def test_analyze_directory(soundfiles_dir,call_back, do_long_tests):
if not do_long_tests:
return
kwargs = {'rep': 'mfcc','win_len': 0.025,
'time_step': 0.01, 'num_coeffs': 13,
'freq_lims': (0,8000),'return_rep':True,
'use_multi':True}
scores,reps = analyze_directory(soundfiles_dir, call_back = call_back,**kwargs)
def run(self):
kwargs = self.kwargs
self.results = list()
if kwargs['type'] == 'one':
try:
asim = analyze_directory(kwargs['query'], **kwargs)
except AcousticSimError as e:
self.errorEncountered.emit(e)
return
except Exception as e:
e = PCTPythonError(e)
self.errorEncountered.emit(e)
return
elif kwargs['type'] == 'two':
try:
asim, output_val = acoustic_similarity_directories(
*kwargs['query'], **kwargs)
except AcousticSimError as e:
self.errorEncountered.emit(e)
return
except Exception as e:
e = PCTPythonError(e)
self.errorEncountered.emit(e)
return
#asim[(kwargs['query'][0],kwargs['query'][1])] = output_val
elif kwargs['type'] == 'file':
try:
asim = acoustic_similarity_mapping(kwargs['query'], **kwargs)
except AcousticSimError as e:
self.errorEncountered.emit(e)
return
except Exception as e:
e = PCTPythonError(e)
self.errorEncountered.emit(e)
return
if self.stopped:
return
for k, v in asim.items():
if self.stopped:
return
self.results.append(list(k) + [v])
if kwargs['type'] == 'two':
self.results.append([
os.path.basename(kwargs['query'][0]),
os.path.basename(kwargs['query'][1]), output_val
])
else:
self.results.append(['AVG', 'AVG', sum(asim.values()) / len(asim)])
if self.stopped:
self.finishedCancelling.emit()
return
self.dataReady.emit(self.results)
def test_clustering(test_dir, call_back):
kwargs = {'rep': 'mfcc','win_len': 0.025,
'time_step': 0.01, 'num_coeffs': 13,
'freq_lims': (0,8000),'return_rep':True}
scores,reps = analyze_directory(test_dir, call_back = call_back,**kwargs)
kwargs['cache'] = reps
cn = ClusterNetwork(reps)
cn.cluster(scores,'complete',False)
cn.cluster(scores, 'affinity', False)
cn.cluster(scores, 'affinity', True)
def test_analyze_directory(soundfiles_dir, call_back):
kwargs = {
'rep': 'mfcc',
'win_len': 0.025,
'time_step': 0.01,
'num_coeffs': 13,
'freq_lims': (0, 8000),
'return_rep': True,
'use_multi': True
}
scores, reps = analyze_directory(soundfiles_dir,
call_back=call_back,
**kwargs)
def run(self):
kwargs = self.kwargs
self.results = list()
if kwargs['type'] == 'one':
try:
asim = analyze_directory(kwargs['query'], **kwargs)
except AcousticSimError as e:
self.errorEncountered.emit(e)
return
except Exception as e:
e = PCTPythonError(e)
self.errorEncountered.emit(e)
return
elif kwargs['type'] == 'two':
try:
asim, output_val = acoustic_similarity_directories(*kwargs['query'],**kwargs)
except AcousticSimError as e:
self.errorEncountered.emit(e)
return
except Exception as e:
e = PCTPythonError(e)
self.errorEncountered.emit(e)
return
#asim[(kwargs['query'][0],kwargs['query'][1])] = output_val
elif kwargs['type'] == 'file':
try:
asim = acoustic_similarity_mapping(kwargs['query'], **kwargs)
except AcousticSimError as e:
self.errorEncountered.emit(e)
return
except Exception as e:
e = PCTPythonError(e)
self.errorEncountered.emit(e)
return
if self.stopped:
return
for k,v in asim.items():
if self.stopped:
return
self.results.append(list(k) + [v])
if kwargs['type'] == 'two':
self.results.append([os.path.basename(kwargs['query'][0]),os.path.basename(kwargs['query'][1]), output_val])
else:
self.results.append(['AVG', 'AVG',sum(asim.values())/len(asim)])
if self.stopped:
self.finishedCancelling.emit()
return
self.dataReady.emit(self.results)
def main():
kwargs = {'rep':'mfcc',
'match_function':word_distance,
'use_multi':True,
'num_cores':4,
'return_rep': True,
}
scores, reps = analyze_directory(data_dir, **kwargs)
cn = ClusterNetwork(reps)
cluster_method = 'affinity'
one_cluster = False
cn.cluster(scores,cluster_method,one_cluster)
save_output(cn)
def test_clustering(test_dir, call_back):
kwargs = {
'rep': 'mfcc',
'win_len': 0.025,
'time_step': 0.01,
'num_coeffs': 13,
'freq_lims': (0, 8000),
'return_rep': True
}
scores, reps = analyze_directory(test_dir, call_back=call_back, **kwargs)
kwargs['cache'] = reps
cn = ClusterNetwork(reps)
cn.cluster(scores, 'complete', False)
cn.cluster(scores, 'affinity', False)
cn.cluster(scores, 'affinity', True)
def run(self):
token_path = self.settings['path']
cluster_method = self.settings['cluster_alg']
one_cluster = self.settings['one_cluster']
if self.redo_scores:
self.updateProgress.emit("Recalculating distances...")
kwarg_dict = self.settings.acousticsim_kwarg()
kwarg_dict['cache'] = self.cn.reps
kwarg_dict['return_rep'] = False
scores = analyze_directory(token_path,**kwarg_dict)
else:
scores = None
self.updateProgress.emit("Clustering tokens...")
self.cn.cluster(scores,cluster_method,one_cluster)
self.updateProgress.emit("Ready")
self.dataReady.emit(self.cn)
def find_durations():
kwargs = {'rep':'mfcc',
'match_function':word_distance,
'use_multi':True,
'num_cores':4,
'return_rep': True,
}
scores, reps = analyze_directory(data_dir, **kwargs)
out = list()
with open('output_dur_word.txt', 'w') as f:
writer = csv.writer(f, delimiter = '\t')
writer.writerow(['Label','Duration'])
for v in reps.values():
if 'Exp3' not in v._true_label:
continue
base, _ = os.path.splitext(v._filepath)
one_textgrid = base + '.TextGrid'
one_begin,one_end = get_vowel_points(one_textgrid,
tier_name = 'word', vowel_label = 'word')
dur = one_end - one_begin
writer.writerow([v._true_label,str(dur)])
def run(self):
token_path = self.settings['path']
if not os.path.exists(token_path):
token_path = ''
if not token_path:
return
self.updateProgress.emit("Processing acoustic tokens, please be patient...")
kwarg_dict = self.settings.acousticsim_kwarg()
scores,self.reps = analyze_directory(token_path, call_back = self.updateProgress.emit,**kwarg_dict)
self.updateProgress.emit("Finished acoustic processing! Clustering tokens now...")
cn = ClusterNetwork(self.reps)
cluster_method = self.settings['cluster_alg']
one_cluster = self.settings['one_cluster']
cn.cluster(scores,cluster_method,one_cluster)
self.updateProgress.emit("Ready")
self.dataReady.emit(cn)
def test_analyze_directory(call_back):
kwargs = {'rep': 'mfcc','win_len': 0.025,
'time_step': 0.01, 'num_coeffs': 13,
'freq_lims': (0,8000),'return_rep':True,
'use_multi':True}
scores,reps = analyze_directory(TEST_DIR, call_back = call_back,**kwargs)