Merge tag 'upstream/2017.05.18.1'

[youtubedl] / youtube_dl / extractor / bilibili.py

diff --git a/youtube_dl/extractor/bilibili.py b/youtube_dl/extractor/bilibili.py
index 75d744852edc382721cee8556067f89ccb0092df..1e3f25515d3596f7b196a16ab867b7b5e8c891ab 100644 (file)
--- a/youtube_dl/extractor/bilibili.py
+++ b/youtube_dl/extractor/bilibili.py
@@ -1,96 +1,245 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
 # coding: utf-8
 from __future__ import unicode_literals
 

+import hashlib

 import re
 
 from .common import InfoExtractor
 import re
 
 from .common import InfoExtractor

+from ..compat import (
+    compat_parse_qs,
+    compat_urlparse,
+)

 from ..utils import (
 from ..utils import (

+    ExtractorError,

     int_or_none,
     int_or_none,

-    unified_strdate,
+    float_or_none,
+    parse_iso8601,
+    smuggle_url,
+    strip_jsonp,
+    unified_timestamp,
+    unsmuggle_url,
+    urlencode_postdata,

 )
 
 
 class BiliBiliIE(InfoExtractor):
 )
 
 
 class BiliBiliIE(InfoExtractor):

-    _VALID_URL = r'http://www\.bilibili\.(?:tv|com)/video/av(?P<id>[0-9]+)/'
+    _VALID_URL = r'https?://(?:www\.|bangumi\.|)bilibili\.(?:tv|com)/(?:video/av|anime/(?P<anime_id>\d+)/play#)(?P<id>\d+)'

 
 

-    _TEST = {
+    _TESTS = [{

         'url': 'http://www.bilibili.tv/video/av1074402/',
         'url': 'http://www.bilibili.tv/video/av1074402/',

-        'md5': '2c301e4dab317596e837c3e7633e7d86',
+        'md5': '9fa226fe2b8a9a4d5a69b4c6a183417e',

         'info_dict': {
             'id': '1074402',
         'info_dict': {
             'id': '1074402',

-            'ext': 'flv',
+            'ext': 'mp4',

             'title': '【金坷垃】金泡沫',
             'title': '【金坷垃】金泡沫',

-            'duration': 308,
+            'description': 'md5:ce18c2a2d2193f0df2917d270f2e5923',
+            'duration': 308.315,
+            'timestamp': 1398012660,

             'upload_date': '20140420',
             'upload_date': '20140420',

-            'thumbnail': 're:^https?://.+\.jpg',
+            'thumbnail': r're:^https?://.+\.jpg',
+            'uploader': '菊子桑',
+            'uploader_id': '156160',
+        },
+    }, {
+        # Tested in BiliBiliBangumiIE
+        'url': 'http://bangumi.bilibili.com/anime/1869/play#40062',
+        'only_matching': True,
+    }, {
+        'url': 'http://bangumi.bilibili.com/anime/5802/play#100643',
+        'md5': '3f721ad1e75030cc06faf73587cfec57',
+        'info_dict': {
+            'id': '100643',
+            'ext': 'mp4',
+            'title': 'CHAOS;CHILD',
+            'description': '如果你是神明，并且能够让妄想成为现实。那你会进行怎么样的妄想？是淫靡的世界？独裁社会？毁灭性的制裁？还是……2015年，涩谷。从6年前发生的大灾害“涩谷地震”之后复兴了的这个街区里新设立的私立高中...',

         },
         },

-    }
+        'skip': 'Geo-restricted to China',
+    }]
+
+    _APP_KEY = '84956560bc028eb7'
+    _BILIBILI_KEY = '94aba54af9065f71de72f5508f1cd42e'
+
+    def _report_error(self, result):
+        if 'message' in result:
+            raise ExtractorError('%s said: %s' % (self.IE_NAME, result['message']), expected=True)
+        elif 'code' in result:
+            raise ExtractorError('%s returns error %d' % (self.IE_NAME, result['code']), expected=True)
+        else:
+            raise ExtractorError('Can\'t extract Bangumi episode ID')

 
     def _real_extract(self, url):
 
     def _real_extract(self, url):

-        video_id = self._match_id(url)
-        webpage = self._download_webpage(url, video_id)
+        url, smuggled_data = unsmuggle_url(url, {})

 
 

-        video_code = self._search_regex(
-            r'(?s)<div itemprop="video".*?>(.*?)</div>', webpage, 'video code')
+        mobj = re.match(self._VALID_URL, url)
+        video_id = mobj.group('id')
+        anime_id = mobj.group('anime_id')
+        webpage = self._download_webpage(url, video_id)

 
 

-        title = self._html_search_meta(
-            'media:title', video_code, 'title', fatal=True)
-        duration_str = self._html_search_meta(
-            'duration', video_code, 'duration')
-        if duration_str is None:
-            duration = None
+        if 'anime/' not in url:
+            cid = compat_parse_qs(self._search_regex(
+                [r'EmbedPlayer\([^)]+,\s*"([^"]+)"\)',
+                 r'<iframe[^>]+src="https://secure\.bilibili\.com/secure,([^"]+)"'],
+                webpage, 'player parameters'))['cid'][0]

         else:
         else:

-            duration_mobj = re.match(
-                r'^T(?:(?P<hours>[0-9]+)H)?(?P<minutes>[0-9]+)M(?P<seconds>[0-9]+)S$',
-                duration_str)
-            duration = (
-                int_or_none(duration_mobj.group('hours'), default=0) * 3600 +
-                int(duration_mobj.group('minutes')) * 60 +
-                int(duration_mobj.group('seconds')))
-        upload_date = unified_strdate(self._html_search_meta(
-            'uploadDate', video_code, fatal=False))
-        thumbnail = self._html_search_meta(
-            'thumbnailUrl', video_code, 'thumbnail', fatal=False)
-
-        cid = self._search_regex(r'cid=(\d+)', webpage, 'cid')
-
-        lq_doc = self._download_xml(
-            'http://interface.bilibili.com/v_cdn_play?appkey=1&cid=%s' % cid,
-            video_id,
-            note='Downloading LQ video info'
-        )
-        lq_durl = lq_doc.find('./durl')
-        formats = [{
-            'format_id': 'lq',
-            'quality': 1,
-            'url': lq_durl.find('./url').text,
-            'filesize': int_or_none(
-                lq_durl.find('./size'), get_attr='text'),
-        }]
-
-        hq_doc = self._download_xml(
-            'http://interface.bilibili.com/playurl?appkey=1&cid=%s' % cid,
-            video_id,
-            note='Downloading HQ video info',
-            fatal=False,
-        )
-        if hq_doc is not False:
-            hq_durl = hq_doc.find('./durl')
-            formats.append({
-                'format_id': 'hq',
-                'quality': 2,
-                'ext': 'flv',
-                'url': hq_durl.find('./url').text,
-                'filesize': int_or_none(
-                    hq_durl.find('./size'), get_attr='text'),
+            if 'no_bangumi_tip' not in smuggled_data:
+                self.to_screen('Downloading episode %s. To download all videos in anime %s, re-run youtube-dl with %s' % (
+                    video_id, anime_id, compat_urlparse.urljoin(url, '//bangumi.bilibili.com/anime/%s' % anime_id)))
+            headers = {
+                'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8',
+            }
+            headers.update(self.geo_verification_headers())
+
+            js = self._download_json(
+                'http://bangumi.bilibili.com/web_api/get_source', video_id,
+                data=urlencode_postdata({'episode_id': video_id}),
+                headers=headers)
+            if 'result' not in js:
+                self._report_error(js)
+            cid = js['result']['cid']
+
+        payload = 'appkey=%s&cid=%s&otype=json&quality=2&type=mp4' % (self._APP_KEY, cid)
+        sign = hashlib.md5((payload + self._BILIBILI_KEY).encode('utf-8')).hexdigest()
+
+        video_info = self._download_json(
+            'http://interface.bilibili.com/playurl?%s&sign=%s' % (payload, sign),
+            video_id, note='Downloading video info page',
+            headers=self.geo_verification_headers())
+
+        if 'durl' not in video_info:
+            self._report_error(video_info)
+
+        entries = []
+
+        for idx, durl in enumerate(video_info['durl']):
+            formats = [{
+                'url': durl['url'],
+                'filesize': int_or_none(durl['size']),
+            }]
+            for backup_url in durl.get('backup_url', []):
+                formats.append({
+                    'url': backup_url,
+                    # backup URLs have lower priorities
+                    'preference': -2 if 'hd.mp4' in backup_url else -3,
+                })
+
+            for a_format in formats:
+                a_format.setdefault('http_headers', {}).update({
+                    'Referer': url,
+                })
+
+            self._sort_formats(formats)
+
+            entries.append({
+                'id': '%s_part%s' % (video_id, idx),
+                'duration': float_or_none(durl.get('length'), 1000),
+                'formats': formats,

             })
 
             })
 

-        self._sort_formats(formats)
-        return {
+        title = self._html_search_regex('<h1[^>]+title="([^"]+)">', webpage, 'title')
+        description = self._html_search_meta('description', webpage)
+        timestamp = unified_timestamp(self._html_search_regex(
+            r'<time[^>]+datetime="([^"]+)"', webpage, 'upload time', default=None))
+        thumbnail = self._html_search_meta(['og:image', 'thumbnailUrl'], webpage)
+
+        # TODO 'view_count' requires deobfuscating Javascript
+        info = {

             'id': video_id,
             'title': title,
             'id': video_id,
             'title': title,

-            'formats': formats,
-            'duration': duration,
-            'upload_date': upload_date,
+            'description': description,
+            'timestamp': timestamp,

             'thumbnail': thumbnail,
             'thumbnail': thumbnail,

+            'duration': float_or_none(video_info.get('timelength'), scale=1000),

         }
         }

+
+        uploader_mobj = re.search(
+            r'<a[^>]+href="(?:https?:)?//space\.bilibili\.com/(?P<id>\d+)"[^>]+title="(?P<name>[^"]+)"',
+            webpage)
+        if uploader_mobj:
+            info.update({
+                'uploader': uploader_mobj.group('name'),
+                'uploader_id': uploader_mobj.group('id'),
+            })
+
+        for entry in entries:
+            entry.update(info)
+
+        if len(entries) == 1:
+            return entries[0]
+        else:
+            for idx, entry in enumerate(entries):
+                entry['id'] = '%s_part%d' % (video_id, (idx + 1))
+
+            return {
+                '_type': 'multi_video',
+                'id': video_id,
+                'title': title,
+                'description': description,
+                'entries': entries,
+            }
+
+
+class BiliBiliBangumiIE(InfoExtractor):
+    _VALID_URL = r'https?://bangumi\.bilibili\.com/anime/(?P<id>\d+)'
+
+    IE_NAME = 'bangumi.bilibili.com'
+    IE_DESC = 'BiliBili番剧'
+
+    _TESTS = [{
+        'url': 'http://bangumi.bilibili.com/anime/1869',
+        'info_dict': {
+            'id': '1869',
+            'title': '混沌武士',
+            'description': 'md5:6a9622b911565794c11f25f81d6a97d2',
+        },
+        'playlist_count': 26,
+    }, {
+        'url': 'http://bangumi.bilibili.com/anime/1869',
+        'info_dict': {
+            'id': '1869',
+            'title': '混沌武士',
+            'description': 'md5:6a9622b911565794c11f25f81d6a97d2',
+        },
+        'playlist': [{
+            'md5': '91da8621454dd58316851c27c68b0c13',
+            'info_dict': {
+                'id': '40062',
+                'ext': 'mp4',
+                'title': '混沌武士',
+                'description': '故事发生在日本的江户时代。风是一个小酒馆的打工女。一日，酒馆里来了一群恶霸，虽然他们的举动令风十分不满，但是毕竟风只是一届女流，无法对他们采取什么行动，只能在心里嘟哝。这时，酒家里又进来了个“不良份子...',
+                'timestamp': 1414538739,
+                'upload_date': '20141028',
+                'episode': '疾风怒涛 Tempestuous Temperaments',
+                'episode_number': 1,
+            },
+        }],
+        'params': {
+            'playlist_items': '1',
+        },
+    }]
+
+    @classmethod
+    def suitable(cls, url):
+        return False if BiliBiliIE.suitable(url) else super(BiliBiliBangumiIE, cls).suitable(url)
+
+    def _real_extract(self, url):
+        bangumi_id = self._match_id(url)
+
+        # Sometimes this API returns a JSONP response
+        season_info = self._download_json(
+            'http://bangumi.bilibili.com/jsonp/seasoninfo/%s.ver' % bangumi_id,
+            bangumi_id, transform_source=strip_jsonp)['result']
+
+        entries = [{
+            '_type': 'url_transparent',
+            'url': smuggle_url(episode['webplay_url'], {'no_bangumi_tip': 1}),
+            'ie_key': BiliBiliIE.ie_key(),
+            'timestamp': parse_iso8601(episode.get('update_time'), delimiter=' '),
+            'episode': episode.get('index_title'),
+            'episode_number': int_or_none(episode.get('index')),
+        } for episode in season_info['episodes']]
+
+        entries = sorted(entries, key=lambda entry: entry.get('episode_number'))
+
+        return self.playlist_result(
+            entries, bangumi_id,
+            season_info.get('bangumi_title'), season_info.get('evaluate'))