X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fdownloader%2Fhls.py;h=0f2c06f40414fd94154c69d2aad365880dcb0e20;hb=HEAD;hp=d0a5f7ba4aa717ed70b7cfa8afe9ff4e22fc2814;hpb=f59746372a1aae3bc80462d34a7e96200be22ab5;p=youtube-dl diff --git a/youtube_dl/downloader/hls.py b/youtube_dl/downloader/hls.py index d0a5f7ba4..0f2c06f40 100644 --- a/youtube_dl/downloader/hls.py +++ b/youtube_dl/downloader/hls.py @@ -1,6 +1,5 @@ from __future__ import unicode_literals -import os.path import re import binascii try: @@ -18,8 +17,6 @@ from ..compat import ( compat_struct_pack, ) from ..utils import ( - encodeFilename, - sanitize_open, parse_m3u8_attributes, update_url_query, ) @@ -62,12 +59,12 @@ class HlsFD(FragmentFD): man_url = info_dict['url'] self.to_screen('[%s] Downloading m3u8 manifest' % self.FD_NAME) - manifest = self.ydl.urlopen(self._prepare_url(info_dict, man_url)).read() - - s = manifest.decode('utf-8', 'ignore') + urlh = self.ydl.urlopen(self._prepare_url(info_dict, man_url)) + man_url = urlh.geturl() + s = urlh.read().decode('utf-8', 'ignore') if not self.can_download(s, info_dict): - if info_dict.get('extra_param_to_segment_url'): + if info_dict.get('extra_param_to_segment_url') or info_dict.get('_decryption_key_url'): self.report_error('pycrypto not found. Please install it.') return False self.report_warning( @@ -78,15 +75,36 @@ class HlsFD(FragmentFD): fd.add_progress_hook(ph) return fd.real_download(filename, info_dict) - total_frags = 0 + def is_ad_fragment_start(s): + return (s.startswith('#ANVATO-SEGMENT-INFO') and 'type=ad' in s + or s.startswith('#UPLYNK-SEGMENT') and s.endswith(',ad')) + + def is_ad_fragment_end(s): + return (s.startswith('#ANVATO-SEGMENT-INFO') and 'type=master' in s + or s.startswith('#UPLYNK-SEGMENT') and s.endswith(',segment')) + + media_frags = 0 + ad_frags = 0 + ad_frag_next = False for line in s.splitlines(): line = line.strip() - if line and not line.startswith('#'): - total_frags += 1 + if not line: + continue + if line.startswith('#'): + if is_ad_fragment_start(line): + ad_frag_next = True + elif is_ad_fragment_end(line): + ad_frag_next = False + continue + if ad_frag_next: + ad_frags += 1 + continue + media_frags += 1 ctx = { 'filename': filename, - 'total_frags': total_frags, + 'total_frags': media_frags, + 'ad_frags': ad_frags, } self._prepare_and_start_frag_download(ctx) @@ -103,64 +121,65 @@ class HlsFD(FragmentFD): media_sequence = 0 decrypt_info = {'METHOD': 'NONE'} byte_range = {} - frags_filenames = [] + frag_index = 0 + ad_frag_next = False for line in s.splitlines(): line = line.strip() if line: if not line.startswith('#'): + if ad_frag_next: + continue + frag_index += 1 + if frag_index <= ctx['fragment_index']: + continue frag_url = ( line if re.match(r'^https?://', line) else compat_urlparse.urljoin(man_url, line)) - frag_name = 'Frag%d' % i - frag_filename = '%s-%s' % (ctx['tmpfilename'], frag_name) if extra_query: frag_url = update_url_query(frag_url, extra_query) count = 0 headers = info_dict.get('http_headers', {}) if byte_range: - headers['Range'] = 'bytes=%d-%d' % (byte_range['start'], byte_range['end']) + headers['Range'] = 'bytes=%d-%d' % (byte_range['start'], byte_range['end'] - 1) while count <= fragment_retries: try: - success = ctx['dl'].download(frag_filename, { - 'url': frag_url, - 'http_headers': headers, - }) + success, frag_content = self._download_fragment( + ctx, frag_url, info_dict, headers) 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 as err: # 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). + # See https://github.com/ytdl-org/youtube-dl/issues/10165, + # https://github.com/ytdl-org/youtube-dl/issues/10448). count += 1 if count <= fragment_retries: - self.report_retry_fragment(err, frag_name, count, fragment_retries) + self.report_retry_fragment(err, frag_index, count, fragment_retries) if count > fragment_retries: if skip_unavailable_fragments: i += 1 media_sequence += 1 - self.report_skip_fragment(frag_name) + self.report_skip_fragment(frag_index) continue self.report_error( 'giving up after %s fragment retries' % fragment_retries) return False if decrypt_info['METHOD'] == 'AES-128': iv = decrypt_info.get('IV') or compat_struct_pack('>8xq', media_sequence) + decrypt_info['KEY'] = decrypt_info.get('KEY') or self.ydl.urlopen( + self._prepare_url(info_dict, info_dict.get('_decryption_key_url') or decrypt_info['URI'])).read() frag_content = AES.new( decrypt_info['KEY'], AES.MODE_CBC, iv).decrypt(frag_content) - ctx['dest_stream'].write(frag_content) - frags_filenames.append(frag_sanitized) + self._append_fragment(ctx, frag_content) # We only download the first fragment during the test if test: break i += 1 media_sequence += 1 elif line.startswith('#EXT-X-KEY'): + decrypt_url = decrypt_info.get('URI') decrypt_info = parse_m3u8_attributes(line[11:]) if decrypt_info['METHOD'] == 'AES-128': if 'IV' in decrypt_info: @@ -170,7 +189,8 @@ class HlsFD(FragmentFD): man_url, decrypt_info['URI']) if extra_query: decrypt_info['URI'] = update_url_query(decrypt_info['URI'], extra_query) - decrypt_info['KEY'] = self.ydl.urlopen(decrypt_info['URI']).read() + if decrypt_url != decrypt_info['URI']: + decrypt_info['KEY'] = None elif line.startswith('#EXT-X-MEDIA-SEQUENCE'): media_sequence = int(line[22:]) elif line.startswith('#EXT-X-BYTERANGE'): @@ -180,10 +200,11 @@ class HlsFD(FragmentFD): 'start': sub_range_start, 'end': sub_range_start + int(splitted_byte_range[0]), } + elif is_ad_fragment_start(line): + ad_frag_next = True + elif is_ad_fragment_end(line): + ad_frag_next = False self._finish_frag_download(ctx) - for frag_file in frags_filenames: - os.remove(encodeFilename(frag_file)) - return True