From aa9dc24f5a84265b66a25b9fa70c06e4fae299c6 Mon Sep 17 00:00:00 2001 From: Yen Chi Hsuan Date: Sun, 24 Apr 2016 23:52:17 +0800 Subject: [PATCH] [douyutv] Improve extraction and update tests The JSON API sometimes return HTML pages with errors --- youtube_dl/extractor/douyutv.py | 33 ++++++++++++++++++++++++--------- 1 file changed, 24 insertions(+), 9 deletions(-) diff --git a/youtube_dl/extractor/douyutv.py b/youtube_dl/extractor/douyutv.py index 3915cb1829..ce69627558 100644 --- a/youtube_dl/extractor/douyutv.py +++ b/youtube_dl/extractor/douyutv.py @@ -18,7 +18,7 @@ class DouyuTVIE(InfoExtractor): 'display_id': 'iseven', 'ext': 'flv', 'title': 're:^清晨醒脑!T-ara根本停不下来! [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$', - 'description': 'md5:f34981259a03e980a3c6404190a3ed61', + 'description': 're:.*m7show@163\.com.*', 'thumbnail': 're:^https?://.*\.jpg$', 'uploader': '7师傅', 'uploader_id': '431925', @@ -43,7 +43,7 @@ class DouyuTVIE(InfoExtractor): 'params': { 'skip_download': True, }, - 'skip': 'Romm not found', + 'skip': 'Room not found', }, { 'url': 'http://www.douyutv.com/17732', 'info_dict': { @@ -51,7 +51,7 @@ class DouyuTVIE(InfoExtractor): 'display_id': '17732', 'ext': 'flv', 'title': 're:^清晨醒脑!T-ara根本停不下来! [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}$', - 'description': 'md5:f34981259a03e980a3c6404190a3ed61', + 'description': 're:.*m7show@163\.com.*', 'thumbnail': 're:^https?://.*\.jpg$', 'uploader': '7师傅', 'uploader_id': '431925', @@ -75,13 +75,28 @@ class DouyuTVIE(InfoExtractor): room_id = self._html_search_regex( r'"room_id"\s*:\s*(\d+),', page, 'room id') - prefix = 'room/%s?aid=android&client_sys=android&time=%d' % ( - room_id, int(time.time())) + config = 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() - auth = hashlib.md5((prefix + '1231').encode('ascii')).hexdigest() - config = self._download_json( - 'http://www.douyutv.com/api/v1/%s&auth=%s' % (prefix, auth), - video_id) + config_page = self._download_webpage( + 'http://www.douyutv.com/api/v1/%s&auth=%s' % (prefix, auth), + video_id) + try: + config = self._parse_json(config_page, 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. ' + 'Waiting for %(timeout)s seconds before retrying') + continue + else: + break + if config is None: + raise ExtractorError('Unable to fetch API result') data = config['data']