[weibo] Misc improvements
authorYen Chi Hsuan <yan12125@gmail.com>
Tue, 9 Jan 2018 10:12:55 +0000 (18:12 +0800)
committerYen Chi Hsuan <yan12125@gmail.com>
Tue, 9 Jan 2018 10:12:55 +0000 (18:12 +0800)
youtube_dl/extractor/weibo.py

index cbe0c3228babefda67619dc48605fefa4ce864dc..3cb4d71a6c2381fe1169b11785a2cc950d2b8437 100644 (file)
@@ -8,7 +8,8 @@ import random
 import re
 
 from ..compat import (
-    compat_urlparse,
+    compat_parse_qs,
+    compat_str,
 )
 from ..utils import (
     js_to_json,
@@ -31,70 +32,71 @@ class WeiboIE(InfoExtractor):
     def _real_extract(self, url):
         video_id = self._match_id(url)
         # to get Referer url for genvisitor
-        webpage, urlh = self._download_webpage_handle(url, video_id, note="first visit the page")
+        webpage, urlh = self._download_webpage_handle(url, video_id)
 
         visitor_url = urlh.geturl()
-        headers = {
-            'Referer': visitor_url
-        }
 
-        fp = {
-            "os": "2",
-            "browser": "Gecko57,0,0,0",
-            "fonts": "undefined",
-            "screenInfo": "1440*900*24",
-            "plugins": ""
-        }
-        data = urlencode_postdata({
-            "cb": "gen_callback",
-            "fp": json.dumps(fp),
-        })
-
-        genvisitor_url = 'https://passport.weibo.com/visitor/genvisitor'
-        webpage = self._download_webpage(genvisitor_url, video_id, data=data, headers=headers, note="gen visitor")
-
-        p = strip_jsonp(webpage)
-        i1 = p.find('{')
-        i2 = p.rfind('}')
-        j = p[i1:i2 + 1]  # get JSON object
-        d = json.loads(j)
-        tid = d["data"]["tid"]
-        cnfd = "%03d" % d["data"]["confidence"]
-
-        query = {
-            'a': 'incarnate',
-            't': tid,
-            'w': 2,
-            'c': cnfd,
-            'cb': 'cross_domain',
-            'from': 'weibo',
-            '_rand': random.random()
-        }
-        gencallback_url = "https://passport.weibo.com/visitor/visitor"
-        self._download_webpage(gencallback_url, video_id, note="gen callback", query=query)
+        if 'passport.weibo.com' in visitor_url:
+            # first visit
+            visitor_data = self._download_json(
+                'https://passport.weibo.com/visitor/genvisitor', video_id,
+                note='Generating first-visit data',
+                transform_source=strip_jsonp,
+                headers={'Referer': visitor_url},
+                data=urlencode_postdata({
+                    'cb': 'gen_callback',
+                    'fp': json.dumps({
+                        'os': '2',
+                        'browser': 'Gecko57,0,0,0',
+                        'fonts': 'undefined',
+                        'screenInfo': '1440*900*24',
+                        'plugins': '',
+                    }),
+                }))
+
+            tid = visitor_data['data']['tid']
+            cnfd = '%03d' % visitor_data['data']['confidence']
+
+            self._download_webpage(
+                'https://passport.weibo.com/visitor/visitor', video_id,
+                note='Running first-visit callback',
+                query={
+                    'a': 'incarnate',
+                    't': tid,
+                    'w': 2,
+                    'c': cnfd,
+                    'cb': 'cross_domain',
+                    'from': 'weibo',
+                    '_rand': random.random(),
+                })
+
+            webpage = self._download_webpage(
+                url, video_id, note='Revisiting webpage')
+
+        title = self._html_search_regex(
+            r'<title>(.+?)</title>', webpage, 'title')
+
+        video_formats = compat_parse_qs(self._search_regex(
+            r'video-sources=\\\"(.+?)\"', webpage, 'video_sources'))
 
-        webpage = self._download_webpage(url, video_id, note="retry to visit the page")
+        formats = []
+        supported_resolutions = (480, 720)
+        for res in supported_resolutions:
+            vid_urls = video_formats.get(compat_str(res))
+            if not vid_urls or not isinstance(vid_urls, list):
+                continue
 
-        title = self._html_search_regex(r'<title>(.+?)</title>', webpage, 'title')
+            vid_url = vid_urls[0]
+            formats.append({
+                'url': vid_url,
+                'height': res,
+            })
 
-        video_sources_text = self._search_regex(r'video-sources=\\\"(.+?)\"', webpage, 'video_sources')
+        self._sort_formats(formats)
 
-        video_formats = compat_urlparse.parse_qs(video_sources_text)
+        uploader = self._og_search_property(
+            'nick-name', webpage, 'uploader', default=None)
 
-        formats = []
-        supported_resolutions = ('720', '480')
-        for res in supported_resolutions:
-            f = video_formats.get(res)
-            if isinstance(f, list):
-                if len(f) > 0:
-                    vid_url = f[0]
-                    formats.append({
-                        'url': vid_url,
-                        'format': 'mp4',
-                        'height': int(res),
-                    })
-        self._sort_formats(formats)
-        uploader = self._og_search_property('nick-name', webpage, 'uploader', default=None)
         return {
             'id': video_id,
             'title': title,
@@ -118,12 +120,17 @@ class WeiboMobileIE(InfoExtractor):
     def _real_extract(self, url):
         video_id = self._match_id(url)
         # to get Referer url for genvisitor
-        webpage = self._download_webpage(url, video_id, note="visit the page")
-        js_code = self._search_regex(r'var\s+\$render_data\s*=\s*\[({.*})\]\[0\] \|\| {};', webpage, 'js_code', flags=re.DOTALL)
-        weibo_info = self._parse_json(js_code, video_id, transform_source=js_to_json)
-        page_info = weibo_info.get('status').get('page_info')
-        title = weibo_info.get('status').get('status_title')
-        uploader = weibo_info.get('status').get('user').get('screen_name')
+        webpage = self._download_webpage(url, video_id, note='visit the page')
+
+        weibo_info = self._parse_json(self._search_regex(
+            r'var\s+\$render_data\s*=\s*\[({.*})\]\[0\]\s*\|\|\s*{};',
+            webpage, 'js_code', flags=re.DOTALL),
+            video_id, transform_source=js_to_json)
+
+        status_data = weibo_info.get('status', {})
+        page_info = status_data.get('page_info')
+        title = status_data['status_title']
+        uploader = status_data.get('user', {}).get('screen_name')
 
         return {
             'id': video_id,