From c653326a1425f4c271f387fde7a706bf4b52a7a3 Mon Sep 17 00:00:00 2001 From: =?utf8?q?Sergey=20M=E2=80=A4?= Date: Thu, 20 Jul 2017 22:49:52 +0700 Subject: [PATCH] [funnyordie] Extract more metadata (closes #13677) --- youtube_dl/extractor/funnyordie.py | 64 ++++++++++++++++++++++++++---- 1 file changed, 56 insertions(+), 8 deletions(-) diff --git a/youtube_dl/extractor/funnyordie.py b/youtube_dl/extractor/funnyordie.py index 49409369c..f85e7de14 100644 --- a/youtube_dl/extractor/funnyordie.py +++ b/youtube_dl/extractor/funnyordie.py @@ -1,10 +1,14 @@ from __future__ import unicode_literals -import json import re from .common import InfoExtractor -from ..utils import ExtractorError +from ..utils import ( + ExtractorError, + float_or_none, + int_or_none, + unified_timestamp, +) class FunnyOrDieIE(InfoExtractor): @@ -18,6 +22,10 @@ class FunnyOrDieIE(InfoExtractor): 'title': 'Heart-Shaped Box: Literal Video Version', 'description': 'md5:ea09a01bc9a1c46d9ab696c01747c338', 'thumbnail': r're:^http:.*\.jpg$', + 'uploader': 'DASjr', + 'timestamp': 1317904928, + 'upload_date': '20111006', + 'duration': 318.3, }, }, { 'url': 'http://www.funnyordie.com/embed/e402820827', @@ -27,6 +35,8 @@ class FunnyOrDieIE(InfoExtractor): 'title': 'Please Use This Song (Jon Lajoie)', 'description': 'Please use this to sell something. www.jonlajoie.com', 'thumbnail': r're:^http:.*\.jpg$', + 'timestamp': 1398988800, + 'upload_date': '20140502', }, 'params': { 'skip_download': True, @@ -100,15 +110,53 @@ class FunnyOrDieIE(InfoExtractor): 'url': 'http://www.funnyordie.com%s' % src, }] - post_json = self._search_regex( - r'fb_post\s*=\s*(\{.*?\});', webpage, 'post details') - post = json.loads(post_json) + timestamp = unified_timestamp(self._html_search_meta( + 'uploadDate', webpage, 'timestamp', default=None)) + + uploader = self._html_search_regex( + r']+\bclass=["\']channel-preview-name[^>]+>(.+?)