]> git.bitcoin.ninja Git - youtube-dl/commitdiff
[douyutv] Use new api
authorsteven7851 <steven7851@msn.com>
Thu, 25 Aug 2016 23:32:54 +0000 (07:32 +0800)
committerGitHub <noreply@github.com>
Thu, 25 Aug 2016 23:32:54 +0000 (07:32 +0800)
use lapi for flv info, and html5 api for room info
#10153 #10318

youtube_dl/extractor/douyutv.py

index ce6962755831a8c6f853271ad49112fee4fcbc63..33efc993eeb89b76af7779ca67de35ad92620150 100644 (file)
@@ -3,9 +3,10 @@ from __future__ import unicode_literals
 
 import hashlib
 import time
+import uuid
 from .common import InfoExtractor
 from ..utils import (ExtractorError, unescapeHTML)
-from ..compat import (compat_str, compat_basestring)
+from ..compat import (compat_str, compat_basestring, compat_urllib_parse_urlencode)
 
 
 class DouyuTVIE(InfoExtractor):
@@ -21,7 +22,6 @@ class DouyuTVIE(InfoExtractor):
             'description': 're:.*m7show@163\.com.*',
             'thumbnail': 're:^https?://.*\.jpg$',
             'uploader': '7师傅',
-            'uploader_id': '431925',
             'is_live': True,
         },
         'params': {
@@ -37,7 +37,6 @@ class DouyuTVIE(InfoExtractor):
             'description': 'md5:746a2f7a253966a06755a912f0acc0d2',
             'thumbnail': 're:^https?://.*\.jpg$',
             'uploader': 'douyu小漠',
-            'uploader_id': '3769985',
             'is_live': True,
         },
         'params': {
@@ -54,7 +53,6 @@ class DouyuTVIE(InfoExtractor):
             'description': 're:.*m7show@163\.com.*',
             'thumbnail': 're:^https?://.*\.jpg$',
             'uploader': '7师傅',
-            'uploader_id': '431925',
             'is_live': True,
         },
         'params': {
@@ -75,19 +73,39 @@ class DouyuTVIE(InfoExtractor):
             room_id = self._html_search_regex(
                 r'"room_id"\s*:\s*(\d+),', page, 'room id')
 
-        config = None
+        room_url = 'http://m.douyu.com/html5/live?roomId=%s' % room_id
+        room_content = self._download_webpage(room_url, video_id)
+        room_json = self._parse_json(room_content, video_id, fatal=False)
+
+        room = room_json['data']
+
+        show_status = room.get('show_status')
+        # 1 = live, 2 = offline
+        if show_status == '2':
+            raise ExtractorError(
+                'Live stream is offline', expected=True)
+
+        flv_json = None
         # Douyu API sometimes returns error "Unable to load the requested class: eticket_redis_cache"
         # Retry with different parameters - same parameters cause same errors
         for i in range(5):
-            prefix = 'room/%s?aid=android&client_sys=android&time=%d' % (
-                room_id, int(time.time()))
-            auth = hashlib.md5((prefix + '1231').encode('ascii')).hexdigest()
+            tt = int(time.time() / 60)
+            did = uuid.uuid4().hex.upper()
+
+            # Decompile core.swf in webpage by ffdec "Search SWFs in memory"
+            # core.swf is encrypted originally, but ffdec can dump memory to get the decrypted one
+            # If API changes in the future, just use this way to update
+            sign_content = '{room_id}{did}A12Svb&%1UUmf@hC{tt}'.format(room_id = room_id, did = did, tt = tt)
+            sign = hashlib.md5((sign_content).encode('utf-8')).hexdigest()
 
-            config_page = self._download_webpage(
-                'http://www.douyutv.com/api/v1/%s&auth=%s' % (prefix, auth),
-                video_id)
+            payload = {'cdn': 'ws', 'rate': '0', 'tt': tt, 'did': did, 'sign': sign}
+            flv_data = compat_urllib_parse_urlencode(payload)
+
+            flv_request_url = 'http://www.douyu.com/lapi/live/getPlay/%s' % room_id
+            flv_content = self._download_webpage(flv_request_url, video_id, data=flv_data,
+                headers={'Content-Type': 'application/x-www-form-urlencoded'})
             try:
-                config = self._parse_json(config_page, video_id, fatal=False)
+                flv_json = self._parse_json(flv_content, video_id, fatal=False)
             except ExtractorError:
                 # Wait some time before retrying to get a different time() value
                 self._sleep(1, video_id, msg_template='%(video_id)s: Error occurs. '
@@ -95,54 +113,35 @@ class DouyuTVIE(InfoExtractor):
                 continue
             else:
                 break
-        if config is None:
+        if flv_json is None:
             raise ExtractorError('Unable to fetch API result')
 
-        data = config['data']
+        flv = flv_json['data']
 
-        error_code = config.get('error', 0)
+        error_code = flv_json.get('error', 0)
         if error_code is not 0:
             error_desc = 'Server reported error %i' % error_code
-            if isinstance(data, (compat_str, compat_basestring)):
-                error_desc += ': ' + data
+            if isinstance(flv, (compat_str, compat_basestring)):
+                error_desc += ': ' + flv
             raise ExtractorError(error_desc, expected=True)
 
-        show_status = data.get('show_status')
-        # 1 = live, 2 = offline
-        if show_status == '2':
-            raise ExtractorError(
-                'Live stream is offline', expected=True)
-
-        base_url = data['rtmp_url']
-        live_path = data['rtmp_live']
-
-        title = self._live_title(unescapeHTML(data['room_name']))
-        description = data.get('show_details')
-        thumbnail = data.get('room_src')
-
-        uploader = data.get('nickname')
-        uploader_id = data.get('owner_uid')
+        base_url = flv['rtmp_url']
+        live_path = flv['rtmp_live']
 
-        multi_formats = data.get('rtmp_multi_bitrate')
-        if not isinstance(multi_formats, dict):
-            multi_formats = {}
-        multi_formats['live'] = live_path
+        video_url = '%s/%s' % (base_url, live_path)
 
-        formats = [{
-            'url': '%s/%s' % (base_url, format_path),
-            'format_id': format_id,
-            'preference': 1 if format_id == 'live' else 0,
-        } for format_id, format_path in multi_formats.items()]
-        self._sort_formats(formats)
+        title = self._live_title(unescapeHTML(room['room_name']))
+        description = room.get('notice')
+        thumbnail = room.get('room_src')
+        uploader = room.get('nickname')
 
         return {
             'id': room_id,
             'display_id': video_id,
+            'url': video_url,
             'title': title,
             'description': description,
             'thumbnail': thumbnail,
             'uploader': uploader,
-            'uploader_id': uploader_id,
-            'formats': formats,
             'is_live': True,
         }