From 9e9bc793f3abddc4824cfcb13f569163fb0a4ba7 Mon Sep 17 00:00:00 2001
From: xavier <xavier.beynon@gmail.com>
Date: Thu, 23 Oct 2014 23:54:59 -0500
Subject: [PATCH] Finished audiomack extractor

---
 youtube_dl/extractor/audiomack.py | 82 ++++++++++++++++++++-----------
 1 file changed, 53 insertions(+), 29 deletions(-)

diff --git a/youtube_dl/extractor/audiomack.py b/youtube_dl/extractor/audiomack.py
index c5214f401..2ececa998 100644
--- a/youtube_dl/extractor/audiomack.py
+++ b/youtube_dl/extractor/audiomack.py
@@ -1,43 +1,67 @@
+# Xavier Beynon 2014
 # coding: utf-8
 from __future__ import unicode_literals
 
 from .common import InfoExtractor
+from .soundcloud import SoundcloudIE
 import datetime
 import time
-import urllib.request
-import json
 
 
 class AudiomackIE(InfoExtractor):
     _VALID_URL = r'https?://(?:www\.)?audiomack\.com/song/(?P<id>[\w/-]+)'
-    _TEST = {
-        'url': 'https://www.audiomack.com/song/crewneckkramer/story-i-tell',
-        'info_dict': {
-            'id': 'story-i-tell',
-            'ext': 'mp3',
-            'title': 'story-i-tell'
+    IE_NAME = 'audiomack'
+    _TESTS = [
+        #hosted on audiomack
+        {
+            'url': 'http://www.audiomack.com/song/roosh-williams/extraordinary',
+            'file': 'Roosh Williams - Extraordinary.mp3',
+            'info_dict':
+            {
+                'ext': 'mp3',
+                'title': 'Roosh Williams - Extraordinary'
+            }
+        },
+        #hosted on soundcloud via audiomack
+        {
+            'url': 'http://www.audiomack.com/song/xclusiveszone/take-kare',
+            'file': '172419696.mp3',
+            'info_dict':
+            {
+                'ext': 'mp3',
+                'title': 'Young Thug ft Lil Wayne - Take Kare',
+                "upload_date": "20141016",
+                "description": "New track produced by London On Da Track called “Take Kare\"\n\nhttp://instagram.com/theyoungthugworld\nhttps://www.facebook.com/ThuggerThuggerCashMoney\n",
+                "uploader": "Young Thug World"
+            }
         }
-    }
+    ]
 
     def _real_extract(self, url):
-        # TODO more code goes here, for example ...
-        #webpage = self._download_webpage(url, video_id)
-        #title = self._html_search_regex(r'<h1>(.*?)</h1>', webpage, 'title')
-	
-        assert("/song/" in url)
-        songurl = url[url.index("/song/")+5:]
-        title = songurl[songurl.rindex("/")+1:]
-        video_id = title
-        t = int(time.mktime(datetime.datetime.now().timetuple()))
-        s = "http://www.audiomack.com/api/music/url/song"+songurl+"?_="+str(t)
-        f = urllib.request.urlopen(s)
-        j = f.read(1000).decode("utf-8")
-        data = json.loads(j)
+        #id is what follows /song/ in url, usually the uploader name + title
+        id = url[url.index("/song/")+5:]
 
-        return {
-            'id': video_id,
-            'title': title,
-            'url' : data["url"],
-            'ext' : 'mp3'
-            # TODO more properties (see youtube_dl/extractor/common.py)
-        }   
+        #Call the api, which gives us a json doc with the real url inside
+        rightnow = int(time.mktime(datetime.datetime.now().timetuple()))
+        apiresponse = self._download_json("http://www.audiomack.com/api/music/url/song"+id+"?_="+str(rightnow), id)
+        if not url in apiresponse:
+            raise Exception("Unable to deduce api url of song")
+        realurl = apiresponse["url"]
+
+        #Audiomack wraps a lot of soundcloud tracks in their branded wrapper
+        # - if so, pass the work off to the soundcloud extractor
+        if SoundcloudIE.suitable(realurl):
+            sc = SoundcloudIE(downloader=self._downloader)
+            return sc._real_extract(realurl)
+        else:
+            #Pull out metadata
+            page = self._download_webpage(url, id)
+            artist = self._html_search_regex(r'<span class="artist">(.*)</span>', page, "artist")
+            songtitle = self._html_search_regex(r'<h1 class="profile-title song-title"><span class="artist">.*</span>(.*)</h1>', page, "title")
+            title = artist+" - "+songtitle
+            return {
+                'id': title,  # ignore id, which is not useful in song name
+                'title': title,
+                'url': realurl,
+                'ext': 'mp3'
+            }