2014-09-24 12:16:56 +00:00
|
|
|
from __future__ import unicode_literals
|
|
|
|
|
2016-02-19 18:29:24 +00:00
|
|
|
import os.path
|
2014-09-24 12:16:56 +00:00
|
|
|
import re
|
2016-01-10 19:09:53 +00:00
|
|
|
import binascii
|
|
|
|
try:
|
|
|
|
from Crypto.Cipher import AES
|
|
|
|
can_decrypt_frag = True
|
|
|
|
except ImportError:
|
|
|
|
can_decrypt_frag = False
|
2013-09-23 15:59:27 +00:00
|
|
|
|
2015-07-28 20:28:30 +00:00
|
|
|
from .fragment import FragmentFD
|
2016-05-01 07:56:51 +00:00
|
|
|
from .external import FFmpegFD
|
2015-07-28 20:28:30 +00:00
|
|
|
|
2016-01-10 19:09:53 +00:00
|
|
|
from ..compat import (
|
2016-08-26 21:55:55 +00:00
|
|
|
compat_urllib_error,
|
2016-01-10 19:09:53 +00:00
|
|
|
compat_urlparse,
|
|
|
|
compat_struct_pack,
|
|
|
|
)
|
2014-12-13 11:24:42 +00:00
|
|
|
from ..utils import (
|
2013-09-23 15:59:27 +00:00
|
|
|
encodeFilename,
|
2015-08-30 20:28:36 +00:00
|
|
|
sanitize_open,
|
2016-01-10 19:09:53 +00:00
|
|
|
parse_m3u8_attributes,
|
2016-08-13 21:53:07 +00:00
|
|
|
update_url_query,
|
2013-09-23 15:59:27 +00:00
|
|
|
)
|
|
|
|
|
|
|
|
|
2016-02-19 18:29:24 +00:00
|
|
|
class HlsFD(FragmentFD):
|
|
|
|
""" A limited implementation that does not require ffmpeg """
|
2014-09-24 12:16:56 +00:00
|
|
|
|
2015-07-28 20:28:30 +00:00
|
|
|
FD_NAME = 'hlsnative'
|
|
|
|
|
2016-05-01 07:56:51 +00:00
|
|
|
@staticmethod
|
|
|
|
def can_download(manifest):
|
|
|
|
UNSUPPORTED_FEATURES = (
|
2016-01-10 19:09:53 +00:00
|
|
|
r'#EXT-X-KEY:METHOD=(?!NONE|AES-128)', # encrypted streams [1]
|
2016-05-01 07:56:51 +00:00
|
|
|
r'#EXT-X-BYTERANGE', # playlists composed of byte ranges of media files [2]
|
2016-06-04 20:16:05 +00:00
|
|
|
|
2016-05-09 14:45:03 +00:00
|
|
|
# Live streams heuristic does not always work (e.g. geo restricted to Germany
|
|
|
|
# http://hls-geo.daserste.de/i/videoportal/Film/c_620000/622873/format,716451,716457,716450,716458,716459,.mp4.csmil/index_4_av.m3u8?null=0)
|
2016-05-09 16:16:33 +00:00
|
|
|
# r'#EXT-X-MEDIA-SEQUENCE:(?!0$)', # live streams [3]
|
2016-06-04 20:16:05 +00:00
|
|
|
|
|
|
|
# This heuristic also is not correct since segments may not be appended as well.
|
2016-06-04 20:31:10 +00:00
|
|
|
# Twitch vods of finished streams have EXT-X-PLAYLIST-TYPE:EVENT despite
|
|
|
|
# no segments will definitely be appended to the end of the playlist.
|
2016-06-04 20:16:05 +00:00
|
|
|
# r'#EXT-X-PLAYLIST-TYPE:EVENT', # media segments may be appended to the end of
|
2016-06-04 20:21:43 +00:00
|
|
|
# # event media playlists [4]
|
2016-06-04 20:16:05 +00:00
|
|
|
|
2016-05-01 07:56:51 +00:00
|
|
|
# 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.2.4
|
|
|
|
# 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.2.2
|
|
|
|
# 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.2
|
2016-05-09 14:55:37 +00:00
|
|
|
# 4. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.5
|
2016-05-01 07:56:51 +00:00
|
|
|
)
|
2016-01-10 19:09:53 +00:00
|
|
|
check_results = [not re.search(feature, manifest) for feature in UNSUPPORTED_FEATURES]
|
2016-06-20 14:55:17 +00:00
|
|
|
check_results.append(can_decrypt_frag or '#EXT-X-KEY:METHOD=AES-128' not in manifest)
|
2016-01-10 19:09:53 +00:00
|
|
|
return all(check_results)
|
2016-05-01 07:56:51 +00:00
|
|
|
|
2014-09-24 12:16:56 +00:00
|
|
|
def real_download(self, filename, info_dict):
|
2015-07-28 20:28:30 +00:00
|
|
|
man_url = info_dict['url']
|
|
|
|
self.to_screen('[%s] Downloading m3u8 manifest' % self.FD_NAME)
|
|
|
|
manifest = self.ydl.urlopen(man_url).read()
|
2014-09-24 12:16:56 +00:00
|
|
|
|
2015-07-28 20:28:30 +00:00
|
|
|
s = manifest.decode('utf-8', 'ignore')
|
2016-05-01 07:56:51 +00:00
|
|
|
|
|
|
|
if not self.can_download(s):
|
|
|
|
self.report_warning(
|
|
|
|
'hlsnative has detected features it does not support, '
|
|
|
|
'extraction will be delegated to ffmpeg')
|
|
|
|
fd = FFmpegFD(self.ydl, self.params)
|
|
|
|
for ph in self._progress_hooks:
|
|
|
|
fd.add_progress_hook(ph)
|
|
|
|
return fd.real_download(filename, info_dict)
|
|
|
|
|
2016-01-10 19:09:53 +00:00
|
|
|
total_frags = 0
|
2014-09-24 12:16:56 +00:00
|
|
|
for line in s.splitlines():
|
|
|
|
line = line.strip()
|
|
|
|
if line and not line.startswith('#'):
|
2016-01-10 19:09:53 +00:00
|
|
|
total_frags += 1
|
2014-09-24 12:16:56 +00:00
|
|
|
|
2015-07-28 20:28:30 +00:00
|
|
|
ctx = {
|
2014-09-24 12:16:56 +00:00
|
|
|
'filename': filename,
|
2016-01-10 19:09:53 +00:00
|
|
|
'total_frags': total_frags,
|
2015-07-28 20:28:30 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
self._prepare_and_start_frag_download(ctx)
|
|
|
|
|
2016-08-26 21:55:55 +00:00
|
|
|
fragment_retries = self.params.get('fragment_retries', 0)
|
|
|
|
skip_unavailable_fragments = self.params.get('skip_unavailable_fragments', True)
|
|
|
|
test = self.params.get('test', False)
|
|
|
|
|
2016-08-28 16:51:53 +00:00
|
|
|
extra_query = None
|
2016-08-13 21:53:07 +00:00
|
|
|
extra_param_to_segment_url = info_dict.get('extra_param_to_segment_url')
|
2016-08-28 16:51:53 +00:00
|
|
|
if extra_param_to_segment_url:
|
|
|
|
extra_query = compat_urlparse.parse_qs(extra_param_to_segment_url)
|
2016-01-10 19:09:53 +00:00
|
|
|
i = 0
|
|
|
|
media_sequence = 0
|
|
|
|
decrypt_info = {'METHOD': 'NONE'}
|
2015-07-28 20:28:30 +00:00
|
|
|
frags_filenames = []
|
2016-01-10 19:09:53 +00:00
|
|
|
for line in s.splitlines():
|
|
|
|
line = line.strip()
|
|
|
|
if line:
|
|
|
|
if not line.startswith('#'):
|
|
|
|
frag_url = (
|
|
|
|
line
|
|
|
|
if re.match(r'^https?://', line)
|
|
|
|
else compat_urlparse.urljoin(man_url, line))
|
2016-08-26 21:55:55 +00:00
|
|
|
frag_name = 'Frag%d' % i
|
|
|
|
frag_filename = '%s-%s' % (ctx['tmpfilename'], frag_name)
|
2016-08-28 16:51:53 +00:00
|
|
|
if extra_query:
|
|
|
|
frag_url = update_url_query(frag_url, extra_query)
|
2016-08-26 21:55:55 +00:00
|
|
|
count = 0
|
|
|
|
while count <= fragment_retries:
|
|
|
|
try:
|
|
|
|
success = ctx['dl'].download(frag_filename, {'url': frag_url})
|
|
|
|
if not success:
|
|
|
|
return False
|
|
|
|
down, frag_sanitized = sanitize_open(frag_filename, 'rb')
|
|
|
|
frag_content = down.read()
|
|
|
|
down.close()
|
|
|
|
break
|
|
|
|
except compat_urllib_error.HTTPError:
|
|
|
|
# Unavailable (possibly temporary) fragments may be served.
|
|
|
|
# First we try to retry then either skip or abort.
|
|
|
|
# See https://github.com/rg3/youtube-dl/issues/10165,
|
|
|
|
# https://github.com/rg3/youtube-dl/issues/10448).
|
|
|
|
count += 1
|
|
|
|
if count <= fragment_retries:
|
|
|
|
self.report_retry_fragment(frag_name, count, fragment_retries)
|
|
|
|
if count > fragment_retries:
|
|
|
|
if skip_unavailable_fragments:
|
|
|
|
i += 1
|
|
|
|
media_sequence += 1
|
|
|
|
self.report_skip_fragment(frag_name)
|
|
|
|
continue
|
|
|
|
self.report_error(
|
|
|
|
'giving up after %s fragment retries' % fragment_retries)
|
2016-01-10 19:09:53 +00:00
|
|
|
return False
|
|
|
|
if decrypt_info['METHOD'] == 'AES-128':
|
2016-06-20 14:55:17 +00:00
|
|
|
iv = decrypt_info.get('IV') or compat_struct_pack('>8xq', media_sequence)
|
|
|
|
frag_content = AES.new(
|
|
|
|
decrypt_info['KEY'], AES.MODE_CBC, iv).decrypt(frag_content)
|
2016-01-10 19:09:53 +00:00
|
|
|
ctx['dest_stream'].write(frag_content)
|
|
|
|
frags_filenames.append(frag_sanitized)
|
|
|
|
# We only download the first fragment during the test
|
2016-08-26 21:55:55 +00:00
|
|
|
if test:
|
2016-01-10 19:09:53 +00:00
|
|
|
break
|
|
|
|
i += 1
|
|
|
|
media_sequence += 1
|
|
|
|
elif line.startswith('#EXT-X-KEY'):
|
|
|
|
decrypt_info = parse_m3u8_attributes(line[11:])
|
|
|
|
if decrypt_info['METHOD'] == 'AES-128':
|
|
|
|
if 'IV' in decrypt_info:
|
2016-08-25 07:37:41 +00:00
|
|
|
decrypt_info['IV'] = binascii.unhexlify(decrypt_info['IV'][2:].zfill(32))
|
2016-01-10 19:09:53 +00:00
|
|
|
if not re.match(r'^https?://', decrypt_info['URI']):
|
2016-06-20 14:55:17 +00:00
|
|
|
decrypt_info['URI'] = compat_urlparse.urljoin(
|
|
|
|
man_url, decrypt_info['URI'])
|
2016-08-28 16:51:53 +00:00
|
|
|
if extra_query:
|
|
|
|
decrypt_info['URI'] = update_url_query(decrypt_info['URI'], extra_query)
|
2016-01-10 19:09:53 +00:00
|
|
|
decrypt_info['KEY'] = self.ydl.urlopen(decrypt_info['URI']).read()
|
|
|
|
elif line.startswith('#EXT-X-MEDIA-SEQUENCE'):
|
|
|
|
media_sequence = int(line[22:])
|
2015-07-28 20:28:30 +00:00
|
|
|
|
|
|
|
self._finish_frag_download(ctx)
|
|
|
|
|
|
|
|
for frag_file in frags_filenames:
|
2015-08-30 20:28:36 +00:00
|
|
|
os.remove(encodeFilename(frag_file))
|
2015-07-28 20:28:30 +00:00
|
|
|
|
2014-09-24 12:16:56 +00:00
|
|
|
return True
|