From 2a46a27e6c704f7b99242243747fa7d8ac68cefc Mon Sep 17 00:00:00 2001 From: Eric Wong Date: Mon, 15 Jun 2015 01:07:48 -0700 Subject: [PATCH] [thisamericanlife] Add a new extractor --- youtube_dl/extractor/__init__.py | 1 + youtube_dl/extractor/thisamericanlife.py | 32 ++++++++++++++++++++++++ 2 files changed, 33 insertions(+) create mode 100644 youtube_dl/extractor/thisamericanlife.py diff --git a/youtube_dl/extractor/__init__.py b/youtube_dl/extractor/__init__.py index d44339200b..aba62db532 100644 --- a/youtube_dl/extractor/__init__.py +++ b/youtube_dl/extractor/__init__.py @@ -569,6 +569,7 @@ from .tf1 import TF1IE from .theonion import TheOnionIE from .theplatform import ThePlatformIE from .thesixtyone import TheSixtyOneIE +from .thisamericanlife import ThisAmericanLifeIE from .thisav import ThisAVIE from .tinypic import TinyPicIE from .tlc import TlcIE, TlcDeIE diff --git a/youtube_dl/extractor/thisamericanlife.py b/youtube_dl/extractor/thisamericanlife.py new file mode 100644 index 0000000000..6118afa1af --- /dev/null +++ b/youtube_dl/extractor/thisamericanlife.py @@ -0,0 +1,32 @@ +# coding: utf-8 +from __future__ import unicode_literals + +from .common import InfoExtractor + + +class ThisAmericanLifeIE(InfoExtractor): + _VALID_URL = r'https?://(?:www\.)?thisamericanlife\.org/radio-archives/episode/(?P\d+)' + _TEST = { + 'url': 'http://www.thisamericanlife.org/radio-archives/episode/487/harper-high-school-part-one', + 'md5': '5cda28076c9f9d1fd0b0f5cff5959948', + 'info_dict': { + 'id': '487', + 'title': '487: Harper High School, Part One', + 'url' : 'http://stream.thisamericanlife.org/487/stream/487_64k.m3u8', + 'ext': 'aac', + } + } + + def _real_extract(self, url): + video_id = self._match_id(url) + webpage = self._download_webpage(url, video_id) + + title = self._html_search_regex(r']*>(.*?)', webpage, 'title') + media_url = 'http://stream.thisamericanlife.org/' + video_id + '/stream/' + video_id + '_64k.m3u8' + + return { + 'id': video_id, + 'title': title, + 'url': media_url, + 'ext': 'aac', + }