_0_av.m3u8?null=0', + 'manifest_url': 'http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_0_av.m3u8?null=0', + 'ext': 'mp4', + 'format_id': '180', + 'protocol': 'm3u8', + 'acodec': 'mp4a.40.2', + 'vcodec': 'avc1.66.30', + 'tbr': 180, + 'width': 256, + 'height': 144, + }, { + 'url': 'http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_1_av.m3u8?null=0', + 'manifest_url': 'http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_1_av.m3u8?null=0', + 'ext': 'mp4', + 'format_id': '303', + 'protocol': 'm3u8', + 'acodec': 'mp4a.40.2', + 'vcodec': 'avc1.66.30', + 'tbr': 303, + 'width': 320, + 'height': 180, + }, { + 'url': 'http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_2_av.m3u8?null=0', + 'manifest_url': 'http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_2_av.m3u8?null=0', + 'ext': 'mp4', + 'format_id': '575', + 'protocol': 'm3u8', + 'acodec': 'mp4a.40.2', + 'vcodec': 'avc1.66.30', + 'tbr': 575, + 'width': 512, + 'height': 288, + }, { + 'url': 'http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_3_av.m3u8?null=0', + 'manifest_url': 'http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_3_av.m3u8?null=0', + 'ext': 'mp4', + 'format_id': '831', + 'protocol': 'm3u8', + 'acodec': 'mp4a.40.2', + 'vcodec': 'avc1.77.30', + 'tbr': 831, + 'width': 704, + 'height': 396, + }, { + 'url': 'http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_4_av.m3u8?null=0', + 'manifest_url': 'http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_4_av.m3u8?null=0', + 'ext': 'mp4', + 'protocol': 'm3u8', + 'format_id': '1467', + 'acodec': 'mp4a.40.2', + 'vcodec': 'avc1.77.30', + 'tbr': 1467, + 'width': 1024, + 'height': 576, + }] + ), + ( + # https://github.com/rg3/youtube-dl/issues/11995 + # http://teamcoco.com/video/clueless-gamer-super-bowl-for-honor + 'teamcoco_11995', + 'http://ak.storage-w.teamcococdn.com/cdn/2017-02/98599/ed8f/main.m3u8', + [{ + 'url': 'http://ak.storage-w.teamcococdn.com/cdn/2017-02/98599/ed8f/main.m3u8', + 'ext': 'mp4', + 'format_id': 'meta', + 'format_note': 'Quality selection URL', + 'protocol': 'm3u8', + 'preference': -100, + 'resolution': 'multiple', + }, { + 'url': 'http://ak.storage-w.teamcococdn.com/cdn/2017-02/98599/ed8f/hls/CONAN_020217_Highlight_show-audio-160k_v4.m3u8', + 'ext': 'mp4', + 'format_id': 'audio-0-Default', + 'protocol': 'm3u8', + 'vcodec': 'none', + }, { + 'url': 'http://ak.storage-w.teamcococdn.com/cdn/2017-02/98599/ed8f/hls/CONAN_020217_Highlight_show-audio-64k_v4.m3u8', + 'ext': 'mp4', + 'format_id': 'audio-1-Default', + 'protocol': 'm3u8', + 'vcodec': 'none', + }, { + 'url': 'http://ak.storage-w.teamcococdn.com/cdn/2017-02/98599/ed8f/hls/CONAN_020217_Highlight_show-audio-64k_v4.m3u8', + 'manifest_url': 'http://ak.storage-w.teamcococdn.com/cdn/2017-02/98599/ed8f/hls/CONAN_020217_Highlight_show-audio-64k_v4.m3u8', + 'ext': 'mp4', + 'format_id': '71', + 'protocol': 'm3u8', + 'acodec': 'mp4a.40.5', + 'vcodec': 'none', + 'tbr': 71, + }, { + 'url': 'http://ak.storage-w.teamcococdn.com/cdn/2017-02/98599/ed8f/hls/CONAN_020217_Highlight_show-400k_v4.m3u8', + 'manifest_url': 'http://ak.storage-w.teamcococdn.com/cdn/2017-02/98599/ed8f/hls/CONAN_020217_Highlight_show-400k_v4.m3u8', + 'ext': 'mp4', + 'format_id': '413', + 'protocol': 'm3u8', + 'acodec': 'none', + 'vcodec': 'avc1.42001e', + 'tbr': 413, + 'width': 400, + 'height': 224, + }, { + 'url': 'http://ak.storage-w.teamcococdn.com/cdn/2017-02/98599/ed8f/hls/CONAN_020217_Highlight_show-400k_v4.m3u8', + 'manifest_url': 'http://ak.storage-w.teamcococdn.com/cdn/2017-02/98599/ed8f/hls/CONAN_020217_Highlight_show-400k_v4.m3u8', + 'ext': 'mp4', + 'format_id': '522', + 'protocol': 'm3u8', + 'acodec': 'none', + 'vcodec': 'avc1.42001e', + 'tbr': 522, + 'width': 400, + 'height': 224, + }, { + 'url': 'http://ak.storage-w.teamcococdn.com/cdn/2017-02/98599/ed8f/hls/CONAN_020217_Highlight_show-1m_v4.m3u8', + 'manifest_url': 'http://ak.storage-w.teamcococdn.com/cdn/2017-02/98599/ed8f/hls/CONAN_020217_Highlight_show-1m_v4.m3u8', + 'ext': 'mp4', + 'format_id': '1205', + 'protocol': 'm3u8', + 'acodec': 'none', + 'vcodec': 'avc1.4d001e', + 'tbr': 1205, + 'width': 640, + 'height': 360, + }, { + 'url': 'http://ak.storage-w.teamcococdn.com/cdn/2017-02/98599/ed8f/hls/CONAN_020217_Highlight_show-2m_v4.m3u8', + 'manifest_url': 'http://ak.storage-w.teamcococdn.com/cdn/2017-02/98599/ed8f/hls/CONAN_020217_Highlight_show-2m_v4.m3u8', + 'ext': 'mp4', + 'format_id': '2374', + 'protocol': 'm3u8', + 'acodec': 'none', + 'vcodec': 'avc1.4d001f', + 'tbr': 2374, + 'width': 1024, + 'height': 576, + }] + ), + ( + # https://github.com/rg3/youtube-dl/issues/12211 + # http://video.toggle.sg/en/series/whoopie-s-world/ep3/478601 + 'toggle_mobile_12211', + 'http://cdnapi.kaltura.com/p/2082311/sp/208231100/playManifest/protocol/http/entryId/0_89q6e8ku/format/applehttp/tags/mobile_sd/f/a.m3u8', + [{ + 'url': 'http://cdnapi.kaltura.com/p/2082311/sp/208231100/playManifest/protocol/http/entryId/0_89q6e8ku/format/applehttp/tags/mobile_sd/f/a.m3u8', + 'ext': 'mp4', + 'format_id': 'meta', + 'format_note': 'Quality selection URL', + 'protocol': 'm3u8', + 'preference': -100, + 'resolution': 'multiple' + }, { + 'url': 'http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/2/pv/1/flavorId/0_sa2ntrdg/name/a.mp4/index.m3u8', + 'ext': 'mp4', + 'format_id': 'audio-English', + 'protocol': 'm3u8', + 'language': 'eng', + 'vcodec': 'none', + }, { + 'url': 'http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/2/pv/1/flavorId/0_r7y0nitg/name/a.mp4/index.m3u8', + 'ext': 'mp4', + 'format_id': 'audio-Undefined', + 'protocol': 'm3u8', + 'language': 'und', + 'vcodec': 'none', + }, { + 'url': 'http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/2/pv/1/flavorId/0_qlk9hlzr/name/a.mp4/index.m3u8', + 'manifest_url': 'http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/2/pv/1/flavorId/0_qlk9hlzr/name/a.mp4/index.m3u8', + 'ext': 'mp4', + 'format_id': '155', + 'protocol': 'm3u8', + 'tbr': 155, + 'width': 320, + 'height': 180, + }, { + 'url': 'http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/2/pv/1/flavorId/0_oefackmi/name/a.mp4/index.m3u8', + 'manifest_url': 'http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/2/pv/1/flavorId/0_oefackmi/name/a.mp4/index.m3u8', + 'ext': 'mp4', + 'format_id': '502', + 'protocol': 'm3u8', + 'tbr': 502, + 'width': 480, + 'height': 270, + }, { + 'url': 'http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/12/pv/1/flavorId/0_vyg9pj7k/name/a.mp4/index.m3u8', + 'manifest_url': 'http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/12/pv/1/flavorId/0_vyg9pj7k/name/a.mp4/index.m3u8', + 'ext': 'mp4', + 'format_id': '827', + 'protocol': 'm3u8', + 'tbr': 827, + 'width': 640, + 'height': 360, + }, { + 'url': 'http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/12/pv/1/flavorId/0_50n4psvx/name/a.mp4/index.m3u8', + 'manifest_url': 'http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/12/pv/1/flavorId/0_50n4psvx/name/a.mp4/index.m3u8', + 'ext': 'mp4', + 'format_id': '1396', + 'protocol': 'm3u8', + 'tbr': 1396, + 'width': 854, + 'height': 480, + }] + ), + ( + # http://www.twitch.tv/riotgames/v/6528877 + 'twitch_vod', + 'https://usher.ttvnw.net/vod/6528877?allow_source=true&allow_audio_only=true&allow_spectre=true&player=twitchweb&nauth=%7B%22user_id%22%3Anull%2C%22vod_id%22%3A6528877%2C%22expires%22%3A1492887874%2C%22chansub%22%3A%7B%22restricted_bitrates%22%3A%5B%5D%7D%2C%22privileged%22%3Afalse%2C%22https_required%22%3Afalse%7D&nauthsig=3e29296a6824a0f48f9e731383f77a614fc79bee', + [{ + 'url': 'https://usher.ttvnw.net/vod/6528877?allow_source=true&allow_audio_only=true&allow_spectre=true&player=twitchweb&nauth=%7B%22user_id%22%3Anull%2C%22vod_id%22%3A6528877%2C%22expires%22%3A1492887874%2C%22chansub%22%3A%7B%22restricted_bitrates%22%3A%5B%5D%7D%2C%22privileged%22%3Afalse%2C%22https_required%22%3Afalse%7D&nauthsig=3e29296a6824a0f48f9e731383f77a614fc79bee', + 'ext': 'mp4', + 'format_id': 'meta', + 'format_note': 'Quality selection URL', + 'protocol': 'm3u8', + 'preference': -100, + 'resolution': 'multiple' + }, { + 'url': 'https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/audio_only/index-muted-HM49I092CC.m3u8', + 'manifest_url': 'https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/audio_only/index-muted-HM49I092CC.m3u8', + 'ext': 'mp4', + 'format_id': 'Audio Only', + 'protocol': 'm3u8', + 'acodec': 'mp4a.40.2', + 'vcodec': 'none', + 'tbr': 182, + }, { + 'url': 'https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/mobile/index-muted-HM49I092CC.m3u8', + 'manifest_url': 'https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/mobile/index-muted-HM49I092CC.m3u8', + 'ext': 'mp4', + 'format_id': 'Mobile', + 'protocol': 'm3u8', + 'acodec': 'mp4a.40.2', + 'vcodec': 'avc1.42C00D', + 'tbr': 280, + 'width': 400, + 'height': 226, + }, { + 'url': 'https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/low/index-muted-HM49I092CC.m3u8', + 'manifest_url': 'https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/low/index-muted-HM49I092CC.m3u8', + 'ext': 'mp4', + 'format_id': 'Low', + 'protocol': 'm3u8', + 'acodec': 'mp4a.40.2', + 'vcodec': 'avc1.42C01E', + 'tbr': 628, + 'width': 640, + 'height': 360, + }, { + 'url': 'https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/medium/index-muted-HM49I092CC.m3u8', + 'manifest_url': 'https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/medium/index-muted-HM49I092CC.m3u8', + 'ext': 'mp4', + 'format_id': 'Medium', + 'protocol': 'm3u8', + 'acodec': 'mp4a.40.2', + 'vcodec': 'avc1.42C01E', + 'tbr': 893, + 'width': 852, + 'height': 480, + }, { + 'url': 'https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/high/index-muted-HM49I092CC.m3u8', + 'manifest_url': 'https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/high/index-muted-HM49I092CC.m3u8', + 'ext': 'mp4', + 'format_id': 'High', + 'protocol': 'm3u8', + 'acodec': 'mp4a.40.2', + 'vcodec': 'avc1.42C01F', + 'tbr': 1603, + 'width': 1280, + 'height': 720, + }, { + 'url': 'https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/chunked/index-muted-HM49I092CC.m3u8', + 'manifest_url': 'https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/chunked/index-muted-HM49I092CC.m3u8', + 'ext': 'mp4', + 'format_id': 'Source', + 'protocol': 'm3u8', + 'acodec': 'mp4a.40.2', + 'vcodec': 'avc1.100.31', + 'tbr': 3214, + 'width': 1280, + 'height': 720, + }] + ) + ] + + for m3u8_file, m3u8_url, expected_formats in _TEST_CASES: + with io.open('./test/testdata/m3u8/%s.m3u8' % m3u8_file, + mode='r', encoding='utf-8') as f: + formats = self.ie._parse_m3u8_formats( + f.read(), m3u8_url, ext='mp4') + self.ie._sort_formats(formats) + expect_value(self, formats, expected_formats, None) + if __name__ == '__main__': unittest.main() diff --git a/test/testdata/m3u8/pluzz_francetv_11507.m3u8 b/test/testdata/m3u8/pluzz_francetv_11507.m3u8 new file mode 100644 index 0000000000..0809f5aa07 --- /dev/null +++ b/test/testdata/m3u8/pluzz_francetv_11507.m3u8 @@ -0,0 +1,14 @@ +#EXTM3U + #EXT-X-VERSION:5 + #EXT-X-MEDIA:TYPE=SUBTITLES,GROUP-ID="subs",NAME="Francais",DEFAULT=NO,FORCED=NO,URI="http://replayftv-pmd.francetv.fr/subtitles/2017/16/156589847-1492488987.m3u8",LANGUAGE="fra" + #EXT-X-MEDIA:TYPE=AUDIO,GROUP-ID="aac",LANGUAGE="fra",NAME="Francais",DEFAULT=YES, AUTOSELECT=YES +#EXT-X-STREAM-INF:SUBTITLES="subs",AUDIO="aac",PROGRAM-ID=1,BANDWIDTH=180000,RESOLUTION=256x144,CODECS="avc1.66.30, mp4a.40.2" +http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_0_av.m3u8?null=0 +#EXT-X-STREAM-INF:SUBTITLES="subs",AUDIO="aac",PROGRAM-ID=1,BANDWIDTH=303000,RESOLUTION=320x180,CODECS="avc1.66.30, mp4a.40.2" +http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_1_av.m3u8?null=0 +#EXT-X-STREAM-INF:SUBTITLES="subs",AUDIO="aac",PROGRAM-ID=1,BANDWIDTH=575000,RESOLUTION=512x288,CODECS="avc1.66.30, mp4a.40.2" +http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_2_av.m3u8?null=0 +#EXT-X-STREAM-INF:SUBTITLES="subs",AUDIO="aac",PROGRAM-ID=1,BANDWIDTH=831000,RESOLUTION=704x396,CODECS="avc1.77.30, mp4a.40.2" +http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_3_av.m3u8?null=0 +#EXT-X-STREAM-INF:SUBTITLES="subs",AUDIO="aac",PROGRAM-ID=1,BANDWIDTH=1467000,RESOLUTION=1024x576,CODECS="avc1.77.30, mp4a.40.2" +http://replayftv-vh.akamaihd.net/i/streaming-adaptatif_france-dom-tom/2017/S16/J2/156589847-58f59130c1f52-,standard1,standard2,standard3,standard4,standard5,.mp4.csmil/index_4_av.m3u8?null=0 diff --git a/test/testdata/m3u8/teamcoco_11995.m3u8 b/test/testdata/m3u8/teamcoco_11995.m3u8 new file mode 100644 index 0000000000..a6e4216978 --- /dev/null +++ b/test/testdata/m3u8/teamcoco_11995.m3u8 @@ -0,0 +1,16 @@ +#EXTM3U +#EXT-X-MEDIA:TYPE=AUDIO,GROUP-ID="audio-0",NAME="Default",AUTOSELECT=YES,DEFAULT=YES,URI="hls/CONAN_020217_Highlight_show-audio-160k_v4.m3u8" +#EXT-X-MEDIA:TYPE=AUDIO,GROUP-ID="audio-1",NAME="Default",AUTOSELECT=YES,DEFAULT=YES,URI="hls/CONAN_020217_Highlight_show-audio-64k_v4.m3u8" +#EXT-X-I-FRAME-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=37862000,CODECS="avc1.4d001f",URI="hls/CONAN_020217_Highlight_show-2m_iframe.m3u8" +#EXT-X-I-FRAME-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=18750000,CODECS="avc1.4d001e",URI="hls/CONAN_020217_Highlight_show-1m_iframe.m3u8" +#EXT-X-I-FRAME-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=6535000,CODECS="avc1.42001e",URI="hls/CONAN_020217_Highlight_show-400k_iframe.m3u8" +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=2374000,RESOLUTION=1024x576,CODECS="avc1.4d001f,mp4a.40.2",AUDIO="audio-0" +hls/CONAN_020217_Highlight_show-2m_v4.m3u8 +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=1205000,RESOLUTION=640x360,CODECS="avc1.4d001e,mp4a.40.2",AUDIO="audio-0" +hls/CONAN_020217_Highlight_show-1m_v4.m3u8 +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=522000,RESOLUTION=400x224,CODECS="avc1.42001e,mp4a.40.2",AUDIO="audio-0" +hls/CONAN_020217_Highlight_show-400k_v4.m3u8 +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=413000,RESOLUTION=400x224,CODECS="avc1.42001e,mp4a.40.5",AUDIO="audio-1" +hls/CONAN_020217_Highlight_show-400k_v4.m3u8 +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=71000,CODECS="mp4a.40.5",AUDIO="audio-1" +hls/CONAN_020217_Highlight_show-audio-64k_v4.m3u8 diff --git a/test/testdata/m3u8/toggle_mobile_12211.m3u8 b/test/testdata/m3u8/toggle_mobile_12211.m3u8 new file mode 100644 index 0000000000..69604e683c --- /dev/null +++ b/test/testdata/m3u8/toggle_mobile_12211.m3u8 @@ -0,0 +1,13 @@ +#EXTM3U +#EXT-X-VERSION:4 +#EXT-X-MEDIA:TYPE=AUDIO,GROUP-ID="audio",LANGUAGE="eng",NAME="English",URI="http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/2/pv/1/flavorId/0_sa2ntrdg/name/a.mp4/index.m3u8" +#EXT-X-MEDIA:TYPE=AUDIO,GROUP-ID="audio",LANGUAGE="und",NAME="Undefined",URI="http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/2/pv/1/flavorId/0_r7y0nitg/name/a.mp4/index.m3u8" + +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=155648,RESOLUTION=320x180,AUDIO="audio" +http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/2/pv/1/flavorId/0_qlk9hlzr/name/a.mp4/index.m3u8 +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=502784,RESOLUTION=480x270,AUDIO="audio" +http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/2/pv/1/flavorId/0_oefackmi/name/a.mp4/index.m3u8 +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=827392,RESOLUTION=640x360,AUDIO="audio" +http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/12/pv/1/flavorId/0_vyg9pj7k/name/a.mp4/index.m3u8 +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=1396736,RESOLUTION=854x480,AUDIO="audio" +http://k.toggle.sg/fhls/p/2082311/sp/208231100/serveFlavor/entryId/0_89q6e8ku/v/12/pv/1/flavorId/0_50n4psvx/name/a.mp4/index.m3u8 diff --git a/test/testdata/m3u8/twitch_vod.m3u8 b/test/testdata/m3u8/twitch_vod.m3u8 new file mode 100644 index 0000000000..7617277ca8 --- /dev/null +++ b/test/testdata/m3u8/twitch_vod.m3u8 @@ -0,0 +1,20 @@ +#EXTM3U +#EXT-X-TWITCH-INFO:ORIGIN="s3",CLUSTER="edgecast_vod",REGION="EU",MANIFEST-CLUSTER="edgecast_vod",USER-IP="109.171.17.81" +#EXT-X-MEDIA:TYPE=VIDEO,GROUP-ID="chunked",NAME="Source",AUTOSELECT=YES,DEFAULT=YES +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=3214134,CODECS="avc1.100.31,mp4a.40.2",RESOLUTION="1280x720",VIDEO="chunked" +https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/chunked/index-muted-HM49I092CC.m3u8 +#EXT-X-MEDIA:TYPE=VIDEO,GROUP-ID="high",NAME="High",AUTOSELECT=YES,DEFAULT=YES +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=1603789,CODECS="avc1.42C01F,mp4a.40.2",RESOLUTION="1280x720",VIDEO="high" +https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/high/index-muted-HM49I092CC.m3u8 +#EXT-X-MEDIA:TYPE=VIDEO,GROUP-ID="medium",NAME="Medium",AUTOSELECT=YES,DEFAULT=YES +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=893387,CODECS="avc1.42C01E,mp4a.40.2",RESOLUTION="852x480",VIDEO="medium" +https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/medium/index-muted-HM49I092CC.m3u8 +#EXT-X-MEDIA:TYPE=VIDEO,GROUP-ID="low",NAME="Low",AUTOSELECT=YES,DEFAULT=YES +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=628347,CODECS="avc1.42C01E,mp4a.40.2",RESOLUTION="640x360",VIDEO="low" +https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/low/index-muted-HM49I092CC.m3u8 +#EXT-X-MEDIA:TYPE=VIDEO,GROUP-ID="mobile",NAME="Mobile",AUTOSELECT=YES,DEFAULT=YES +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=280474,CODECS="avc1.42C00D,mp4a.40.2",RESOLUTION="400x226",VIDEO="mobile" +https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/mobile/index-muted-HM49I092CC.m3u8 +#EXT-X-MEDIA:TYPE=VIDEO,GROUP-ID="audio_only",NAME="Audio Only",AUTOSELECT=NO,DEFAULT=NO +#EXT-X-STREAM-INF:PROGRAM-ID=1,BANDWIDTH=182725,CODECS="mp4a.40.2",VIDEO="audio_only" +https://vod.edgecast.hls.ttvnw.net/e5da31ab49_riotgames_15001215120_261543898/audio_only/index-muted-HM49I092CC.m3u8 diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index 61d97ab72e..359c549c5c 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -1312,17 +1312,25 @@ class InfoExtractor(object): entry_protocol='m3u8', preference=None, m3u8_id=None, note=None, errnote=None, fatal=True, live=False): - res = self._download_webpage_handle( m3u8_url, video_id, note=note or 'Downloading m3u8 information', errnote=errnote or 'Failed to download m3u8 information', fatal=fatal) + if res is False: return [] + m3u8_doc, urlh = res m3u8_url = urlh.geturl() + return self._parse_m3u8_formats( + m3u8_doc, m3u8_url, ext=ext, entry_protocol=entry_protocol, + preference=preference, m3u8_id=m3u8_id, live=live) + + def _parse_m3u8_formats(self, m3u8_doc, m3u8_url, ext=None, + entry_protocol='m3u8', preference=None, + m3u8_id=None, live=False): if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access return [] @@ -1333,19 +1341,21 @@ class InfoExtractor(object): if re.match(r'^https?://', u) else compat_urlparse.urljoin(m3u8_url, u)) - # We should try extracting formats only from master playlists [1], i.e. - # playlists that describe available qualities. On the other hand media - # playlists [2] should be returned as is since they contain just the media - # without qualities renditions. + # References: + # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-21 + # 2. https://github.com/rg3/youtube-dl/issues/12211 + + # We should try extracting formats only from master playlists [1, 4.3.4], + # i.e. playlists that describe available qualities. On the other hand + # media playlists [1, 4.3.3] should be returned as is since they contain + # just the media without qualities renditions. # Fortunately, master playlist can be easily distinguished from media - # playlist based on particular tags availability. As of [1, 2] master - # playlist tags MUST NOT appear in a media playist and vice versa. - # As of [3] #EXT-X-TARGETDURATION tag is REQUIRED for every media playlist - # and MUST NOT appear in master playlist thus we can clearly detect media - # playlist with this criterion. - # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.4 - # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3 - # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.1 + # playlist based on particular tags availability. As of [1, 4.3.3, 4.3.4] + # master playlist tags MUST NOT appear in a media playist and vice versa. + # As of [1, 4.3.3.1] #EXT-X-TARGETDURATION tag is REQUIRED for every + # media playlist and MUST NOT appear in master playlist thus we can + # clearly detect media playlist with this criterion. + if '#EXT-X-TARGETDURATION' in m3u8_doc: # media playlist, return as is return [{ 'url': m3u8_url, @@ -1354,52 +1364,67 @@ class InfoExtractor(object): 'protocol': entry_protocol, 'preference': preference, }] - audio_in_video_stream = {} - last_info = {} - last_media = {} + + groups = {} + last_stream_inf = {} + + def extract_media(x_media_line): + media = parse_m3u8_attributes(x_media_line) + # As per [1, 4.3.4.1] TYPE, GROUP-ID and NAME are REQUIRED + media_type, group_id, name = media.get('TYPE'), media.get('GROUP-ID'), media.get('NAME') + if not (media_type and group_id and name): + return + groups.setdefault(group_id, []).append(media) + if media_type not in ('VIDEO', 'AUDIO'): + return + media_url = media.get('URI') + if media_url: + format_id = [] + for v in (group_id, name): + if v: + format_id.append(v) + f = { + 'format_id': '-'.join(format_id), + 'url': format_url(media_url), + 'language': media.get('LANGUAGE'), + 'ext': ext, + 'protocol': entry_protocol, + 'preference': preference, + } + if media_type == 'AUDIO': + f['vcodec'] = 'none' + formats.append(f) + + def build_stream_name(): + # Despite specification does not mention NAME attribute for + # EXT-X-STREAM-INF it still sometimes may be present + stream_name = last_stream_inf.get('NAME') + if stream_name: + return stream_name + # If there is no NAME in EXT-X-STREAM-INF it will be obtained + # from corresponding rendition group + stream_group_id = last_stream_inf.get('VIDEO') + if not stream_group_id: + return + stream_group = groups.get(stream_group_id) + if not stream_group: + return stream_group_id + rendition = stream_group[0] + return rendition.get('NAME') or stream_group_id + for line in m3u8_doc.splitlines(): if line.startswith('#EXT-X-STREAM-INF:'): - last_info = parse_m3u8_attributes(line) + last_stream_inf = parse_m3u8_attributes(line) elif line.startswith('#EXT-X-MEDIA:'): - media = parse_m3u8_attributes(line) - media_type = media.get('TYPE') - if media_type in ('VIDEO', 'AUDIO'): - group_id = media.get('GROUP-ID') - media_url = media.get('URI') - if media_url: - format_id = [] - for v in (group_id, media.get('NAME')): - if v: - format_id.append(v) - f = { - 'format_id': '-'.join(format_id), - 'url': format_url(media_url), - 'language': media.get('LANGUAGE'), - 'ext': ext, - 'protocol': entry_protocol, - 'preference': preference, - } - if media_type == 'AUDIO': - f['vcodec'] = 'none' - if group_id and not audio_in_video_stream.get(group_id): - audio_in_video_stream[group_id] = False - formats.append(f) - else: - # When there is no URI in EXT-X-MEDIA let this tag's - # data be used by regular URI lines below - last_media = media - if media_type == 'AUDIO' and group_id: - audio_in_video_stream[group_id] = True + extract_media(line) elif line.startswith('#') or not line.strip(): continue else: - tbr = int_or_none(last_info.get('AVERAGE-BANDWIDTH') or last_info.get('BANDWIDTH'), scale=1000) + tbr = int_or_none(last_stream_inf.get('AVERAGE-BANDWIDTH') or last_stream_inf.get('BANDWIDTH'), scale=1000) format_id = [] if m3u8_id: format_id.append(m3u8_id) - # Despite specification does not mention NAME attribute for - # EXT-X-STREAM-INF it still sometimes may be present - stream_name = last_info.get('NAME') or last_media.get('NAME') + stream_name = build_stream_name() # Bandwidth of live streams may differ over time thus making # format_id unpredictable. So it's better to keep provided # format_id intact. @@ -1412,11 +1437,11 @@ class InfoExtractor(object): 'manifest_url': manifest_url, 'tbr': tbr, 'ext': ext, - 'fps': float_or_none(last_info.get('FRAME-RATE')), + 'fps': float_or_none(last_stream_inf.get('FRAME-RATE')), 'protocol': entry_protocol, 'preference': preference, } - resolution = last_info.get('RESOLUTION') + resolution = last_stream_inf.get('RESOLUTION') if resolution: mobj = re.search(r'(?P\d+)[xX](?P\d+)', resolution) if mobj: @@ -1432,13 +1457,26 @@ class InfoExtractor(object): 'vbr': vbr, 'abr': abr, }) - f.update(parse_codecs(last_info.get('CODECS'))) - if audio_in_video_stream.get(last_info.get('AUDIO')) is False and f['vcodec'] != 'none': - # TODO: update acodec for audio only formats with the same GROUP-ID - f['acodec'] = 'none' + codecs = parse_codecs(last_stream_inf.get('CODECS')) + f.update(codecs) + audio_group_id = last_stream_inf.get('AUDIO') + # As per [1, 4.3.4.1.1] any EXT-X-STREAM-INF tag which + # references a rendition group MUST have a CODECS attribute. + # However, this is not always respected, for example, [2] + # contains EXT-X-STREAM-INF tag which references AUDIO + # rendition group but does not have CODECS and despite + # referencing audio group an audio group, it represents + # a complete (with audio and video) format. So, for such cases + # we will ignore references to rendition groups and treat them + # as complete formats. + if audio_group_id and codecs and f.get('vcodec') != 'none': + audio_group = groups.get(audio_group_id) + if audio_group and audio_group[0].get('URI'): + # TODO: update acodec for audio only formats with + # the same GROUP-ID + f['acodec'] = 'none' formats.append(f) - last_info = {} - last_media = {} + last_stream_inf = {} return formats @staticmethod