def extract_and_compress(btfeat,npicks,winsize,finaldim,seed=3232343,randproj=None):
"""
From a btfeat matrix, usually 12xLENGTH
Extracts 'npicks' windows of size 'winsize' equally spaced
Flatten these picks, pass them through a random projection, final
size is 'finaldim'
Returns matrix npicks x finaldim, or 0 x finaldim if problem
(btfeats not long enough for instance)
We could return less than npicks if not long enough!
For speed, we can compute the random projection once and pass it as an
argument.
"""
# features length
ftlen = btfeat.shape[1]
ndim = btfeat.shape[0]
# too small case
if ftlen < winsize:
return np.zeros((0,finaldim))
# random projection
if randproj is None:
randproj = RANDPROJ.proj_point5(ndim * winsize, finaldim, seed=seed)
# not big enough for number of picks, last one too large return just 1
if ftlen < int(ftlen * (npicks *1./(npicks+1))) + winsize:
pos = int( (ftlen-winsize) / 2.) # middle
picks = [ btfeat[:,pos:pos+winsize] ]
# regular case, picks will contain npicks
else:
picks = []
for k in range(1,npicks+1):
pos = int(ftlen * (k *1./(npicks+1)))
picks.append( btfeat[:,pos:pos+winsize] )
# project / compress these
projections = map(lambda x: np.dot(x.flatten(),randproj).reshape(1,finaldim), picks)
return np.concatenate(projections)
def avgcov_and_compress(feats,finaldim,seed=3232343,randproj=None):
"""
From a features matrix 12x.... (beat-aligned or not)
Compute the correlation matrix (12x12) and projects it (except diagonal)
to the given final dim
RETURN
vector 1xfinaldim or 0xfinaldim is problem
"""
# features length
ftlen = feats.shape[1]
ndim = feats.shape[0]
# too small case
if ftlen < 3:
return np.zeros((0,finaldim))
# random projection
if randproj is None:
# 12 + 78, 78=13*12/2
randproj = RANDPROJ.proj_point5(90, finaldim, seed=seed)
# corr
avg = np.average(feats,1)
cov = np.cov(feats)
covflat = []
for k in range(12):
covflat.extend( np.diag(cov,k) )
covflat = np.array(covflat)
feats = np.concatenate([avg,covflat])
# compress
return np.dot(feats.flatten(),randproj).reshape(1,finaldim)
def avgcov_and_compress(feats, finaldim, seed=3232343, randproj=None):
"""
From a features matrix 12x.... (beat-aligned or not)
Compute the correlation matrix (12x12) and projects it (except diagonal)
to the given final dim
RETURN
vector 1xfinaldim or 0xfinaldim is problem
"""
# features length
ftlen = feats.shape[1]
ndim = feats.shape[0]
# too small case
if ftlen < 3:
return np.zeros((0, finaldim))
# random projection
if randproj is None:
# 12 + 78, 78=13*12/2
randproj = RANDPROJ.proj_point5(90, finaldim, seed=seed)
# corr
avg = np.average(feats, 1)
cov = np.cov(feats)
covflat = []
for k in range(12):
covflat.extend(np.diag(cov, k))
covflat = np.array(covflat)
feats = np.concatenate([avg, covflat])
# compress
return np.dot(feats.flatten(), randproj).reshape(1, finaldim)
def extract_and_compress(btfeat,
npicks,
winsize,
finaldim,
seed=3232343,
randproj=None):
"""
From a btfeat matrix, usually 12xLENGTH
Extracts 'npicks' windows of size 'winsize' equally spaced
Flatten these picks, pass them through a random projection, final
size is 'finaldim'
Returns matrix npicks x finaldim, or 0 x finaldim if problem
(btfeats not long enough for instance)
We could return less than npicks if not long enough!
For speed, we can compute the random projection once and pass it as an
argument.
"""
# features length
ftlen = btfeat.shape[1]
ndim = btfeat.shape[0]
# too small case
if ftlen < winsize:
return np.zeros((0, finaldim))
# random projection
if randproj is None:
randproj = RANDPROJ.proj_point5(ndim * winsize, finaldim, seed=seed)
# not big enough for number of picks, last one too large return just 1
if ftlen < int(ftlen * (npicks * 1. / (npicks + 1))) + winsize:
pos = int((ftlen - winsize) / 2.) # middle
picks = [btfeat[:, pos:pos + winsize]]
# regular case, picks will contain npicks
else:
picks = []
for k in range(1, npicks + 1):
pos = int(ftlen * (k * 1. / (npicks + 1)))
picks.append(btfeat[:, pos:pos + winsize])
# # project / compress these
projections = map(
lambda x: np.dot(x.flatten(), randproj).reshape(1, finaldim), picks)
return np.concatenate(projections)
def cov_and_compress(feats,finaldim,seed=3232343,randproj=None):
"""
From a features matrix 12x.... (beat-aligned or not)
Compute the correlation matrix (12x12) and projects it (except diagonal)
to the given final dim
RETURN
vector 1xfinaldim or 0xfinaldim is problem
"""
# features length
ftlen = feats.shape[1]
ndim = feats.shape[0]
# too small case
if ftlen < 3:
return np.zeros((0,finaldim))
# random projection
if randproj is None:
# 12*12
randproj = RANDPROJ.proj_point5(144, finaldim, seed=seed)
# corr
fcov = np.cov(feats)
# compress
return np.dot(fcov.flatten(),randproj).reshape(1,finaldim)
def cov_and_compress(feats, finaldim, seed=3232343, randproj=None):
"""
From a features matrix 12x.... (beat-aligned or not)
Compute the correlation matrix (12x12) and projects it (except diagonal)
to the given final dim
RETURN
vector 1xfinaldim or 0xfinaldim is problem
"""
# features length
ftlen = feats.shape[1]
ndim = feats.shape[0]
# too small case
if ftlen < 3:
return np.zeros((0, finaldim))
# random projection
if randproj is None:
# 12*12
randproj = RANDPROJ.proj_point5(144, finaldim, seed=seed)
# corr
fcov = np.cov(feats)
# compress
return np.dot(fcov.flatten(), randproj).reshape(1, finaldim)
def process_filelist_test(filelist=None,
model=None,
tmpfilename=None,
npicks=None,
winsize=None,
finaldim=None,
K=1,
typecompress='picks'):
"""
Main function, process all files in the list (as long as their artist
is in testartist)
INPUT
filelist - a list of song files
model - h5 file containing feats and year for all train songs
tmpfilename - where to save our processed features
npicks - number of segments to pick per song
winsize - size of each segment we pick
finaldim - how many values do we keep
K - param of KNN (default 1)
typecompress - feature type, 'picks', 'corrcoeff' or 'cov'
must be the same as in training
"""
# sanity check
for arg in locals().values():
assert not arg is None, 'process_filelist_test, missing an argument, something still None'
if os.path.isfile(tmpfilename):
print 'ERROR: file', tmpfilename, 'already exists.'
return
if not os.path.isfile(model):
print 'ERROR: model', model, 'does not exist.'
return
# create kdtree
h5model = tables.openFile(model, mode='r')
assert h5model.root.data.feats.shape[
1] == finaldim, 'inconsistency in final dim'
kd = ANN.kdtree(h5model.root.data.feats)
# create outputfile
output = tables.openFile(tmpfilename, mode='a')
group = output.createGroup("/", 'data', 'TMP FILE FOR YEAR RECOGNITION')
output.createEArray(group,
'year_real',
tables.IntAtom(shape=()), (0, ),
'',
expectedrows=len(filelist))
output.createEArray(group,
'year_pred',
tables.Float64Atom(shape=()), (0, ),
'',
expectedrows=len(filelist))
# random projection
ndim = 12 # fixed in this dataset
if typecompress == 'picks':
randproj = RANDPROJ.proj_point5(ndim * winsize, finaldim)
elif typecompress == 'corrcoeff' or typecompress == 'cov':
randproj = RANDPROJ.proj_point5(ndim * ndim, finaldim)
elif typecompress == 'avgcov':
randproj = RANDPROJ.proj_point5(90, finaldim)
else:
assert False, 'Unknown type of compression: ' + str(typecompress)
# go through files
cnt_f = 0
for f in filelist:
cnt_f += 1
if cnt_f % 5000 == 0:
print 'TESTING FILE #' + str(cnt_f)
# check file
h5 = GETTERS.open_h5_file_read(f)
artist_id = GETTERS.get_artist_id(h5)
year = GETTERS.get_year(h5)
track_id = GETTERS.get_track_id(h5)
h5.close()
if year <= 0: # probably useless but...
continue
if typecompress == 'picks':
# we have a train artist with a song year, we're good
bttimbre = get_bttimbre(f)
if bttimbre is None:
continue
# we even have normal features, awesome!
processed_feats = CBTF.extract_and_compress(bttimbre,
npicks,
winsize,
finaldim,
randproj=randproj)
elif typecompress == 'corrcoeff':
h5 = GETTERS.open_h5_file_read(f)
timbres = GETTERS.get_segments_timbre(h5).T
h5.close()
processed_feats = CBTF.corr_and_compress(timbres,
finaldim,
randproj=randproj)
elif typecompress == 'cov':
h5 = GETTERS.open_h5_file_read(f)
timbres = GETTERS.get_segments_timbre(h5).T
h5.close()
processed_feats = CBTF.cov_and_compress(timbres,
finaldim,
randproj=randproj)
elif typecompress == 'avgcov':
h5 = GETTERS.open_h5_file_read(f)
timbres = GETTERS.get_segments_timbre(h5).T
h5.close()
processed_feats = CBTF.avgcov_and_compress(timbres,
finaldim,
randproj=randproj)
else:
assert False, 'Unknown type of compression: ' + str(typecompress)
if processed_feats is None:
continue
if processed_feats.shape[0] == 0:
continue
# do prediction
year_pred = do_prediction(processed_feats, kd, h5model, K)
# add pred and ground truth to output
if not year_pred is None:
output.root.data.year_real.append([year])
output.root.data.year_pred.append([year_pred])
# close output and model
del kd
h5model.close()
output.close()
# done
return
def process_filelist_train(filelist=None,testartists=None,tmpfilename=None,
npicks=None,winsize=None,finaldim=None,typecompress='picks'):
"""
Main function, process all files in the list (as long as their artist
is not in testartist)
INPUT
filelist - a list of song files
testartists - set of artist ID that we should not use
tmpfilename - where to save our processed features
npicks - number of segments to pick per song
winsize - size of each segment we pick
finaldim - how many values do we keep
typecompress - one of 'picks' (win of btchroma), 'corrcoef' (correlation coefficients),
'cov' (covariance)
"""
# sanity check
for arg in locals().values():
assert not arg is None,'process_filelist_train, missing an argument, something still None'
if os.path.isfile(tmpfilename):
print 'ERROR: file',tmpfilename,'already exists.'
return
# create outputfile
output = tables.openFile(tmpfilename, mode='a')
group = output.createGroup("/",'data','TMP FILE FOR YEAR RECOGNITION')
output.createEArray(group,'feats',tables.Float64Atom(shape=()),(0,finaldim),'',
expectedrows=len(filelist))
output.createEArray(group,'year',tables.IntAtom(shape=()),(0,),'',
expectedrows=len(filelist))
output.createEArray(group,'track_id',tables.StringAtom(18,shape=()),(0,),'',
expectedrows=len(filelist))
# random projection
ndim = 12 # fixed in this dataset
if typecompress == 'picks':
randproj = RANDPROJ.proj_point5(ndim * winsize, finaldim)
elif typecompress == 'corrcoeff' or typecompress == 'cov':
randproj = RANDPROJ.proj_point5(ndim * ndim, finaldim)
elif typecompress == 'avgcov':
randproj = RANDPROJ.proj_point5(90, finaldim)
else:
assert False,'Unknown type of compression: '+str(typecompress)
# iterate over files
cnt_f = 0
for f in filelist:
cnt_f += 1
# verbose
if cnt_f % 50000 == 0:
print 'training... checking file #',cnt_f
# check file
h5 = GETTERS.open_h5_file_read(f)
artist_id = GETTERS.get_artist_id(h5)
year = GETTERS.get_year(h5)
track_id = GETTERS.get_track_id(h5)
h5.close()
if year <= 0 or artist_id in testartists:
continue
# we have a train artist with a song year, we're good
bttimbre = get_bttimbre(f)
if typecompress == 'picks':
if bttimbre is None:
continue
# we even have normal features, awesome!
processed_feats = CBTF.extract_and_compress(bttimbre,npicks,winsize,finaldim,
randproj=randproj)
elif typecompress == 'corrcoeff':
h5 = GETTERS.open_h5_file_read(f)
timbres = GETTERS.get_segments_timbre(h5).T
h5.close()
processed_feats = CBTF.corr_and_compress(timbres,finaldim,randproj=randproj)
elif typecompress == 'cov':
h5 = GETTERS.open_h5_file_read(f)
timbres = GETTERS.get_segments_timbre(h5).T
h5.close()
processed_feats = CBTF.cov_and_compress(timbres,finaldim,randproj=randproj)
elif typecompress == 'avgcov':
h5 = GETTERS.open_h5_file_read(f)
timbres = GETTERS.get_segments_timbre(h5).T
h5.close()
processed_feats = CBTF.avgcov_and_compress(timbres,finaldim,randproj=randproj)
else:
assert False,'Unknown type of compression: '+str(typecompress)
# save them to tmp file
n_p_feats = processed_feats.shape[0]
output.root.data.year.append( np.array( [year] * n_p_feats ) )
output.root.data.track_id.append( np.array( [track_id] * n_p_feats ) )
output.root.data.feats.append( processed_feats )
# we're done, close output
output.close()
return
def process_filelist_train(filelist=None,
testartists=None,
tmpfilename=None,
npicks=None,
winsize=None,
finaldim=None,
typecompress='picks'):
"""
Main function, process all files in the list (as long as their artist
is not in testartist)
INPUT
filelist - a list of song files
testartists - set of artist ID that we should not use
tmpfilename - where to save our processed features
npicks - number of segments to pick per song
winsize - size of each segment we pick
finaldim - how many values do we keep
typecompress - one of 'picks' (win of btchroma), 'corrcoef' (correlation coefficients),
'cov' (covariance)
"""
# sanity check
for arg in locals().values():
assert not arg is None, 'process_filelist_train, missing an argument, something still None'
if os.path.isfile(tmpfilename):
print 'ERROR: file', tmpfilename, 'already exists.'
return
# create outputfile
output = tables.openFile(tmpfilename, mode='a')
group = output.createGroup("/", 'data', 'TMP FILE FOR YEAR RECOGNITION')
output.createEArray(group,
'feats',
tables.Float64Atom(shape=()), (0, finaldim),
'',
expectedrows=len(filelist))
output.createEArray(group,
'year',
tables.IntAtom(shape=()), (0, ),
'',
expectedrows=len(filelist))
output.createEArray(group,
'track_id',
tables.StringAtom(18, shape=()), (0, ),
'',
expectedrows=len(filelist))
# random projection
ndim = 12 # fixed in this dataset
if typecompress == 'picks':
randproj = RANDPROJ.proj_point5(ndim * winsize, finaldim)
elif typecompress == 'corrcoeff' or typecompress == 'cov':
randproj = RANDPROJ.proj_point5(ndim * ndim, finaldim)
elif typecompress == 'avgcov':
randproj = RANDPROJ.proj_point5(90, finaldim)
else:
assert False, 'Unknown type of compression: ' + str(typecompress)
# iterate over files
cnt_f = 0
for f in filelist:
cnt_f += 1
# verbose
if cnt_f % 50000 == 0:
print 'training... checking file #', cnt_f
# check file
h5 = GETTERS.open_h5_file_read(f)
artist_id = GETTERS.get_artist_id(h5)
year = GETTERS.get_year(h5)
track_id = GETTERS.get_track_id(h5)
h5.close()
if year <= 0 or artist_id in testartists:
continue
# we have a train artist with a song year, we're good
bttimbre = get_bttimbre(f)
if typecompress == 'picks':
if bttimbre is None:
continue
# we even have normal features, awesome!
processed_feats = CBTF.extract_and_compress(bttimbre,
npicks,
winsize,
finaldim,
randproj=randproj)
elif typecompress == 'corrcoeff':
h5 = GETTERS.open_h5_file_read(f)
timbres = GETTERS.get_segments_timbre(h5).T
h5.close()
processed_feats = CBTF.corr_and_compress(timbres,
finaldim,
randproj=randproj)
elif typecompress == 'cov':
h5 = GETTERS.open_h5_file_read(f)
timbres = GETTERS.get_segments_timbre(h5).T
h5.close()
processed_feats = CBTF.cov_and_compress(timbres,
finaldim,
randproj=randproj)
elif typecompress == 'avgcov':
h5 = GETTERS.open_h5_file_read(f)
timbres = GETTERS.get_segments_timbre(h5).T
h5.close()
processed_feats = CBTF.avgcov_and_compress(timbres,
finaldim,
randproj=randproj)
else:
assert False, 'Unknown type of compression: ' + str(typecompress)
# save them to tmp file
n_p_feats = processed_feats.shape[0]
output.root.data.year.append(np.array([year] * n_p_feats))
output.root.data.track_id.append(np.array([track_id] * n_p_feats))
output.root.data.feats.append(processed_feats)
# we're done, close output
output.close()
return
def process_filelist_test(filelist=None,model=None,tmpfilename=None,
npicks=None,winsize=None,finaldim=None,K=1,
typecompress='picks'):
"""
Main function, process all files in the list (as long as their artist
is in testartist)
INPUT
filelist - a list of song files
model - h5 file containing feats and year for all train songs
tmpfilename - where to save our processed features
npicks - number of segments to pick per song
winsize - size of each segment we pick
finaldim - how many values do we keep
K - param of KNN (default 1)
typecompress - feature type, 'picks', 'corrcoeff' or 'cov'
must be the same as in training
"""
# sanity check
for arg in locals().values():
assert not arg is None,'process_filelist_test, missing an argument, something still None'
if os.path.isfile(tmpfilename):
print 'ERROR: file',tmpfilename,'already exists.'
return
if not os.path.isfile(model):
print 'ERROR: model',model,'does not exist.'
return
# create kdtree
h5model = tables.openFile(model, mode='r')
assert h5model.root.data.feats.shape[1]==finaldim,'inconsistency in final dim'
kd = ANN.kdtree(h5model.root.data.feats)
# create outputfile
output = tables.openFile(tmpfilename, mode='a')
group = output.createGroup("/",'data','TMP FILE FOR YEAR RECOGNITION')
output.createEArray(group,'year_real',tables.IntAtom(shape=()),(0,),'',
expectedrows=len(filelist))
output.createEArray(group,'year_pred',tables.Float64Atom(shape=()),(0,),'',
expectedrows=len(filelist))
# random projection
ndim = 12 # fixed in this dataset
if typecompress == 'picks':
randproj = RANDPROJ.proj_point5(ndim * winsize, finaldim)
elif typecompress == 'corrcoeff' or typecompress=='cov':
randproj = RANDPROJ.proj_point5(ndim * ndim, finaldim)
elif typecompress == 'avgcov':
randproj = RANDPROJ.proj_point5(90, finaldim)
else:
assert False,'Unknown type of compression: '+str(typecompress)
# go through files
cnt_f = 0
for f in filelist:
cnt_f += 1
if cnt_f % 5000 == 0:
print 'TESTING FILE #'+str(cnt_f)
# check file
h5 = GETTERS.open_h5_file_read(f)
artist_id = GETTERS.get_artist_id(h5)
year = GETTERS.get_year(h5)
track_id = GETTERS.get_track_id(h5)
h5.close()
if year <= 0: # probably useless but...
continue
if typecompress == 'picks':
# we have a train artist with a song year, we're good
bttimbre = get_bttimbre(f)
if bttimbre is None:
continue
# we even have normal features, awesome!
processed_feats = CBTF.extract_and_compress(bttimbre,npicks,winsize,finaldim,
randproj=randproj)
elif typecompress == 'corrcoeff':
h5 = GETTERS.open_h5_file_read(f)
timbres = GETTERS.get_segments_timbre(h5).T
h5.close()
processed_feats = CBTF.corr_and_compress(timbres,finaldim,randproj=randproj)
elif typecompress == 'cov':
h5 = GETTERS.open_h5_file_read(f)
timbres = GETTERS.get_segments_timbre(h5).T
h5.close()
processed_feats = CBTF.cov_and_compress(timbres,finaldim,randproj=randproj)
elif typecompress == 'avgcov':
h5 = GETTERS.open_h5_file_read(f)
timbres = GETTERS.get_segments_timbre(h5).T
h5.close()
processed_feats = CBTF.avgcov_and_compress(timbres,finaldim,randproj=randproj)
else:
assert False,'Unknown type of compression: '+str(typecompress)
if processed_feats is None:
continue
if processed_feats.shape[0] == 0:
continue
# do prediction
year_pred = do_prediction(processed_feats,kd,h5model,K)
# add pred and ground truth to output
if not year_pred is None:
output.root.data.year_real.append( [year] )
output.root.data.year_pred.append( [year_pred] )
# close output and model
del kd
h5model.close()
output.close()
# done
return
