Fix extraction from youtube.

[youtubedl] / youtube_dl / downloader / hls.py

diff --git a/youtube_dl/downloader/hls.py b/youtube_dl/downloader/hls.py
index 7373ec05fd0d4a1d983f48668229b21d98977581..4def8e2d579c17956f4a95ebb3de914d457c59f4 100644 (file)
--- a/youtube_dl/downloader/hls.py
+++ b/youtube_dl/downloader/hls.py
@@ -1,6 +1,5 @@
 from __future__ import unicode_literals
 
 from __future__ import unicode_literals
 

-import os.path

 import re
 import binascii
 try:
 import re
 import binascii
 try:


@@ -18,8 +17,6 @@ from ..compat import (
     compat_struct_pack,
 )
 from ..utils import (
     compat_struct_pack,
 )
 from ..utils import (

-    encodeFilename,
-    sanitize_open,

     parse_m3u8_attributes,
     update_url_query,
 )
     parse_m3u8_attributes,
     update_url_query,
 )


@@ -34,7 +31,7 @@ class HlsFD(FragmentFD):
     def can_download(manifest, info_dict):
         UNSUPPORTED_FEATURES = (
             r'#EXT-X-KEY:METHOD=(?!NONE|AES-128)',  # encrypted streams [1]
     def can_download(manifest, info_dict):
         UNSUPPORTED_FEATURES = (
             r'#EXT-X-KEY:METHOD=(?!NONE|AES-128)',  # encrypted streams [1]

-            r'#EXT-X-BYTERANGE',  # playlists composed of byte ranges of media files [2]
+            # r'#EXT-X-BYTERANGE',  # playlists composed of byte ranges of media files [2]

 
             # 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)
 
             # 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)


@@ -52,7 +49,9 @@ class HlsFD(FragmentFD):
             # 4. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.5
         )
         check_results = [not re.search(feature, manifest) for feature in UNSUPPORTED_FEATURES]
             # 4. https://tools.ietf.org/html/draft-pantos-http-live-streaming-17#section-4.3.3.5
         )
         check_results = [not re.search(feature, manifest) for feature in UNSUPPORTED_FEATURES]

-        check_results.append(can_decrypt_frag or '#EXT-X-KEY:METHOD=AES-128' not in manifest)
+        is_aes128_enc = '#EXT-X-KEY:METHOD=AES-128' in manifest
+        check_results.append(can_decrypt_frag or not is_aes128_enc)
+        check_results.append(not (is_aes128_enc and r'#EXT-X-BYTERANGE' in manifest))

         check_results.append(not info_dict.get('is_live'))
         return all(check_results)
 
         check_results.append(not info_dict.get('is_live'))
         return all(check_results)
 


@@ -60,11 +59,14 @@ class HlsFD(FragmentFD):
         man_url = info_dict['url']
         self.to_screen('[%s] Downloading m3u8 manifest' % self.FD_NAME)
 
         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 not self.can_download(s, info_dict):

+            if info_dict.get('extra_param_to_segment_url'):
+                self.report_error('pycrypto not found. Please install it.')
+                return False

             self.report_warning(
                 'hlsnative has detected features it does not support, '
                 'extraction will be delegated to ffmpeg')
             self.report_warning(
                 'hlsnative has detected features it does not support, '
                 'extraction will be delegated to ffmpeg')


@@ -73,15 +75,36 @@ class HlsFD(FragmentFD):
                 fd.add_progress_hook(ph)
             return fd.real_download(filename, info_dict)
 
                 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()
         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,
 
         ctx = {
             'filename': filename,

-            'total_frags': total_frags,
+            'total_frags': media_frags,
+            'ad_frags': ad_frags,

         }
 
         self._prepare_and_start_frag_download(ctx)
         }
 
         self._prepare_and_start_frag_download(ctx)


@@ -97,31 +120,34 @@ class HlsFD(FragmentFD):
         i = 0
         media_sequence = 0
         decrypt_info = {'METHOD': 'NONE'}
         i = 0
         media_sequence = 0
         decrypt_info = {'METHOD': 'NONE'}

-        frags_filenames = []
+        byte_range = {}
+        frag_index = 0
+        ad_frag_next = False

         for line in s.splitlines():
             line = line.strip()
             if line:
                 if not line.startswith('#'):
         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_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
                     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'])

                     while count <= fragment_retries:
                         try:
                     while count <= fragment_retries:
                         try:

-                            success = ctx['dl'].download(frag_filename, {
-                                'url': frag_url,
-                                'http_headers': info_dict.get('http_headers'),
-                            })
+                            success, frag_content = self._download_fragment(
+                                ctx, frag_url, info_dict, headers)

                             if not success:
                                 return False
                             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.
                             break
                         except compat_urllib_error.HTTPError as err:
                             # Unavailable (possibly temporary) fragments may be served.


@@ -130,28 +156,30 @@ class HlsFD(FragmentFD):
                             # https://github.com/rg3/youtube-dl/issues/10448).
                             count += 1
                             if count <= fragment_retries:
                             # https://github.com/rg3/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
                     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)
                             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, decrypt_info['URI'])).read()

                         frag_content = AES.new(
                             decrypt_info['KEY'], AES.MODE_CBC, iv).decrypt(frag_content)
                         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'):
                     # 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:
                     decrypt_info = parse_m3u8_attributes(line[11:])
                     if decrypt_info['METHOD'] == 'AES-128':
                         if 'IV' in decrypt_info:


@@ -161,13 +189,22 @@ class HlsFD(FragmentFD):
                                 man_url, decrypt_info['URI'])
                         if extra_query:
                             decrypt_info['URI'] = update_url_query(decrypt_info['URI'], extra_query)
                                 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-MEDIA-SEQUENCE'):
                     media_sequence = int(line[22:])

+                elif line.startswith('#EXT-X-BYTERANGE'):
+                    splitted_byte_range = line[17:].split('@')
+                    sub_range_start = int(splitted_byte_range[1]) if len(splitted_byte_range) == 2 else byte_range['end']
+                    byte_range = {
+                        '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)
 
 
         self._finish_frag_download(ctx)
 

-        for frag_file in frags_filenames:
-            os.remove(encodeFilename(frag_file))
-

         return True
         return True