]>
Commit | Line | Data |
---|---|---|
f0b5d6af PH |
1 | from __future__ import unicode_literals |
2 | ||
12b84ac8 | 3 | import os.path |
f0b5d6af | 4 | import re |
e154c651 | 5 | import binascii |
6 | try: | |
7 | from Crypto.Cipher import AES | |
8 | can_decrypt_frag = True | |
9 | except ImportError: | |
10 | can_decrypt_frag = False | |
3bc2ddcc | 11 | |
f9a5affa | 12 | from .fragment import FragmentFD |
0d66bd0e | 13 | from .external import FFmpegFD |
f9a5affa | 14 | |
e154c651 | 15 | from ..compat import ( |
25afc2a7 | 16 | compat_urllib_error, |
e154c651 | 17 | compat_urlparse, |
18 | compat_struct_pack, | |
19 | ) | |
1cc79574 | 20 | from ..utils import ( |
3bc2ddcc | 21 | encodeFilename, |
fcd9e423 | 22 | sanitize_open, |
e154c651 | 23 | parse_m3u8_attributes, |
aaf44a2f | 24 | update_url_query, |
3bc2ddcc JMF |
25 | ) |
26 | ||
27 | ||
12b84ac8 | 28 | class HlsFD(FragmentFD): |
29 | """ A limited implementation that does not require ffmpeg """ | |
f0b5d6af | 30 | |
f9a5affa S |
31 | FD_NAME = 'hlsnative' |
32 | ||
0d66bd0e | 33 | @staticmethod |
6f126d90 | 34 | def can_download(manifest, info_dict): |
0d66bd0e | 35 | UNSUPPORTED_FEATURES = ( |
e154c651 | 36 | r'#EXT-X-KEY:METHOD=(?!NONE|AES-128)', # encrypted streams [1] |
f5974637 | 37 | # r'#EXT-X-BYTERANGE', # playlists composed of byte ranges of media files [2] |
1e236d7e | 38 | |
c15c47d1 S |
39 | # Live streams heuristic does not always work (e.g. geo restricted to Germany |
40 | # 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) | |
2937590e | 41 | # r'#EXT-X-MEDIA-SEQUENCE:(?!0$)', # live streams [3] |
1e236d7e S |
42 | |
43 | # This heuristic also is not correct since segments may not be appended as well. | |
633b444f S |
44 | # Twitch vods of finished streams have EXT-X-PLAYLIST-TYPE:EVENT despite |
45 | # no segments will definitely be appended to the end of the playlist. | |
1e236d7e | 46 | # r'#EXT-X-PLAYLIST-TYPE:EVENT', # media segments may be appended to the end of |
51c4d85c | 47 | # # event media playlists [4] |
1e236d7e | 48 | |
0d66bd0e S |
49 | # 1. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.2.4 |
50 | # 2. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.2.2 | |
51 | # 3. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.2 | |
6104cc29 | 52 | # 4. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.5 |
0d66bd0e | 53 | ) |
e154c651 | 54 | check_results = [not re.search(feature, manifest) for feature in UNSUPPORTED_FEATURES] |
f5974637 RA |
55 | is_aes128_enc = '#EXT-X-KEY:METHOD=AES-128' in manifest |
56 | check_results.append(can_decrypt_frag or not is_aes128_enc) | |
57 | check_results.append(not (is_aes128_enc and r'#EXT-X-BYTERANGE' in manifest)) | |
2bfaf89b | 58 | check_results.append(not info_dict.get('is_live')) |
e154c651 | 59 | return all(check_results) |
0d66bd0e | 60 | |
f0b5d6af | 61 | def real_download(self, filename, info_dict): |
f9a5affa S |
62 | man_url = info_dict['url'] |
63 | self.to_screen('[%s] Downloading m3u8 manifest' % self.FD_NAME) | |
69035555 S |
64 | |
65 | manifest = self.ydl.urlopen(self._prepare_url(info_dict, man_url)).read() | |
f0b5d6af | 66 | |
f9a5affa | 67 | s = manifest.decode('utf-8', 'ignore') |
0d66bd0e | 68 | |
6f126d90 | 69 | if not self.can_download(s, info_dict): |
bfa1073e RA |
70 | if info_dict.get('extra_param_to_segment_url'): |
71 | self.report_error('pycrypto not found. Please install it.') | |
72 | return False | |
2bfaf89b RA |
73 | self.report_warning( |
74 | 'hlsnative has detected features it does not support, ' | |
75 | 'extraction will be delegated to ffmpeg') | |
76 | fd = FFmpegFD(self.ydl, self.params) | |
77 | for ph in self._progress_hooks: | |
78 | fd.add_progress_hook(ph) | |
79 | return fd.real_download(filename, info_dict) | |
0d66bd0e | 80 | |
e154c651 | 81 | total_frags = 0 |
f0b5d6af PH |
82 | for line in s.splitlines(): |
83 | line = line.strip() | |
84 | if line and not line.startswith('#'): | |
e154c651 | 85 | total_frags += 1 |
f0b5d6af | 86 | |
f9a5affa | 87 | ctx = { |
f0b5d6af | 88 | 'filename': filename, |
e154c651 | 89 | 'total_frags': total_frags, |
f9a5affa S |
90 | } |
91 | ||
92 | self._prepare_and_start_frag_download(ctx) | |
93 | ||
25afc2a7 S |
94 | fragment_retries = self.params.get('fragment_retries', 0) |
95 | skip_unavailable_fragments = self.params.get('skip_unavailable_fragments', True) | |
96 | test = self.params.get('test', False) | |
97 | ||
b8079a40 | 98 | extra_query = None |
aaf44a2f | 99 | extra_param_to_segment_url = info_dict.get('extra_param_to_segment_url') |
b8079a40 RA |
100 | if extra_param_to_segment_url: |
101 | extra_query = compat_urlparse.parse_qs(extra_param_to_segment_url) | |
e154c651 | 102 | i = 0 |
103 | media_sequence = 0 | |
104 | decrypt_info = {'METHOD': 'NONE'} | |
f5974637 | 105 | byte_range = {} |
f9a5affa | 106 | frags_filenames = [] |
e154c651 | 107 | for line in s.splitlines(): |
108 | line = line.strip() | |
109 | if line: | |
110 | if not line.startswith('#'): | |
111 | frag_url = ( | |
112 | line | |
113 | if re.match(r'^https?://', line) | |
114 | else compat_urlparse.urljoin(man_url, line)) | |
25afc2a7 S |
115 | frag_name = 'Frag%d' % i |
116 | frag_filename = '%s-%s' % (ctx['tmpfilename'], frag_name) | |
b8079a40 RA |
117 | if extra_query: |
118 | frag_url = update_url_query(frag_url, extra_query) | |
25afc2a7 | 119 | count = 0 |
f5974637 RA |
120 | headers = info_dict.get('http_headers', {}) |
121 | if byte_range: | |
122 | headers['Range'] = 'bytes=%d-%d' % (byte_range['start'], byte_range['end']) | |
25afc2a7 S |
123 | while count <= fragment_retries: |
124 | try: | |
69035555 S |
125 | success = ctx['dl'].download(frag_filename, { |
126 | 'url': frag_url, | |
f5974637 | 127 | 'http_headers': headers, |
69035555 | 128 | }) |
25afc2a7 S |
129 | if not success: |
130 | return False | |
131 | down, frag_sanitized = sanitize_open(frag_filename, 'rb') | |
132 | frag_content = down.read() | |
133 | down.close() | |
134 | break | |
2e99cd30 | 135 | except compat_urllib_error.HTTPError as err: |
25afc2a7 S |
136 | # Unavailable (possibly temporary) fragments may be served. |
137 | # First we try to retry then either skip or abort. | |
138 | # See https://github.com/rg3/youtube-dl/issues/10165, | |
139 | # https://github.com/rg3/youtube-dl/issues/10448). | |
140 | count += 1 | |
141 | if count <= fragment_retries: | |
2e99cd30 | 142 | self.report_retry_fragment(err, frag_name, count, fragment_retries) |
25afc2a7 S |
143 | if count > fragment_retries: |
144 | if skip_unavailable_fragments: | |
145 | i += 1 | |
146 | media_sequence += 1 | |
147 | self.report_skip_fragment(frag_name) | |
148 | continue | |
149 | self.report_error( | |
150 | 'giving up after %s fragment retries' % fragment_retries) | |
e154c651 | 151 | return False |
e154c651 | 152 | if decrypt_info['METHOD'] == 'AES-128': |
8369a4fe S |
153 | iv = decrypt_info.get('IV') or compat_struct_pack('>8xq', media_sequence) |
154 | frag_content = AES.new( | |
155 | decrypt_info['KEY'], AES.MODE_CBC, iv).decrypt(frag_content) | |
e154c651 | 156 | ctx['dest_stream'].write(frag_content) |
157 | frags_filenames.append(frag_sanitized) | |
158 | # We only download the first fragment during the test | |
25afc2a7 | 159 | if test: |
e154c651 | 160 | break |
161 | i += 1 | |
162 | media_sequence += 1 | |
163 | elif line.startswith('#EXT-X-KEY'): | |
164 | decrypt_info = parse_m3u8_attributes(line[11:]) | |
165 | if decrypt_info['METHOD'] == 'AES-128': | |
166 | if 'IV' in decrypt_info: | |
07ea9c9b | 167 | decrypt_info['IV'] = binascii.unhexlify(decrypt_info['IV'][2:].zfill(32)) |
e154c651 | 168 | if not re.match(r'^https?://', decrypt_info['URI']): |
8369a4fe S |
169 | decrypt_info['URI'] = compat_urlparse.urljoin( |
170 | man_url, decrypt_info['URI']) | |
b8079a40 RA |
171 | if extra_query: |
172 | decrypt_info['URI'] = update_url_query(decrypt_info['URI'], extra_query) | |
e154c651 | 173 | decrypt_info['KEY'] = self.ydl.urlopen(decrypt_info['URI']).read() |
174 | elif line.startswith('#EXT-X-MEDIA-SEQUENCE'): | |
175 | media_sequence = int(line[22:]) | |
f5974637 RA |
176 | elif line.startswith('#EXT-X-BYTERANGE'): |
177 | splitted_byte_range = line[17:].split('@') | |
178 | sub_range_start = int(splitted_byte_range[1]) if len(splitted_byte_range) == 2 else byte_range['end'] | |
179 | byte_range = { | |
180 | 'start': sub_range_start, | |
181 | 'end': sub_range_start + int(splitted_byte_range[0]), | |
182 | } | |
f9a5affa S |
183 | |
184 | self._finish_frag_download(ctx) | |
185 | ||
186 | for frag_file in frags_filenames: | |
fcd9e423 | 187 | os.remove(encodeFilename(frag_file)) |
f9a5affa | 188 | |
f0b5d6af | 189 | return True |