Update upstream source from tag 'upstream/2019.06.08'

[youtubedl] / youtube_dl / extractor / vk.py

diff --git a/youtube_dl/extractor/vk.py b/youtube_dl/extractor/vk.py
index b52d15ac6f0709dce8a60c2f42d690b49a56d391..f57ed228870b1d094c8c5b42e88f6983671670f6 100644 (file)
--- a/youtube_dl/extractor/vk.py
+++ b/youtube_dl/extractor/vk.py
@@ -3,13 +3,9 @@ from __future__ import unicode_literals
 
 import collections
 import re
 
 import collections
 import re

-import sys

 
 from .common import InfoExtractor
 
 from .common import InfoExtractor

-from ..compat import (
-    compat_str,
-    compat_urlparse,
-)
+from ..compat import compat_urlparse

 from ..utils import (
     clean_html,
     ExtractorError,
 from ..utils import (
     clean_html,
     ExtractorError,


@@ -48,24 +44,9 @@ class VKBaseIE(InfoExtractor):
             'pass': password.encode('cp1251'),
         })
 
             'pass': password.encode('cp1251'),
         })
 

-        # https://new.vk.com/ serves two same remixlhk cookies in Set-Cookie header
-        # and expects the first one to be set rather than second (see
-        # https://github.com/rg3/youtube-dl/issues/9841#issuecomment-227871201).
-        # As of RFC6265 the newer one cookie should be set into cookie store
-        # what actually happens.
-        # We will workaround this VK issue by resetting the remixlhk cookie to
-        # the first one manually.
-        for header, cookies in url_handle.headers.items():
-            if header.lower() != 'set-cookie':
-                continue
-            if sys.version_info[0] >= 3:
-                cookies = cookies.encode('iso-8859-1')
-            cookies = cookies.decode('utf-8')
-            remixlhk = re.search(r'remixlhk=(.+?);.*?\bdomain=(.+?)(?:[,;]|$)', cookies)
-            if remixlhk:
-                value, domain = remixlhk.groups()
-                self._set_cookie(domain, 'remixlhk', value)
-                break
+        # vk serves two same remixlhk cookies in Set-Cookie header and expects
+        # first one to be actually set
+        self._apply_first_set_cookie_header(url_handle, 'remixlhk')

 
         login_page = self._download_webpage(
             'https://login.vk.com/?act=login', None,
 
         login_page = self._download_webpage(
             'https://login.vk.com/?act=login', None,


@@ -103,7 +84,7 @@ class VKIE(VKBaseIE):
             'url': 'http://vk.com/videos-77521?z=video-77521_162222515%2Fclub77521',
             'md5': '7babad3b85ea2e91948005b1b8b0cb84',
             'info_dict': {
             'url': 'http://vk.com/videos-77521?z=video-77521_162222515%2Fclub77521',
             'md5': '7babad3b85ea2e91948005b1b8b0cb84',
             'info_dict': {

-                'id': '162222515',
+                'id': '-77521_162222515',

                 'ext': 'mp4',
                 'title': 'ProtivoGunz - Хуёвая песня',
                 'uploader': 're:(?:Noize MC|Alexander Ilyashenko).*',
                 'ext': 'mp4',
                 'title': 'ProtivoGunz - Хуёвая песня',
                 'uploader': 're:(?:Noize MC|Alexander Ilyashenko).*',


@@ -117,7 +98,7 @@ class VKIE(VKBaseIE):
             'url': 'http://vk.com/video205387401_165548505',
             'md5': '6c0aeb2e90396ba97035b9cbde548700',
             'info_dict': {
             'url': 'http://vk.com/video205387401_165548505',
             'md5': '6c0aeb2e90396ba97035b9cbde548700',
             'info_dict': {

-                'id': '165548505',
+                'id': '205387401_165548505',

                 'ext': 'mp4',
                 'title': 'No name',
                 'uploader': 'Tom Cruise',
                 'ext': 'mp4',
                 'title': 'No name',
                 'uploader': 'Tom Cruise',


@@ -132,7 +113,7 @@ class VKIE(VKBaseIE):
             'url': 'http://vk.com/video_ext.php?oid=32194266&id=162925554&hash=7d8c2e0d5e05aeaa&hd=1',
             'md5': 'c7ce8f1f87bec05b3de07fdeafe21a0a',
             'info_dict': {
             'url': 'http://vk.com/video_ext.php?oid=32194266&id=162925554&hash=7d8c2e0d5e05aeaa&hd=1',
             'md5': 'c7ce8f1f87bec05b3de07fdeafe21a0a',
             'info_dict': {

-                'id': '162925554',
+                'id': '32194266_162925554',

                 'ext': 'mp4',
                 'uploader': 'Vladimir Gavrin',
                 'title': 'Lin Dan',
                 'ext': 'mp4',
                 'uploader': 'Vladimir Gavrin',
                 'title': 'Lin Dan',


@@ -149,7 +130,7 @@ class VKIE(VKBaseIE):
             'md5': 'a590bcaf3d543576c9bd162812387666',
             'note': 'Only available for registered users',
             'info_dict': {
             'md5': 'a590bcaf3d543576c9bd162812387666',
             'note': 'Only available for registered users',
             'info_dict': {

-                'id': '164049491',
+                'id': '-8871596_164049491',

                 'ext': 'mp4',
                 'uploader': 'Триллеры',
                 'title': '► Бойцовский клуб / Fight Club 1999 [HD 720]',
                 'ext': 'mp4',
                 'uploader': 'Триллеры',
                 'title': '► Бойцовский клуб / Fight Club 1999 [HD 720]',


@@ -163,7 +144,7 @@ class VKIE(VKBaseIE):
             'url': 'http://vk.com/hd_kino_mania?z=video-43215063_168067957%2F15c66b9b533119788d',
             'md5': '4d7a5ef8cf114dfa09577e57b2993202',
             'info_dict': {
             'url': 'http://vk.com/hd_kino_mania?z=video-43215063_168067957%2F15c66b9b533119788d',
             'md5': '4d7a5ef8cf114dfa09577e57b2993202',
             'info_dict': {

-                'id': '168067957',
+                'id': '-43215063_168067957',

                 'ext': 'mp4',
                 'uploader': 'Киномания - лучшее из мира кино',
                 'title': ' ',
                 'ext': 'mp4',
                 'uploader': 'Киномания - лучшее из мира кино',
                 'title': ' ',


@@ -177,7 +158,7 @@ class VKIE(VKBaseIE):
             'md5': '0c45586baa71b7cb1d0784ee3f4e00a6',
             'note': 'ivi.ru embed',
             'info_dict': {
             'md5': '0c45586baa71b7cb1d0784ee3f4e00a6',
             'note': 'ivi.ru embed',
             'info_dict': {

-                'id': '60690',
+                'id': '-43215063_169084319',

                 'ext': 'mp4',
                 'title': 'Книга Илая',
                 'duration': 6771,
                 'ext': 'mp4',
                 'title': 'Книга Илая',
                 'duration': 6771,


@@ -191,7 +172,7 @@ class VKIE(VKBaseIE):
             'url': 'https://vk.com/video30481095_171201961?list=8764ae2d21f14088d4',
             'md5': '091287af5402239a1051c37ec7b92913',
             'info_dict': {
             'url': 'https://vk.com/video30481095_171201961?list=8764ae2d21f14088d4',
             'md5': '091287af5402239a1051c37ec7b92913',
             'info_dict': {

-                'id': '171201961',
+                'id': '30481095_171201961',

                 'ext': 'mp4',
                 'title': 'ТюменцевВВ_09.07.2015',
                 'uploader': 'Anton Ivanov',
                 'ext': 'mp4',
                 'title': 'ТюменцевВВ_09.07.2015',
                 'uploader': 'Anton Ivanov',


@@ -206,10 +187,10 @@ class VKIE(VKBaseIE):
             'url': 'https://vk.com/video276849682_170681728',
             'info_dict': {
                 'id': 'V3K4mi0SYkc',
             'url': 'https://vk.com/video276849682_170681728',
             'info_dict': {
                 'id': 'V3K4mi0SYkc',

-                'ext': 'webm',
+                'ext': 'mp4',

                 'title': "DSWD Awards 'Children's Joy Foundation, Inc.' Certificate of Registration and License to Operate",
                 'description': 'md5:bf9c26cfa4acdfb146362682edd3827a',
                 'title': "DSWD Awards 'Children's Joy Foundation, Inc.' Certificate of Registration and License to Operate",
                 'description': 'md5:bf9c26cfa4acdfb146362682edd3827a',

-                'duration': 179,
+                'duration': 178,

                 'upload_date': '20130116',
                 'uploader': "Children's Joy Foundation Inc.",
                 'uploader_id': 'thecjf',
                 'upload_date': '20130116',
                 'uploader': "Children's Joy Foundation Inc.",
                 'uploader_id': 'thecjf',


@@ -239,7 +220,7 @@ class VKIE(VKBaseIE):
             'url': 'http://vk.com/video-110305615_171782105',
             'md5': 'e13fcda136f99764872e739d13fac1d1',
             'info_dict': {
             'url': 'http://vk.com/video-110305615_171782105',
             'md5': 'e13fcda136f99764872e739d13fac1d1',
             'info_dict': {

-                'id': '171782105',
+                'id': '-110305615_171782105',

                 'ext': 'mp4',
                 'title': 'S-Dance, репетиции к The way show',
                 'uploader': 'THE WAY SHOW | 17 апреля',
                 'ext': 'mp4',
                 'title': 'S-Dance, репетиции к The way show',
                 'uploader': 'THE WAY SHOW | 17 апреля',


@@ -254,14 +235,17 @@ class VKIE(VKBaseIE):
         {
             # finished live stream, postlive_mp4
             'url': 'https://vk.com/videos-387766?z=video-387766_456242764%2Fpl_-387766_-2',
         {
             # finished live stream, postlive_mp4
             'url': 'https://vk.com/videos-387766?z=video-387766_456242764%2Fpl_-387766_-2',

-            'md5': '90d22d051fccbbe9becfccc615be6791',

             'info_dict': {
             'info_dict': {

-                'id': '456242764',
+                'id': '-387766_456242764',

                 'ext': 'mp4',
                 'ext': 'mp4',

-                'title': 'ИгроМир 2016 — день 1',
+                'title': 'ИгроМир 2016 День 1 — Игромания Утром',

                 'uploader': 'Игромания',
                 'duration': 5239,
                 'uploader': 'Игромания',
                 'duration': 5239,

-                'view_count': int,
+                # TODO: use act=show to extract view_count
+                # 'view_count': int,
+                'upload_date': '20160929',
+                'uploader_id': '-387766',
+                'timestamp': 1475137527,

             },
         },
         {
             },
         },
         {


@@ -443,8 +427,8 @@ class VKIE(VKBaseIE):
             format_url = url_or_none(format_url)
             if not format_url or not format_url.startswith(('http', '//', 'rtmp')):
                 continue
             format_url = url_or_none(format_url)
             if not format_url or not format_url.startswith(('http', '//', 'rtmp')):
                 continue

-            if (format_id.startswith(('url', 'cache')) or
-                    format_id in ('extra_data', 'live_mp4', 'postlive_mp4')):
+            if (format_id.startswith(('url', 'cache'))
+                    or format_id in ('extra_data', 'live_mp4', 'postlive_mp4')):

                 height = int_or_none(self._search_regex(
                     r'^(?:url|cache)(\d+)', format_id, 'height', default=None))
                 formats.append({
                 height = int_or_none(self._search_regex(
                     r'^(?:url|cache)(\d+)', format_id, 'height', default=None))
                 formats.append({


@@ -465,7 +449,7 @@ class VKIE(VKBaseIE):
         self._sort_formats(formats)
 
         return {
         self._sort_formats(formats)
 
         return {

-            'id': compat_str(data.get('vid') or video_id),
+            'id': video_id,

             'formats': formats,
             'title': title,
             'thumbnail': data.get('jpg'),
             'formats': formats,
             'title': title,
             'thumbnail': data.get('jpg'),