Push api updates to simplify audiomack, add support for albums
authorYour Name <you@example.com>
Fri, 2 Jan 2015 08:20:04 +0000 (02:20 -0600)
committerYour Name <you@example.com>
Fri, 2 Jan 2015 08:20:04 +0000 (02:20 -0600)
youtube_dl/extractor/__init__.py
youtube_dl/extractor/audiomack.py

index 6c5827f88dfe313c64513ba0acf7d70874620113..e4c51f238f210704e5aaf220349cc2fe6d145f36 100644 (file)
@@ -26,7 +26,7 @@ from .arte import (
     ArteTVEmbedIE,
 )
 from .atresplayer import AtresPlayerIE
-from .audiomack import AudiomackIE
+from .audiomack import AudiomackIE, AudiomackAlbumIE
 from .auengine import AUEngineIE
 from .azubu import AzubuIE
 from .bambuser import BambuserIE, BambuserChannelIE
index 622b209899ec3dac2432ed3c7a1dadcf14537a5e..b6f3039c72b0eb2747862703a2e9bc2b34efddc1 100644 (file)
@@ -9,20 +9,21 @@ import time
 
 
 class AudiomackIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?audiomack\.com/song/(?P<id>[\w/-]+)'
+    _VALID_URL = r'https?://(?:www\.)?audiomack\.com/(song)/(?P<id>[\w/-]+)'
     IE_NAME = 'audiomack'
     _TESTS = [
-        # hosted on audiomack
+        # audiomack
         {
             'url': 'http://www.audiomack.com/song/roosh-williams/extraordinary',
             'info_dict':
             {
-                'id': 'roosh-williams/extraordinary',
-                'ext': 'mp3',
-                'title': 'Roosh Williams - Extraordinary'
+                'id': '310086',
+                "ext": "mp3",
+                "artist": "Roosh Williams",
+                'title': 'Extraordinary'
             }
         },
-        # hosted on soundcloud via audiomack
+        # audiomack through soundcloud
         {
             'add_ie': ['Soundcloud'],
             'url': 'http://www.audiomack.com/song/xclusiveszone/take-kare',
@@ -37,33 +38,88 @@ class AudiomackIE(InfoExtractor):
         },
     ]
 
+    @staticmethod
+    def create_song_dictionary(api_response, album_url_tag, track_no=0):
+        # All keys are the same in audiomack api and InfoExtractor format
+        entry = {key: api_response[key] for key in ["title", "artist", "id", "url"] if key in api_response}
+        # Fudge values in the face of missing metadata
+        if "id" not in entry:
+            entry["id"] = track_no
+        if "title" not in entry:
+            entry["title"] = album_url_tag
+        return entry
+
     def _real_extract(self, url):
-        video_id = self._match_id(url)
+        # URLs end with [uploader name]/[uploader title]
+        # this title is whatever the user types in, and is rarely
+        # the proper song title.  Real metadata is in the api response
+        album_url_tag = self._match_id(url)
 
+        # Request the extended version of the api for extra fields like artist and title
         api_response = self._download_json(
-            "http://www.audiomack.com/api/music/url/song/%s?_=%d" % (
-                video_id, time.time()),
-            video_id)
+            "http://www.audiomack.com/api/music/url/song/%s?extended=1&_=%d" % (
+                album_url_tag, time.time()),
+            album_url_tag)
 
-        if "url" not in api_response:
-            raise ExtractorError("Unable to deduce api url of song")
-        realurl = api_response["url"]
+        # API is inconsistent with errors
+        if "url" not in api_response or not api_response["url"] or "error" in api_response:
+            raise ExtractorError("Invalid url %s", 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):
-            return {'_type': 'url', 'url': realurl, 'ie_key': 'Soundcloud'}
+        # if so, pass the work off to the soundcloud extractor
+        if SoundcloudIE.suitable(api_response["url"]):
+            return {'_type': 'url', 'url': api_response["url"], 'ie_key': 'Soundcloud'}
 
-        webpage = self._download_webpage(url, video_id)
-        artist = self._html_search_regex(
-            r'<span class="artist">(.*?)</span>', webpage, "artist")
-        songtitle = self._html_search_regex(
-            r'<h1 class="profile-title song-title"><span class="artist">.*?</span>(.*?)</h1>',
-            webpage, "title")
-        title = artist + " - " + songtitle
+        return self.create_song_dictionary(api_response, album_url_tag)
 
-        return {
-            'id': video_id,
-            'title': title,
-            'url': realurl,
+
+class AudiomackAlbumIE(InfoExtractor):
+    _VALID_URL = r'https?://(?:www\.)?audiomack\.com/album/(?P<id>[\w/-]+)'
+    IE_NAME = 'audiomack:album'
+    _TESTS = [
+        # Standard album playlist
+        {
+            'url': 'http://www.audiomack.com/album/flytunezcom/tha-tour-part-2-mixtape',
+            "playlist_count": 15,
+            'info_dict':
+            {
+                'id': "812251",
+                'title': "Tha Tour: Part 2 (Official Mixtape)"
+            }
+        },
+        # Album playlist ripped from fakeshoredrive with no metadata
+        {
+            "url": "http://www.audiomack.com/album/fakeshoredrive/ppp-pistol-p-project",
+            "playlist_count": 10
         }
+    ]
+
+    def _real_extract(self, url):
+        # URLs end with [uploader name]/[uploader title]
+        # this title is whatever the user types in, and is rarely
+        # the proper song title.  Real metadata is in the api response
+        album_url_tag = self._match_id(url)
+        result = {"_type": "playlist", "entries": []}
+        # There is no one endpoint for album metadata - instead it is included/repeated in each song's metadata
+        # Therefore we don't know how many songs the album has and must infi-loop until failure
+        track_no = 0
+        while True:
+            # Get song's metadata
+            api_response = self._download_json("http://www.audiomack.com/api/music/url/album/%s/%d?extended=1&_=%d"
+                                               % (album_url_tag, track_no, time.time()), album_url_tag)
+
+            # Total failure, only occurs when url is totally wrong
+            # Won't happen in middle of valid playlist (next case)
+            if "url" not in api_response or "error" in api_response:
+                raise ExtractorError("Invalid url for track %d of album url %s" % (track_no, url))
+            # URL is good but song id doesn't exist - usually means end of playlist
+            elif not api_response["url"]:
+                break
+            else:
+                # Pull out the album metadata and add to result (if it exists)
+                for resultkey, apikey in [("id", "album_id"), ("title", "album_title")]:
+                    if apikey in api_response and resultkey not in result:
+                        result[resultkey] = api_response[apikey]
+                result["entries"].append(AudiomackIE.create_song_dictionary(api_response, album_url_tag, track_no))
+            track_no += 1
+        return result