def echoprint_lookup(file):
# Note that song.identify reads just the first 30 seconds of the file
fp = song.util.codegen(file)
if len(fp) and "code" in fp[0]:
# The version parameter to song/identify indicates the use of echoprint
result = song.identify(query_obj=fp, version="4.11")
pprint.pprint(result)
def echoprint_lookup(file):
# Note that song.identify reads just the first 30 seconds of the file
fp = song.util.codegen(file)
if len(fp) and "code" in fp[0]:
# The version parameter to song/identify indicates the use of echoprint
result = song.identify(query_obj=fp, version="4.11")
pprint.pprint(result)
def main(folder):
if folder.endswith('.mp3'):
files = [folder]
else:
# iterate over all .mp3 files in this directory, and identify them.
files = [fullpath(folder, f) for f in os.listdir(folder) if f.endswith('.mp3')]
updated = 0
for path in files:
print os.path.basename(path)
songs = None
try:
songs = song.identify(path, buckets=['audio_summary'])
except EchoNestAPIError, e:
print "API Error: %s" % e
continue
s = songs and songs.pop()
if not s:
print "Couldn't resolve %s" % path
continue
m = Metadata(s)
m.write_id3(path, create_only=False, replace=False)
updated += 1
def id_song(filename):
metadata = {}
config.ECHO_NEST_API_KEY = API_KEY
song_info = song.identify(filename)
metadata["title"]= song_info[0].title
metadata["artist"] = song_info[0].artist_name
#results = song.search(metadata["title"], metadata["artist"],
#buckets=["tracks", "id:7digital-US"])[0]
return metadata
def fetch_data(filepaths, result=None, overwrite=False, checkpoint_file=''):
"""
Parameters
----------
filepaths : list
Collection of audio files on disk to query against the EchoNest API.
result : dict, or None
Dictionary to add info; will create if None.
overwrite : bool, default=False
If False, will skip any keys contained in `result`.
checkpoint_file : str, or None
Path to write results as they are accumulated; ignored if empty.
Returns
-------
result : dict
Map of filebases to metadata.
"""
throttle = Throttle()
throttle.touch()
if result is None:
result = dict()
filepaths = set(filepaths)
while filepaths:
fpath = filepaths.pop()
key = futil.filebase(fpath)
# If we've already got data and we're not overwriting, move on.
if key in result and not overwrite:
print "[%s] %4d: '%s'" % (time.asctime(), len(filepaths), key)
continue
try:
# Otherwise, let's make some requests.
print "[%s] %4d: '%s'" % (time.asctime(), len(filepaths), key)
song = S.identify(filename=fpath,
codegen_start=0,
codegen_duration=60)
if song:
result[key] = extract_info(song[0])
if checkpoint_file:
with open(checkpoint_file, 'w') as fp:
json.dump(result, fp, indent=2)
throttle.wait()
except S.util.EchoNestAPIError as err:
if err.http_status == 429:
print "You got rate limited braah ... hang on."
throttle.wait(10)
filepaths.add(fpath)
elif err.http_status >= 500:
print "Server error; moving on, dropping key: %s" % key
except socket.error as err:
print "Socket Error %s" % err
filepaths.add(fpath)
throttle.wait(10)
return result
def lookup(file):
# Note that song.identify reads just the first 30 seconds of the file
fp = song.util.codegen(file)
if len(fp) and "code" in fp[0]:
# The version parameter to song/identify indicates the use of echoprint
result = song.identify(query_obj=fp, version="4.11")
print "Got result:", result
print "Artist: %s (%s)" % (result[0].artist_name, result[0].artist_id)
print "Song: %s (%s)" % (result[0].title, result[0].id)
else:
print "Couldn't decode", file
def fetch_data(filepaths, result=None, overwrite=False, checkpoint_file=''):
"""
Parameters
----------
filepaths : list
Collection of audio files on disk to query against the EchoNest API.
result : dict, or None
Dictionary to add info; will create if None.
overwrite : bool, default=False
If False, will skip any keys contained in `result`.
checkpoint_file : str, or None
Path to write results as they are accumulated; ignored if empty.
Returns
-------
result : dict
Map of filebases to metadata.
"""
throttle = Throttle()
throttle.touch()
if result is None:
result = dict()
filepaths = set(filepaths)
while filepaths:
fpath = filepaths.pop()
key = futil.filebase(fpath)
# If we've already got data and we're not overwriting, move on.
if key in result and not overwrite:
print "[%s] %4d: '%s'" % (time.asctime(), len(filepaths), key)
continue
try:
# Otherwise, let's make some requests.
print "[%s] %4d: '%s'" % (time.asctime(), len(filepaths), key)
song = S.identify(
filename=fpath, codegen_start=0, codegen_duration=60)
if song:
result[key] = extract_info(song[0])
if checkpoint_file:
with open(checkpoint_file, 'w') as fp:
json.dump(result, fp, indent=2)
throttle.wait()
except S.util.EchoNestAPIError as err:
if err.http_status == 429:
print "You got rate limited braah ... hang on."
throttle.wait(10)
filepaths.add(fpath)
elif err.http_status >= 500:
print "Server error; moving on, dropping key: %s" % key
except socket.error as err:
print "Socket Error %s" % err
filepaths.add(fpath)
throttle.wait(10)
return result
def update_echonest_tags(self):
"""Obtain echoprint tags (if exist) for sourcefile. Generate echoprint
code if required"""
# Fragments taken from echonest's echonest-codegen lookup.py script
if not self.echoprint_code:
self.echoprint_codegen()
# The version parameter to song/identify indicates the use of echoprint
result = song.identify(query_obj=self.echoprint_code, version="4.11")
if len(result):
self.echoprint_tags = result
def scanFolder(folder):
if isinstance(folder, list):
files = folder
elif os.path.isdir(folder):
files = getNamesAndGenreFromFolder(folder)
else:
return -1
notAtEcho=[]
notScanned=[]
foundTracks = []
retry=[]
i=0
if not os.path.isdir('echoNestResults'):
os.mkdir('echoNestResults')
for f in files:
time = t.localtime()
time = str(time.tm_hour) +'_' + str(time.tm_min) + '_' + str(time.tm_sec)
try:
print 'decodeing'
print f
code = getENMFPcode(f)
try:
print 'looking up on echonest'
track = song.identify(query_obj=code)
except:
print "couldn't connect to server"
retry.append(f)
if not track:
print 'song not at echonest'
notAtEcho.append(f)
continue
print 'track:',track
track=track[0]
trackInfo = [f, track.audio_summary['tempo'], track.audio_summary['time_signature'], track.id]
foundTracks.append(trackInfo)
pickle.dump(foundTracks, open('echoNestResults/'+str(i)+'-'+time+'.pkl', 'w'))
try:
tag = d3.load(f).tag
tag.copyright_url = 'echo'
tag.bpm = track.audio_summary['tempo']
tag.save()
except:
print 'id3 fail'
except :
print "couldn't scan", sys.exc_info()[0]
notScanned.append(f)
i+=1
return foundTracks, notAtEcho, notScanned, retry
def lookup(file):
# Note that song.identify reads just the first 30 seconds of the file
# Also you'll need the codegen binary
fp = song.util.codegen(file)
if len(fp) and "code" in fp[0]:
result = song.identify(query_obj=fp, version="4.12")
if len(result):
artist = result[0].artist_name
title = result[0].title
return (title, artist)
else:
return None
else:
print "Error: Couldn't decode", file
def lookup(file):
# song.identify reads the first 30 seconds of the file
fp = song.util.codegen(file)
if len(fp) and "code" in fp[0]:
# The version parameter to song/identify indicates the use of echoprint
result = song.identify(query_obj=fp, version="4.11")
print "Got result:", result
if len(result):
print "Artist: %s (%s)" % (result[0].artist_name, result[0].artist_id)
print "Song: %s (%s)" % (result[0].title, result[0].id)
else:
print "No match. This track may not be in the database yet."
else:
print "Couldn't decode", file
def lookup(file):
# Note that song.identify reads just the first 30 seconds of the file
fp = song.util.codegen(file)
if len(fp) and "code" in fp[0]:
# The version parameter to song/identify indicates the use of echoprint
result = song.identify(query_obj=fp, version="4.12")
print "Got result:", result
if len(result):
print "Artist: %s (%s)" % (result[0].artist_name, result[0].artist_id)
print "Song: %s (%s)" % (result[0].title, result[0].id)
else:
print "No match. This track may not be in the database yet."
else:
print "Couldn't decode", file
def identify_song(tag):
"""
Uses the Echonest API to try identifying a song
Returns:
None if no results were found from any hosts
Returns a list of artist + title names if successful
"""
result = song.identify(query_obj=tag)
if result:
return [{
'artist': r.artist_name,
'title': r.title
} for r in result]
else:
return None
def store_en_info(enmfp, url):
u = urlparse(url)
s = song.identify(code=enmfp)[0] #TODO deal with more than one match
a = artist.Artist(id=s.artist_id)
#TODO find out if we can get release info too
redis_client.set('en_songs:%s' % s.id, json.dumps(s.__dict__))
redis_client.set('en_artists:%s' % a.id, json.dumps(a.__dict__))
redis_client.set('en_enmfp_by_path:%s' % u.path, enmfp)
redis_client.set('en_song_ids_by_enmfp:%s' % enmfp, s.id)
# Only lookup/store artist terms if we dont have yet
key = 'en_artist_terms:%s' % a.id
if not redis_client.get(key):
redis_client.set(key, json.dumps(a.terms))
return {'song': s, 'artist': a}
def store_en_info(enmfp, url):
u = urlparse(url)
s = song.identify(code=enmfp)[0] #TODO deal with more than one match
a = artist.Artist(id=s.artist_id)
#TODO find out if we can get release info too
redis_client.set('en_songs:%s' % s.id, json.dumps(s.__dict__))
redis_client.set('en_artists:%s' % a.id, json.dumps(a.__dict__))
redis_client.set('en_enmfp_by_path:%s' % u.path, enmfp)
redis_client.set('en_song_ids_by_enmfp:%s' % enmfp, s.id)
# Only lookup/store artist terms if we dont have yet
key = 'en_artist_terms:%s' % a.id
if not redis_client.get(key):
redis_client.set(key, json.dumps(a.terms))
return {'song': s, 'artist': a}
def store_en_info(url):
u = urlparse(url)
# Calc or fetch enmfp from cache
sha = sha256_for_file(u.path)
key = 'en_enmfp_by_sha:%s' % sha
enmfp = redis_client.get(key)
if not enmfp:
enmfp = get_enmfp_json_from_mp3(url)
enmfp = redis_client.set(key, enmfp)
# Do we have this song result in redis?
en_song_id = redis_client.get('en_song_ids_by_enmfp:%s' % enmfp)
if en_song_id:
return {'result': 'exists'} # we've already handled this song, so dont do anyhting for it
# Still here? Go to EN API for the song and artist info and store it
s_results = song.identify(code=enmfp)
if len(s_results) == 0:
error = "*** ID FAIL %s with code %s" % (url, enmfp)
print error
return { 'result': 'error', 'error': error}
else:
s = s_results[0]
print "ID %s" % (url)
a = artist.Artist(id=s.artist_id)
#TODO get release info somehow (via en track query then rosetta lookup?)
# Add artist terms to artist obj for json dump
ao = a.__dict__
ao['terms'] = a.terms
redis_client.set('en_songs:%s' % s.id, json.dumps(s.__dict__))
redis_client.set('en_artists:%s' % a.id, json.dumps(ao))
redis_client.set('en_enmfp_by_path:%s' % u.path, enmfp)
redis_client.set('en_song_ids_by_enmfp:%s' % enmfp, s.id)
# Only lookup/store artist terms if we dont have yet
key = 'en_artist_terms:%s' % a.id
if not redis_client.get(key):
redis_client.set(key, json.dumps(a.terms))
return {'result': 'ok', 'song': s, 'artist': a, 'enmfp': enmfp}
def lookup(file):
# Note that song.identify reads just the first 30 seconds of the file
fp = song.util.codegen(file)
if len(fp) and "code" in fp[0]:
# The version parameter to song/identify indicates the use of echoprint
result = song.identify(query_obj=fp, version="4.11")
# print "Got result:", result
if len(result):
print '{'
print ' "artist": { "name": "%s", "id": "%s" },' % (
result[0].artist_name.encode('utf-8'), result[0].artist_id)
print ' "song": { "title": "%s", "id": "%s" }' % (
result[0].title.encode('utf-8'), result[0].id)
print '}'
else:
print '{ "error": "No match. This track may not be in the database yet" }'
else:
print '{ "error": "Couldn\'t decode %s" }' % file
def extract_using_echonest(file_):
# Get length
try:
madf = mad.MadFile(file_)
length = madf.total_time() // 1000
except:
length = None
# Identify via Echo Nest
config.ECHO_NEST_API_KEY = 'UR4VKX7JXDXAULIWB'
config.CODEGEN_BINARY_OVERRIDE = '/usr/bin/echoprint-codegen'
try:
possible_songs = identify(filename=file_)
if possible_songs:
song = possible_songs[0]
return SongInfo(song.title, song.artist_name, '', length, None, None, None)
else:
return empty_song_info()
except:
return empty_song_info()
def match_song(duration, filename):
"""return artist and title of song in filename"""
if FP_SERVICE == "acoustid":
try:
results = acoustid.match(ACOUSTID_API_KEY, filename)
except:
return (None, None)
try:
score, recording_id, title, artist = results.next()
return Song(artist=artist, title=title)
except:
return (None, None)
elif FP_SERVICE == "echonest":
try:
results = ensong.identify(filename,
codegen_duration=duration, codegen_start=0)
song = results[0]
return Song(artist=song.artist_name, title=song.title)
except:
return (None, None)
return (None, None)
def match_song(duration, filename):
"""return artist and title of song in filename"""
if FP_SERVICE == "acoustid":
try:
results = acoustid.match(ACOUSTID_API_KEY, filename)
except:
return (None, None)
try:
score, recording_id, title, artist = results.next()
return Song(artist=artist, title=title)
except:
return (None, None)
elif FP_SERVICE == "echonest":
try:
results = ensong.identify(filename,
codegen_duration=duration,
codegen_start=0)
song = results[0]
return Song(artist=song.artist_name, title=song.title)
except:
return (None, None)
return (None, None)
def _query_echonest(self, fingerprint):
result = song.identify(code=fingerprint)
return result
def tagSong(filename):
# Get status info
statusfile = os.path.expanduser("~/.scrobbyl")
lines = []
if os.path.exists(statusfile):
fp = open(statusfile, "r")
lines = fp.readlines()
fp.close()
lasttime = 0
lastartist = ""
lasttrack = ""
lastscrobble = 0
if len(lines) == 4:
lasttime = int(lines[0])
lastartist = lines[1].strip()
lasttrack = lines[2].strip()
lastscrobble = int(lines[3])
fp = song.util.codegen(filename)
pprint.pprint(fp)
# Make sure we have a valid fp
if fp == None or len(fp) == 0 or "code" not in fp[0]:
raise Exception("Could not calculate fingerprint!")
result = song.identify(query_obj=fp, version="4.11", buckets="audio_summary")
pprint.pprint(result)
if len(result) == 0:
raise Exception("Song not found in database.")
track = result[0].title
artist = result[0].artist_name
songlength = result[0].audio_summary.duration
# Check to make sure it's not a duplicate
now = time.time()
doscrobble = True
if lasttime:
if (now - lasttime) < 100:
# Check for duplicate song
if lasttrack == track and lastartist == artist:
# if config.getboolean('scrobble_rules','allow_repeat') and prevlength < (now - config.get('runtime_info','last_scrobble')):
if False:
print "Same song, but looks like it repeated..."
else:
print "Same song as last time, skipping..."
doscrobble = False
else:
print "New song!"
else:
print "It's been longer than the songlength since we last checked."
else:
lasttrack = lastartist = "none"
print "No previous song found."
if doscrobble:
print "Last track was %s by %s, now %s by %s. Scrobbling..." % (lasttrack, lastartist, track, artist)
# lastfm.scrobble(artist, track)
lastscrobble = now
fp = open(statusfile, "w")
fp.write("%d\n%s\n%s\n%d" % (now, artist, track, lastscrobble))
fp.close()
return (track, artist, doscrobble)
def tagSong(filename):
#Get status info
statusfile = os.path.expanduser("~/.scrobbyl")
lines = []
if os.path.exists(statusfile):
fp = open(statusfile, "r")
lines = fp.readlines()
fp.close()
lasttime = 0
lastartist = ""
lasttrack = ""
lastscrobble = 0
if len(lines) == 4:
lasttime = int(lines[0])
lastartist = lines[1].strip()
lasttrack = lines[2].strip()
lastscrobble = int(lines[3])
fp = song.util.codegen(filename)
pprint.pprint(fp)
#Make sure we have a valid fp
if fp == None or len(fp) == 0 or "code" not in fp[0]:
raise Exception("Could not calculate fingerprint!")
result = song.identify(query_obj=fp,
version="4.11",
buckets="audio_summary")
pprint.pprint(result)
if len(result) == 0:
raise Exception("Song not found in database.")
track = result[0].title
artist = result[0].artist_name
songlength = result[0].audio_summary.duration
#Check to make sure it's not a duplicate
now = time.time()
doscrobble = True
if lasttime:
if (now - lasttime) < 100:
#Check for duplicate song
if lasttrack == track and lastartist == artist:
#if config.getboolean('scrobble_rules','allow_repeat') and prevlength < (now - config.get('runtime_info','last_scrobble')):
if False:
print "Same song, but looks like it repeated..."
else:
print "Same song as last time, skipping..."
doscrobble = False
else:
print "New song!"
else:
print "It's been longer than the songlength since we last checked."
else:
lasttrack = lastartist = "none"
print "No previous song found."
if (doscrobble):
print "Last track was %s by %s, now %s by %s. Scrobbling..." % (
lasttrack, lastartist, track, artist)
#lastfm.scrobble(artist, track)
lastscrobble = now
fp = open(statusfile, "w")
fp.write("%d\n%s\n%s\n%d" % (now, artist, track, lastscrobble))
fp.close()
return (track, artist, doscrobble)
def lookupEcho(code):
track = song.identify(query_obj=code)
return track
