]> Raphaël G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/niconico.py
Update changelog.
[youtubedl] / youtube_dl / extractor / niconico.py
index c0c139b5df16ce900ba6920a1a004bc433eab4e9..df7f528be2d4c8da7cfe826609608754f2f8e088 100644 (file)
@@ -1,16 +1,27 @@
-# encoding: utf-8
+# coding: utf-8
 from __future__ import unicode_literals
 
-import re
+import json
+import datetime
 
 from .common import InfoExtractor
-from ..utils import (
-    compat_urllib_parse,
-    compat_urllib_request,
+from ..compat import (
+    compat_parse_qs,
     compat_urlparse,
-    unified_strdate,
-    parse_duration,
+)
+from ..utils import (
+    determine_ext,
+    dict_get,
+    ExtractorError,
     int_or_none,
+    float_or_none,
+    parse_duration,
+    parse_iso8601,
+    remove_start,
+    try_get,
+    unified_timestamp,
+    urlencode_postdata,
+    xpath_text,
 )
 
 
@@ -18,130 +29,442 @@ class NiconicoIE(InfoExtractor):
     IE_NAME = 'niconico'
     IE_DESC = 'ニコニコ動画'
 
-    _TEST = {
+    _TESTS = [{
         'url': 'http://www.nicovideo.jp/watch/sm22312215',
         'md5': 'd1a75c0823e2f629128c43e1212760f9',
         'info_dict': {
             'id': 'sm22312215',
             'ext': 'mp4',
             'title': 'Big Buck Bunny',
+            'thumbnail': r're:https?://.*',
             'uploader': 'takuya0301',
             'uploader_id': '2698420',
             'upload_date': '20131123',
+            'timestamp': int,  # timestamp is unstable
             'description': '(c) copyright 2008, Blender Foundation / www.bigbuckbunny.org',
             'duration': 33,
+            'view_count': int,
+            'comment_count': int,
         },
-        'params': {
-            'username': 'ydl.niconico@gmail.com',
-            'password': 'youtube-dl',
+        'skip': 'Requires an account',
+    }, {
+        # File downloaded with and without credentials are different, so omit
+        # the md5 field
+        'url': 'http://www.nicovideo.jp/watch/nm14296458',
+        'info_dict': {
+            'id': 'nm14296458',
+            'ext': 'swf',
+            'title': '【鏡音リン】Dance on media【オリジナル】take2!',
+            'description': 'md5:689f066d74610b3b22e0f1739add0f58',
+            'thumbnail': r're:https?://.*',
+            'uploader': 'りょうた',
+            'uploader_id': '18822557',
+            'upload_date': '20110429',
+            'timestamp': 1304065916,
+            'duration': 209,
         },
-    }
+        'skip': 'Requires an account',
+    }, {
+        # 'video exists but is marked as "deleted"
+        # md5 is unstable
+        'url': 'http://www.nicovideo.jp/watch/sm10000',
+        'info_dict': {
+            'id': 'sm10000',
+            'ext': 'unknown_video',
+            'description': 'deleted',
+            'title': 'ドラえもんエターナル第3話「決戦第3新東京市」<前編>',
+            'thumbnail': r're:https?://.*',
+            'upload_date': '20071224',
+            'timestamp': int,  # timestamp field has different value if logged in
+            'duration': 304,
+            'view_count': int,
+        },
+        'skip': 'Requires an account',
+    }, {
+        'url': 'http://www.nicovideo.jp/watch/so22543406',
+        'info_dict': {
+            'id': '1388129933',
+            'ext': 'mp4',
+            'title': '【第1回】RADIOアニメロミックス ラブライブ!~のぞえりRadio Garden~',
+            'description': 'md5:b27d224bb0ff53d3c8269e9f8b561cf1',
+            'thumbnail': r're:https?://.*',
+            'timestamp': 1388851200,
+            'upload_date': '20140104',
+            'uploader': 'アニメロチャンネル',
+            'uploader_id': '312',
+        },
+        'skip': 'The viewing period of the video you were searching for has expired.',
+    }, {
+        # video not available via `getflv`; "old" HTML5 video
+        'url': 'http://www.nicovideo.jp/watch/sm1151009',
+        'md5': '8fa81c364eb619d4085354eab075598a',
+        'info_dict': {
+            'id': 'sm1151009',
+            'ext': 'mp4',
+            'title': 'マスターシステム本体内蔵のスペハリのメインテーマ(PSG版)',
+            'description': 'md5:6ee077e0581ff5019773e2e714cdd0b7',
+            'thumbnail': r're:https?://.*',
+            'duration': 184,
+            'timestamp': 1190868283,
+            'upload_date': '20070927',
+            'uploader': 'denden2',
+            'uploader_id': '1392194',
+            'view_count': int,
+            'comment_count': int,
+        },
+        'skip': 'Requires an account',
+    }, {
+        # "New" HTML5 video
+        # md5 is unstable
+        'url': 'http://www.nicovideo.jp/watch/sm31464864',
+        'info_dict': {
+            'id': 'sm31464864',
+            'ext': 'mp4',
+            'title': '新作TVアニメ「戦姫絶唱シンフォギアAXZ」PV 最高画質',
+            'description': 'md5:e52974af9a96e739196b2c1ca72b5feb',
+            'timestamp': 1498514060,
+            'upload_date': '20170626',
+            'uploader': 'ゲスト',
+            'uploader_id': '40826363',
+            'thumbnail': r're:https?://.*',
+            'duration': 198,
+            'view_count': int,
+            'comment_count': int,
+        },
+        'skip': 'Requires an account',
+    }, {
+        # Video without owner
+        'url': 'http://www.nicovideo.jp/watch/sm18238488',
+        'md5': 'd265680a1f92bdcbbd2a507fc9e78a9e',
+        'info_dict': {
+            'id': 'sm18238488',
+            'ext': 'mp4',
+            'title': '【実写版】ミュータントタートルズ',
+            'description': 'md5:15df8988e47a86f9e978af2064bf6d8e',
+            'timestamp': 1341160408,
+            'upload_date': '20120701',
+            'uploader': None,
+            'uploader_id': None,
+            'thumbnail': r're:https?://.*',
+            'duration': 5271,
+            'view_count': int,
+            'comment_count': int,
+        },
+        'skip': 'Requires an account',
+    }, {
+        'url': 'http://sp.nicovideo.jp/watch/sm28964488?ss_pos=1&cp_in=wt_tg',
+        'only_matching': True,
+    }]
 
-    _VALID_URL = r'https?://(?:www\.|secure\.)?nicovideo\.jp/watch/((?:[a-z]{2})?[0-9]+)'
+    _VALID_URL = r'https?://(?:www\.|secure\.|sp\.)?nicovideo\.jp/watch/(?P<id>(?:[a-z]{2})?[0-9]+)'
     _NETRC_MACHINE = 'niconico'
-    # Determine whether the downloader uses authentication to download video
-    _AUTHENTICATE = False
 
     def _real_initialize(self):
-        if self._downloader.params.get('username', None) is not None:
-            self._AUTHENTICATE = True
-
-        if self._AUTHENTICATE:
-            self._login()
+        self._login()
 
     def _login(self):
         (username, password) = self._get_login_info()
+        # No authentication to be performed
+        if not username:
+            return True
 
         # Log in
+        login_ok = True
         login_form_strs = {
-            'mail': username,
+            'mail_tel': username,
             'password': password,
         }
-        # Convert to UTF-8 *before* urlencode because Python 2.x's urlencode
-        # chokes on unicode
-        login_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k, v in login_form_strs.items())
-        login_data = compat_urllib_parse.urlencode(login_form).encode('utf-8')
-        request = compat_urllib_request.Request(
-            'https://secure.nicovideo.jp/secure/login', login_data)
-        login_results = self._download_webpage(
-            request, None, note='Logging in', errnote='Unable to log in')
-        if re.search(r'(?i)<h1 class="mb8p4">Log in error</h1>', login_results) is not None:
+        urlh = self._request_webpage(
+            'https://account.nicovideo.jp/api/v1/login', None,
+            note='Logging in', errnote='Unable to log in',
+            data=urlencode_postdata(login_form_strs))
+        if urlh is False:
+            login_ok = False
+        else:
+            parts = compat_urlparse.urlparse(urlh.geturl())
+            if compat_parse_qs(parts.query).get('message', [None])[0] == 'cant_login':
+                login_ok = False
+        if not login_ok:
             self._downloader.report_warning('unable to log in: bad username or password')
-            return False
-        return True
+        return login_ok
+
+    def _extract_format_for_quality(self, api_data, video_id, audio_quality, video_quality):
+        def yesno(boolean):
+            return 'yes' if boolean else 'no'
+
+        session_api_data = api_data['video']['dmcInfo']['session_api']
+        session_api_endpoint = session_api_data['urls'][0]
+
+        format_id = '-'.join(map(lambda s: remove_start(s['id'], 'archive_'), [video_quality, audio_quality]))
+
+        session_response = self._download_json(
+            session_api_endpoint['url'], video_id,
+            query={'_format': 'json'},
+            headers={'Content-Type': 'application/json'},
+            note='Downloading JSON metadata for %s' % format_id,
+            data=json.dumps({
+                'session': {
+                    'client_info': {
+                        'player_id': session_api_data['player_id'],
+                    },
+                    'content_auth': {
+                        'auth_type': session_api_data['auth_types'][session_api_data['protocols'][0]],
+                        'content_key_timeout': session_api_data['content_key_timeout'],
+                        'service_id': 'nicovideo',
+                        'service_user_id': session_api_data['service_user_id']
+                    },
+                    'content_id': session_api_data['content_id'],
+                    'content_src_id_sets': [{
+                        'content_src_ids': [{
+                            'src_id_to_mux': {
+                                'audio_src_ids': [audio_quality['id']],
+                                'video_src_ids': [video_quality['id']],
+                            }
+                        }]
+                    }],
+                    'content_type': 'movie',
+                    'content_uri': '',
+                    'keep_method': {
+                        'heartbeat': {
+                            'lifetime': session_api_data['heartbeat_lifetime']
+                        }
+                    },
+                    'priority': session_api_data['priority'],
+                    'protocol': {
+                        'name': 'http',
+                        'parameters': {
+                            'http_parameters': {
+                                'parameters': {
+                                    'http_output_download_parameters': {
+                                        'use_ssl': yesno(session_api_endpoint['is_ssl']),
+                                        'use_well_known_port': yesno(session_api_endpoint['is_well_known_port']),
+                                    }
+                                }
+                            }
+                        }
+                    },
+                    'recipe_id': session_api_data['recipe_id'],
+                    'session_operation_auth': {
+                        'session_operation_auth_by_signature': {
+                            'signature': session_api_data['signature'],
+                            'token': session_api_data['token'],
+                        }
+                    },
+                    'timing_constraint': 'unlimited'
+                }
+            }))
+
+        resolution = video_quality.get('resolution', {})
+
+        return {
+            'url': session_response['data']['session']['content_uri'],
+            'format_id': format_id,
+            'ext': 'mp4',  # Session API are used in HTML5, which always serves mp4
+            'abr': float_or_none(audio_quality.get('bitrate'), 1000),
+            'vbr': float_or_none(video_quality.get('bitrate'), 1000),
+            'height': resolution.get('height'),
+            'width': resolution.get('width'),
+        }
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        video_id = mobj.group(1)
+        video_id = self._match_id(url)
+
+        # Get video webpage. We are not actually interested in it for normal
+        # cases, but need the cookies in order to be able to download the
+        # info webpage
+        webpage, handle = self._download_webpage_handle(
+            'http://www.nicovideo.jp/watch/' + video_id, video_id)
+        if video_id.startswith('so'):
+            video_id = self._match_id(handle.geturl())
 
-        # Get video webpage. We are not actually interested in it, but need
-        # the cookies in order to be able to download the info webpage
-        self._download_webpage('http://www.nicovideo.jp/watch/' + video_id, video_id)
+        api_data = self._parse_json(self._html_search_regex(
+            'data-api-data="([^"]+)"', webpage,
+            'API data', default='{}'), video_id)
 
-        video_info = self._download_xml(
-            'http://ext.nicovideo.jp/api/getthumbinfo/' + video_id, video_id,
-            note='Downloading video info page')
+        def _format_id_from_url(video_url):
+            return 'economy' if video_real_url.endswith('low') else 'normal'
 
-        if self._AUTHENTICATE:
+        try:
+            video_real_url = api_data['video']['smileInfo']['url']
+        except KeyError:  # Flash videos
             # Get flv info
             flv_info_webpage = self._download_webpage(
-                'http://flapi.nicovideo.jp/api/getflv?v=' + video_id,
+                'http://flapi.nicovideo.jp/api/getflv/' + video_id + '?as3=1',
                 video_id, 'Downloading flv info')
+
+            flv_info = compat_urlparse.parse_qs(flv_info_webpage)
+            if 'url' not in flv_info:
+                if 'deleted' in flv_info:
+                    raise ExtractorError('The video has been deleted.',
+                                         expected=True)
+                elif 'closed' in flv_info:
+                    raise ExtractorError('Niconico videos now require logging in',
+                                         expected=True)
+                elif 'error' in flv_info:
+                    raise ExtractorError('%s reports error: %s' % (
+                        self.IE_NAME, flv_info['error'][0]), expected=True)
+                else:
+                    raise ExtractorError('Unable to find video URL')
+
+            video_info_xml = self._download_xml(
+                'http://ext.nicovideo.jp/api/getthumbinfo/' + video_id,
+                video_id, note='Downloading video info page')
+
+            def get_video_info(items):
+                if not isinstance(items, list):
+                    items = [items]
+                for item in items:
+                    ret = xpath_text(video_info_xml, './/' + item)
+                    if ret:
+                        return ret
+
+            video_real_url = flv_info['url'][0]
+
+            extension = get_video_info('movie_type')
+            if not extension:
+                extension = determine_ext(video_real_url)
+
+            formats = [{
+                'url': video_real_url,
+                'ext': extension,
+                'format_id': _format_id_from_url(video_real_url),
+            }]
         else:
-            # Get external player info
-            ext_player_info = self._download_webpage(
-                'http://ext.nicovideo.jp/thumb_watch/' + video_id, video_id)
-            thumb_play_key = self._search_regex(
-                r'\'thumbPlayKey\'\s*:\s*\'(.*?)\'', ext_player_info, 'thumbPlayKey')
+            formats = []
 
-            # Get flv info
-            flv_info_data = compat_urllib_parse.urlencode({
-                'k': thumb_play_key,
-                'v': video_id
-            })
-            flv_info_request = compat_urllib_request.Request(
-                'http://ext.nicovideo.jp/thumb_watch', flv_info_data,
-                {'Content-Type': 'application/x-www-form-urlencoded'})
-            flv_info_webpage = self._download_webpage(
-                flv_info_request, video_id,
-                note='Downloading flv info', errnote='Unable to download flv info')
+            dmc_info = api_data['video'].get('dmcInfo')
+            if dmc_info:  # "New" HTML5 videos
+                quality_info = dmc_info['quality']
+                for audio_quality in quality_info['audios']:
+                    for video_quality in quality_info['videos']:
+                        if not audio_quality['available'] or not video_quality['available']:
+                            continue
+                        formats.append(self._extract_format_for_quality(
+                            api_data, video_id, audio_quality, video_quality))
 
-        video_real_url = compat_urlparse.parse_qs(flv_info_webpage)['url'][0]
+                self._sort_formats(formats)
+            else:  # "Old" HTML5 videos
+                formats = [{
+                    'url': video_real_url,
+                    'ext': 'mp4',
+                    'format_id': _format_id_from_url(video_real_url),
+                }]
+
+            def get_video_info(items):
+                return dict_get(api_data['video'], items)
 
         # Start extracting information
-        title = video_info.find('.//title').text
-        extension = video_info.find('.//movie_type').text
-        video_format = extension.upper()
-        thumbnail = video_info.find('.//thumbnail_url').text
-        description = video_info.find('.//description').text
-        upload_date = unified_strdate(video_info.find('.//first_retrieve').text.split('+')[0])
-        view_count = int_or_none(video_info.find('.//view_counter').text)
-        comment_count = int_or_none(video_info.find('.//comment_num').text)
-        duration = parse_duration(video_info.find('.//length').text)
-        webpage_url = video_info.find('.//watch_url').text
-
-        if video_info.find('.//ch_id') is not None:
-            uploader_id = video_info.find('.//ch_id').text
-            uploader = video_info.find('.//ch_name').text
-        elif video_info.find('.//user_id') is not None:
-            uploader_id = video_info.find('.//user_id').text
-            uploader = video_info.find('.//user_nickname').text
-        else:
-            uploader_id = uploader = None
+        title = get_video_info('title')
+        if not title:
+            title = self._og_search_title(webpage, default=None)
+        if not title:
+            title = self._html_search_regex(
+                r'<span[^>]+class="videoHeaderTitle"[^>]*>([^<]+)</span>',
+                webpage, 'video title')
+
+        watch_api_data_string = self._html_search_regex(
+            r'<div[^>]+id="watchAPIDataContainer"[^>]+>([^<]+)</div>',
+            webpage, 'watch api data', default=None)
+        watch_api_data = self._parse_json(watch_api_data_string, video_id) if watch_api_data_string else {}
+        video_detail = watch_api_data.get('videoDetail', {})
+
+        thumbnail = (
+            get_video_info(['thumbnail_url', 'thumbnailURL']) or
+            self._html_search_meta('image', webpage, 'thumbnail', default=None) or
+            video_detail.get('thumbnail'))
+
+        description = get_video_info('description')
+
+        timestamp = (parse_iso8601(get_video_info('first_retrieve')) or
+                     unified_timestamp(get_video_info('postedDateTime')))
+        if not timestamp:
+            match = self._html_search_meta('datePublished', webpage, 'date published', default=None)
+            if match:
+                timestamp = parse_iso8601(match.replace('+', ':00+'))
+        if not timestamp and video_detail.get('postedAt'):
+            timestamp = parse_iso8601(
+                video_detail['postedAt'].replace('/', '-'),
+                delimiter=' ', timezone=datetime.timedelta(hours=9))
+
+        view_count = int_or_none(get_video_info(['view_counter', 'viewCount']))
+        if not view_count:
+            match = self._html_search_regex(
+                r'>Views: <strong[^>]*>([^<]+)</strong>',
+                webpage, 'view count', default=None)
+            if match:
+                view_count = int_or_none(match.replace(',', ''))
+        view_count = view_count or video_detail.get('viewCount')
+
+        comment_count = (int_or_none(get_video_info('comment_num')) or
+                         video_detail.get('commentCount') or
+                         try_get(api_data, lambda x: x['thread']['commentCount']))
+        if not comment_count:
+            match = self._html_search_regex(
+                r'>Comments: <strong[^>]*>([^<]+)</strong>',
+                webpage, 'comment count', default=None)
+            if match:
+                comment_count = int_or_none(match.replace(',', ''))
+
+        duration = (parse_duration(
+            get_video_info('length') or
+            self._html_search_meta(
+                'video:duration', webpage, 'video duration', default=None)) or
+            video_detail.get('length') or
+            get_video_info('duration'))
+
+        webpage_url = get_video_info('watch_url') or url
+
+        # Note: cannot use api_data.get('owner', {}) because owner may be set to "null"
+        # in the JSON, which will cause None to be returned instead of {}.
+        owner = try_get(api_data, lambda x: x.get('owner'), dict) or {}
+        uploader_id = get_video_info(['ch_id', 'user_id']) or owner.get('id')
+        uploader = get_video_info(['ch_name', 'user_nickname']) or owner.get('nickname')
 
         return {
             'id': video_id,
-            'url': video_real_url,
             'title': title,
-            'ext': extension,
-            'format': video_format,
+            'formats': formats,
             'thumbnail': thumbnail,
             'description': description,
             'uploader': uploader,
-            'upload_date': upload_date,
+            'timestamp': timestamp,
             'uploader_id': uploader_id,
             'view_count': view_count,
             'comment_count': comment_count,
             'duration': duration,
             'webpage_url': webpage_url,
         }
+
+
+class NiconicoPlaylistIE(InfoExtractor):
+    _VALID_URL = r'https?://(?:www\.)?nicovideo\.jp/mylist/(?P<id>\d+)'
+
+    _TEST = {
+        'url': 'http://www.nicovideo.jp/mylist/27411728',
+        'info_dict': {
+            'id': '27411728',
+            'title': 'AKB48のオールナイトニッポン',
+        },
+        'playlist_mincount': 225,
+    }
+
+    def _real_extract(self, url):
+        list_id = self._match_id(url)
+        webpage = self._download_webpage(url, list_id)
+
+        entries_json = self._search_regex(r'Mylist\.preload\(\d+, (\[.*\])\);',
+                                          webpage, 'entries')
+        entries = json.loads(entries_json)
+        entries = [{
+            '_type': 'url',
+            'ie_key': NiconicoIE.ie_key(),
+            'url': ('http://www.nicovideo.jp/watch/%s' %
+                    entry['item_data']['video_id']),
+        } for entry in entries]
+
+        return {
+            '_type': 'playlist',
+            'title': self._search_regex(r'\s+name: "(.*?)"', webpage, 'title'),
+            'id': list_id,
+            'entries': entries,
+        }