Added 'helldon' to transparently port the thing to the desktop :P
[jamaendo] / jamaendo / api.py
index 1e052a8..7792c33 100644 (file)
@@ -8,12 +8,6 @@
 # modification, are permitted provided that the following conditions are met:
 #     * Redistributions of source code must retain the above copyright
 #       notice, this list of conditions and the following disclaimer.
-#     * Redistributions in binary form must reproduce the above copyright
-#       notice, this list of conditions and the following disclaimer in the
-#       documentation and/or other materials provided with the distribution.
-#     * Neither the name of Jamaendo nor the
-#       names of its contributors may be used to endorse or promote products
-#       derived from this software without specific prior written permission.
 #
 # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
 # ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
 
 # An improved, structured jamendo API wrapper for the N900 with cacheing
 # Image / cover downloads.. and more?
-import urllib, threading, os, gzip, time, simplejson, re
+import urllib, threading, os, time, simplejson, re
+import logging, hashlib
+import pycurl, StringIO
 
 _CACHEDIR = None
 _COVERDIR = None
 _GET2 = '''http://api.jamendo.com/get2/'''
 _MP3URL = _GET2+'stream/track/redirect/?id=%d&streamencoding=mp31'
 _OGGURL = _GET2+'stream/track/redirect/?id=%d&streamencoding=ogg2'
-
-
-def set_cache_dir(cachedir):
-    global _CACHEDIR
-    global _COVERDIR
-    _CACHEDIR = cachedir
-    _COVERDIR = os.path.join(_CACHEDIR, 'covers')
-
-    try:
-        os.makedirs(_CACHEDIR)
-    except OSError:
-        pass
-
-    try:
-        os.makedirs(_COVERDIR)
-    except OSError:
-        pass
+_TORRENTURL = _GET2+'bittorrent/file/redirect/?album_id=%d&type=archive&class=mp32'
+
+try:
+    log = logging.getLogger(__name__)
+except:
+    class StdoutLogger(object):
+        def info(self, s, *args):
+            print s % (args)
+        def debug(self, s, *args):
+            pass#print s % (args)
+    log = StdoutLogger()
 
 # These classes can be partially constructed,
 # and if asked for a property they don't know,
@@ -59,8 +49,22 @@ def set_cache_dir(cachedir):
 
 _ARTIST_FIELDS = ['id', 'name', 'image']
 _ALBUM_FIELDS = ['id', 'name', 'image', 'artist_name', 'artist_id', 'license_url']
-_TRACK_FIELDS = ['id', 'name', 'image', 'artist_name', 'album_name', 'album_id', 'numalbum', 'duration']
+_TRACK_FIELDS = ['id', 'name', 'album_image', 'artist_id', 'artist_name', 'album_name', 'album_id', 'numalbum', 'duration']
 _RADIO_FIELDS = ['id', 'name', 'idstr', 'image']
+_TAG_FIELDS = ['id', 'name']
+
+def curlGET(url):
+    c = pycurl.Curl()
+    s = StringIO.StringIO()
+    c.setopt(pycurl.FOLLOWLOCATION, 1)
+    c.setopt(pycurl.URL, url)
+    c.setopt(pycurl.WRITEFUNCTION, s.write)
+    try:
+        c.perform()
+    finally:
+        c.close()
+    s.seek(0)
+    return s.read()
 
 class LazyQuery(object):
     def set_from_json(self, json):
@@ -104,10 +108,10 @@ class LazyQuery(object):
     def __repr__(self):
         try:
             return u"%s(%s)"%(self.__class__.__name__,
-                              u", ".join(repr(v) for k,v in self.__dict__.iteritems() if not k.startswith('_')))
+                              u", ".join(("%s:%s"%(k,repr(v))) for k,v in self.__dict__.iteritems() if not k.startswith('_')))
         except UnicodeEncodeError:
-            import traceback
-            traceback.print_exc()
+            #import traceback
+            #traceback.print_exc()
             return u"%s(?)"%(self.__class__.__name__)
 
 class Artist(LazyQuery):
@@ -125,6 +129,9 @@ class Artist(LazyQuery):
     def _set_from(self, other):
         return self._set_from_impl(other, 'name', 'image', 'albums')
 
+    def get_data(self):
+        return {'name':self.name, 'image':self.image}
+
 class Album(LazyQuery):
     def __init__(self, ID, json=None):
         self.ID = int(ID)
@@ -137,18 +144,29 @@ class Album(LazyQuery):
         if json:
             self.set_from_json(json)
 
+    def torrent_url(self):
+        return _TORRENTURL%(self.ID)
+
+
     def _needs_load(self):
         return self._needs_load_impl('name', 'image', 'artist_name', 'artist_id', 'license_url', 'tracks')
 
     def _set_from(self, other):
         return self._set_from_impl(other, 'name', 'image', 'artist_name', 'artist_id', 'license_url', 'tracks')
 
+    def get_data(self):
+        return {'name':self.name, 'image':self.image,
+                'artist_name':self.artist_name,
+                'artist_id':self.artist_id,
+                'license_url':self.license_url}
+
 class Track(LazyQuery):
     def __init__(self, ID, json=None):
         self.ID = int(ID)
         self.name = None
-        self.image = None
+        self.artist_id = None
         self.artist_name = None
+        self.album_image = None
         self.album_name = None
         self.album_id = None
         self.numalbum = None
@@ -162,11 +180,21 @@ class Track(LazyQuery):
     def ogg_url(self):
        return _OGGURL%(self.ID)
 
+    def get_data(self):
+        return {'name':self.name,
+                'artist_id':self.artist_id,
+                'artist_name':self.artist_name,
+                'album_image':self.album_image,
+                'album_name':self.album_name,
+                'album_id':self.album_id,
+                'numalbum':self.numalbum,
+                'duration':self.duration}
+
     def _needs_load(self):
-        return self._needs_load_impl('name', 'artist_name', 'album_name', 'album_id', 'numalbum', 'duration')
+        return self._needs_load_impl('name', 'artist_name', 'artist_id', 'album_name', 'album_id', 'numalbum', 'duration')
 
     def _set_from(self, other):
-        return self._set_from_impl(other, 'name', 'image', 'artist_name', 'album_name', 'album_id', 'numalbum', 'duration')
+        return self._set_from_impl(other, 'name', 'album_image', 'artist_name', 'artist_id', 'album_name', 'album_id', 'numalbum', 'duration')
 
 class Radio(LazyQuery):
     def __init__(self, ID, json=None):
@@ -183,6 +211,18 @@ class Radio(LazyQuery):
     def _set_from(self, other):
         return self._set_from_impl(other, 'name', 'idstr', 'image')
 
+class Tag(LazyQuery):
+    def __init__(self, ID, json=None):
+        self.ID = int(ID)
+        self.name = None
+        if json:
+            self.set_from_json(json)
+
+    def _needs_load(self):
+        return self._needs_load_impl('name')
+
+    def _set_from(self, other):
+        return self._set_from_impl(other, 'name')
 
 _artists = {} # id -> Artist()
 _albums = {} # id -> Album()
@@ -195,6 +235,8 @@ _CACHED_ARTISTS = 100
 _CACHED_ALBUMS = 200
 _CACHED_TRACKS = 500
 _CACHED_RADIOS = 10
+# cache sizes, persistant
+_CACHED_COVERS = 2048
 
 # TODO: cache queries?
 
@@ -213,11 +255,12 @@ class Query(object):
         pass
 
     def _geturl(self, url):
-        print "*** %s" % (url)
+        log.info("%s", url)
         Query._ratelimit()
-        f = urllib.urlopen(url)
-        ret = simplejson.load(f)
-        f.close()
+        try:
+            ret = simplejson.loads(curlGET(url))
+        except Exception, e:
+            return None
         return ret
 
     def __str__(self):
@@ -226,6 +269,96 @@ class Query(object):
     def execute(self):
         raise NotImplemented
 
+class CoverFetcher(threading.Thread):
+    def __init__(self):
+        threading.Thread.__init__(self)
+        self.setDaemon(True)
+        self.cond = threading.Condition()
+        self.work = []
+
+    def _retrieve(self, url, fname):
+        BROKEN = 'http://imgjam.com/radios/default/default.100.png'
+        if url == BROKEN:
+            return None
+        f = open(fname, 'wb')
+        c = pycurl.Curl()
+        c.setopt(pycurl.FOLLOWLOCATION, 1)
+        c.setopt(pycurl.URL, str(url))
+        c.setopt(pycurl.WRITEFUNCTION, f.write)
+        try:
+            c.perform()
+        except:
+            fname = None
+        finally:
+            c.close()
+            f.close()
+        log.debug("Coverfetch: %s -> %s", url, fname)
+        return fname
+
+    def _fetch_cover(self, albumid, size):
+        try:
+            coverdir = _COVERDIR if _COVERDIR else '/tmp'
+            to = os.path.join(coverdir, '%d-%d.jpg'%(albumid, size))
+            if not os.path.isfile(to):
+                url = _GET2+'image/album/redirect/?id=%d&imagesize=%d'%(albumid, size)
+                to = self._retrieve(url, to)
+            return to
+        except Exception, e:
+            return None
+
+    def _fetch_image(self, url):
+        try:
+            h = hashlib.md5(url).hexdigest()
+            coverdir = _COVERDIR if _COVERDIR else '/tmp'
+            to = os.path.join(coverdir, h+'.jpg')
+            if not os.path.isfile(to):
+                to = self._retrieve(url, to)
+            return to
+        except Exception, e:
+            return None
+
+    def request_cover(self, albumid, size, cb):
+        self.cond.acquire()
+        self.work.insert(0, (albumid, size, cb))
+        self.cond.notify()
+        self.cond.release()
+
+    def request_images(self, urls, cb):
+        """cb([(url, image)])"""
+        self.cond.acquire()
+        self.work = [('image', url, cb) for url in urls] + self.work
+        self.cond.notify()
+        self.cond.release()
+
+    def run(self):
+        while True:
+            work = []
+            self.cond.acquire()
+            while True:
+                work = self.work
+                if work:
+                    self.work = []
+                    break
+                self.cond.wait()
+            self.cond.release()
+
+            multi = len(work) > 1
+            for job in work:
+                if job[0] == 'image':
+                    self.process_image(job[1], job[2])
+                else:
+                    self.process_cover(*job)
+
+    def process_cover(self, albumid, size, cb):
+        cover = self._fetch_cover(albumid, size)
+        if cover:
+            cb(albumid, size, cover)
+
+    def process_image(self, url, cb):
+        image = self._fetch_image(url)
+        if image:
+            cb([(url, image)])
+
 class CoverCache(object):
     """
     cache and fetch covers
@@ -235,23 +368,47 @@ class CoverCache(object):
     """
     def __init__(self):
         self._covers = {} # (albumid, size) -> file
-        coverdir = _COVERDIR if _COVERDIR else '/tmp'
-        if os.path.isdir(coverdir):
-            covermatch = re.compile(r'(\d+)\-(\d+)\.jpg')
-            for fil in os.listdir(coverdir):
-                fl = os.path.join(coverdir, fil)
-                m = covermatch.match(fil)
-                if m and os.path.isfile(fl):
-                    self._covers[(int(m.group(1)), int(m.group(2)))] = fl
+        self._images = {}
+        self._fetcher = CoverFetcher()
+        self._fetcher.start()
+        if _COVERDIR and os.path.isdir(_COVERDIR):
+            self.prime_cache()
+
+    def prime_cache(self):
+        coverdir = _COVERDIR
+        covermatch = re.compile(r'(\d+)\-(\d+)\.jpg')
+
+        prev_covers = os.listdir(coverdir)
+
+        if len(prev_covers) > _CACHED_COVERS:
+            import random
+            dropn = len(prev_covers) - _CACHED_COVERS
+            todrop = random.sample(prev_covers, dropn)
+            log.warning("Deleting from cache: %s", todrop)
+            for d in todrop:
+                m = covermatch.match(d)
+                if m:
+                    try:
+                        os.unlink(os.path.join(coverdir, d))
+                    except OSError, e:
+                        log.exception('unlinking failed')
+
+        for fil in os.listdir(coverdir):
+            fl = os.path.join(coverdir, fil)
+            m = covermatch.match(fil)
+            if m and os.path.isfile(fl):
+                self._covers[(int(m.group(1)), int(m.group(2)))] = fl
 
     def fetch_cover(self, albumid, size):
-        coverdir = _COVERDIR if _COVERDIR else '/tmp'
-        to = os.path.join(coverdir, '%d-%d.jpg'%(albumid, size))
-        if not os.path.isfile(to):
-            url = _GET2+'image/album/redirect/?id=%d&imagesize=%d'%(albumid, size)
-            urllib.urlretrieve(url, to)
-            self._covers[(albumid, size)] = to
-        return to
+        coverdir = _COVERDIR
+        if coverdir:
+            to = os.path.join(coverdir, '%d-%d.jpg'%(albumid, size))
+            if not os.path.isfile(to):
+                url = _GET2+'image/album/redirect/?id=%d&imagesize=%d'%(albumid, size)
+                urllib.urlretrieve(url, to)
+                self._covers[(albumid, size)] = to
+            return to
+        return None
 
     def get_cover(self, albumid, size):
         cover = self._covers.get((albumid, size), None)
@@ -262,20 +419,61 @@ class CoverCache(object):
     def get_async(self, albumid, size, cb):
         cover = self._covers.get((albumid, size), None)
         if cover:
-            cb(cover)
+            cb(albumid, size, cover)
         else:
-            # TODO
-            cover = self.fetch_cover(albumid, size)
-            cb(cover)
+            def cb2(albumid, size, cover):
+                self._covers[(albumid, size)] = cover
+                cb(albumid, size, cover)
+            self._fetcher.request_cover(albumid, size, cb2)
+
+    def get_images_async(self, url_list, cb):
+        found = []
+        lookup = []
+        for url in url_list:
+            image = self._images.get(url, None)
+            if image:
+                found.append((url, image))
+            else:
+                lookup.append(url)
+        if found:
+            cb(found)
+
+        if lookup:
+            def cb2(results):
+                for url, image in results:
+                    self._images[url] = image
+                cb(results)
+            self._fetcher.request_images(lookup, cb2)
 
 _cover_cache = CoverCache()
 
-def get_album_cover(albumid, size=200):
+def set_cache_dir(cachedir):
+    global _CACHEDIR
+    global _COVERDIR
+    _CACHEDIR = cachedir
+    _COVERDIR = os.path.join(_CACHEDIR, 'covers')
+
+    try:
+        os.makedirs(_CACHEDIR)
+    except OSError:
+        pass
+
+    try:
+        os.makedirs(_COVERDIR)
+    except OSError:
+        pass
+
+    _cover_cache.prime_cache()
+
+def get_album_cover(albumid, size=100):
     return _cover_cache.get_cover(albumid, size)
 
-def get_album_cover_async(cb, albumid, size=200):
+def get_album_cover_async(cb, albumid, size=100):
     _cover_cache.get_async(albumid, size, cb)
 
+def get_images_async(cb, url_list):
+    _cover_cache.get_images_async(url_list, cb)
+
 class CustomQuery(Query):
     def __init__(self, url):
         Query.__init__(self)
@@ -294,11 +492,21 @@ class GetQuery(Query):
             'params' : 'artist_id=%d',
             'constructor' : Artist
             },
+        'artist_list' : {
+            'url' : _GET2+'+'.join(_ALBUM_FIELDS)+'/artist/json/?',
+            'params' : 'artist_id=%s',
+            'constructor' : Album
+            },
         'album' : {
             'url' : _GET2+'+'.join(_ALBUM_FIELDS)+'/album/json/?',
             'params' : 'album_id=%d',
             'constructor' : Album
             },
+        'album_list' : {
+            'url' : _GET2+'+'.join(_ALBUM_FIELDS)+'/album/json/?',
+            'params' : 'album_id=%s',
+            'constructor' : Album
+            },
         'albums' : {
             'url' : _GET2+'+'.join(_ALBUM_FIELDS)+'/album/json/?',
             'params' : 'artist_id=%d',
@@ -309,6 +517,11 @@ class GetQuery(Query):
             'params' : 'id=%d',
             'constructor' : Track
             },
+        'track_list' : {
+            'url' : _GET2+'+'.join(_TRACK_FIELDS)+'/track/json/track_album+album_artist?',
+            'params' : 'id=%s',
+            'constructor' : Track
+            },
         'tracks' : {
             'url' : _GET2+'+'.join(_TRACK_FIELDS)+'/track/json/track_album+album_artist?',
             'params' : 'order=numalbum_asc&album_id=%d',
@@ -316,7 +529,7 @@ class GetQuery(Query):
             },
         'radio' : {
             'url' : _GET2+'+'.join(_TRACK_FIELDS)+'/track/json/radio_track_inradioplaylist+track_album+album_artist/?',
-            'params' : 'order=random_asc&radio_id=%d',
+            'params' : 'order=random_asc&radio_id=%d&n=16',
             'constructor' : [Track]
             },
         'favorite_albums' : {
@@ -324,11 +537,12 @@ class GetQuery(Query):
             'params' : 'user_idstr=%s',
             'constructor' : [Album]
             },
-    #http://api.jamendo.com/get2/id+name+url+image+artist_name/album/jsonpretty/album_user_starred/?user_idstr=sylvinus&n=all
-    #q = SearchQuery('album', user_idstr=user)
-
+        'tag' : {
+            'url' : _GET2+'+'.join(_TRACK_FIELDS)+'/track/json/track_album+album_artist?',
+            'params' : 'tag_id=%d&n=50&order=rating_desc',
+            'constructor' : [Track]
+            },
         }
-#http://api.jamendo.com/get2/id+name+image+artist_name+album_name+album_id+numalbum+duration/track/json/radio_track_inradioplaylist+track_album+album_artist/?order=numradio_asc&radio_id=283
 
     def __init__(self, what, ID):
         Query.__init__(self)
@@ -357,7 +571,7 @@ class GetQuery(Query):
         return self.url + self.params % (self.ID)
 
 class SearchQuery(GetQuery):
-    def __init__(self, what, query=None, order=None, user=None, count=10):
+    def __init__(self, what, query=None, order=None, user=None, count=20):
         GetQuery.__init__(self, what, None)
         self.query = query
         self.order = order
@@ -396,6 +610,20 @@ def _update_cache(cache, new_items):
             old._set_from(item)
         else:
             cache[item.ID] = item
+        if isinstance(item, Artist) and item.albums:
+            for album in item.albums:
+                _update_cache(_albums, album)
+        elif isinstance(item, Album) and item.tracks:
+            for track in item.tracks:
+                _update_cache(_tracks, track)
+    # enforce cache limits here!
+    # also, TODO: save/load cache between sessions
+    # that will require storing a timestamp with
+    # each item, though..
+    # perhaps,
+    # artists: 1 day - changes often
+    # albums: 2-5 days - changes less often (?)
+    # tracks: 1 week - changes rarely, queried often
 
 def get_artist(artist_id):
     """Returns: Artist"""
@@ -410,13 +638,61 @@ def get_artist(artist_id):
             a = a[0]
     return a
 
-def get_albums(artist_id):
+def get_artists(artist_ids):
+    """Returns: [Artist]"""
+    assert(isinstance(artist_ids, list))
+    found = []
+    lookup = []
+    for artist_id in artist_ids:
+        a = _artists.get(artist_id, None)
+        if not a:
+            lookup.append(artist_id)
+        else:
+            found.append(a)
+    if lookup:
+        q = GetQuery('artist_list', '+'.join(str(x) for x in lookup))
+        a = q.execute()
+        if not a:
+            raise JamendoAPIException(str(q))
+        _update_cache(_artists, a)
+        lookup = a
+    return found + lookup
+
+def get_album_list(album_ids):
     """Returns: [Album]"""
+    assert(isinstance(album_ids, list))
+    found = []
+    lookup = []
+    for album_id in album_ids:
+        a = _albums.get(album_id, None)
+        if not a:
+            lookup.append(album_id)
+        else:
+            found.append(a)
+    if lookup:
+        q = GetQuery('album_list', '+'.join(str(x) for x in lookup))
+        a = q.execute()
+        if not a:
+            raise JamendoAPIException(str(q))
+        _update_cache(_albums, a)
+        lookup = a
+    return found + lookup
+
+def get_albums(artist_id):
+    """Returns: [Album]
+    Parameter can either be an artist_id or a list of album ids.
+    """
+    if isinstance(artist_id, list):
+        return get_album_list(artist_id)
+    a = _artists.get(artist_id, None)
+    if a and a.albums:
+        return a.albums
+
     q = GetQuery('albums', artist_id)
     a = q.execute()
     if not a:
         raise JamendoAPIException(str(q))
-    _update_cache(_artists, a)
+    _update_cache(_albums, a)
     return a
 
 def get_album(album_id):
@@ -432,8 +708,36 @@ def get_album(album_id):
             a = a[0]
     return a
 
-def get_tracks(album_id):
+def get_track_list(track_ids):
     """Returns: [Track]"""
+    assert(isinstance(track_ids, list))
+    found = []
+    lookup = []
+    for track_id in track_ids:
+        a = _tracks.get(track_id, None)
+        if not a:
+            lookup.append(track_id)
+        else:
+            found.append(a)
+    if lookup:
+        q = GetQuery('track_list', '+'.join(str(x) for x in lookup))
+        a = q.execute()
+        if not a:
+            raise JamendoAPIException(str(q))
+        _update_cache(_tracks, a)
+        lookup = a
+    return found + lookup
+
+def get_tracks(album_id):
+    """Returns: [Track]
+    Parameter can either be an album_id or a list of track ids.
+    """
+    if isinstance(album_id, list):
+        return get_track_list(album_id)
+    a = _albums.get(album_id, None)
+    if a and a.tracks:
+        return a.tracks
+
     q = GetQuery('tracks', album_id)
     a = q.execute()
     if not a:
@@ -463,6 +767,16 @@ def get_radio_tracks(radio_id):
     _update_cache(_tracks, a)
     return a
 
+#http://api.jamendo.com/get2/id+name/track/plain/?tag_id=327&n=50&order=rating_desc
+def get_tag_tracks(tag_id):
+    """Returns: [Track]"""
+    q = GetQuery('tag', tag_id)
+    a = q.execute()
+    if not a:
+        raise JamendoAPIException(str(q))
+    _update_cache(_tracks, a)
+    return a
+
 def search_artists(query):
     """Returns: [Artist]"""
     q = SearchQuery('artist', query, 'searchweight_desc')
@@ -517,6 +831,33 @@ def tracks_of_the_week():
     _update_cache(_tracks, a)
     return a
 
+def top_artists(order='rating_desc', count=20):
+    """Returns: [Artist]"""
+    q = SearchQuery('artist', order=order, count=count)
+    a = q.execute()
+    if not a:
+        raise JamendoAPIException(str(q))
+    _update_cache(_artists, a)
+    return a
+
+def top_albums(order='rating_desc', count=20):
+    """Returns: [Album]"""
+    q = SearchQuery('album', order=order, count=count)
+    a = q.execute()
+    if not a:
+        raise JamendoAPIException(str(q))
+    _update_cache(_albums, a)
+    return a
+
+def top_tracks(order='rating_desc', count=20):
+    """Returns: [Track]"""
+    q = SearchQuery('track', order=order, count=count)
+    a = q.execute()
+    if not a:
+        raise JamendoAPIException(str(q))
+    _update_cache(_tracks, a)
+    return a
+
 def get_radio(radio_id):
     """Returns: Radio"""
     q = CustomQuery(_GET2+"id+name+idstr+image/radio/json?id=%d"%(radio_id))
@@ -535,6 +876,14 @@ def starred_radios():
         raise JamendoAPIException(str(q))
     return [Radio(int(radio['id']), json=radio) for radio in js]
 
+def top_tags(count=50, order='rating_desc'):
+    """Returns: [Tag]"""
+    q = CustomQuery(_GET2+"id+name/tag/json?n=%d&order=%s"%(count, order))
+    js = q.execute()
+    if not js:
+        raise JamendoAPIException(str(q))
+    return [Tag(int(tag['id']), json=tag) for tag in js]
+
 def favorite_albums(user):
     """Returns: [Album]"""
     q = SearchQuery('favorite_albums', user=user, count=20)
@@ -549,6 +898,7 @@ def favorite_albums(user):
 def _artist_loader(self):
     if self._needs_load():
         artist = get_artist(self.ID)
+        artist.albums = get_albums(self.ID)
         self._set_from(artist)
 Artist.load = _artist_loader