def get_data(path_to_dataset):
classes = np.sort([
c for c in os.listdir(path_to_dataset)
if os.path.isdir(os.path.join(path_to_dataset, c))
])
print("Classes: ", classes)
labels = {}
for i in range(len(classes)):
labels[classes[i]] = i
files_lists = len(labels) * [None]
for i, c in enumerate(classes):
files_lists[i], _ = util.list_files(os.path.join(path_to_dataset, c))
Xdatas, Ylabels = [], []
for i, fl in enumerate(files_lists):
label = [0] * len(classes)
label[i] = 1
for f in fl:
print('load file', f)
r_data, sr = librosa.load(f)
data = vggish_input.waveform_to_examples(r_data, sr)
Xdatas.append(data)
Ylabels.append(label)
return Xdatas, Ylabels
def get_crosstalk_factor_map(dir):
# crossTalkCorrectionFactors.17100.txt
print ' - get crosstalk factor map,"%s"' % dir
fileList = utilities.list_files(dir, 'crossTalkCorrectionFactors',indentation=9)
print ' - list: ', fileList
crosstalks = {}
for fileName in fileList:
f = open(fileName)
lines = f.readlines()
runNo = fileName.replace('/', '.').split('.')[-2]
runDes = '0'
if '-' in runNo:
runNo = runNo.split('-')
if verbosity: print ' - ',runNo
runDes = runNo[-1]
runNo = runNo[0]
runNo = int(runNo)
lines = [i.replace('\t', ' ').replace('\n', '').split() for i in lines]
lines = filter(lambda x: len(x) > 1, lines)
corrections = [float(i[1].strip('%')) / 100 for i in lines]
diaCorrection = corrections[-1]
silCorrections = corrections[:-1]
silCor = reduce(lambda x, y: x + y, silCorrections) / len(silCorrections)
silCor2 = reduce(lambda x, y: x + y, map(lambda x: x ** 2, silCorrections)) / len(silCorrections)
sigSil = math.sqrt(silCor2 - silCor * silCor)
sigSil = round(sigSil, 5) * 100.
silCor = round(silCor, 5) * 100.
diaCorrection = round(diaCorrection, 5) * 100.
key = "%s-%s" % (runNo, runDes)
crosstalks[key] = {'meanSil': silCor, 'sigSil': sigSil, 'meanDia': diaCorrection, 'fileName': fileName,
'runDes': runDes}
if verbosity: print ' - ',runNo, crosstalks[key]
return crosstalks
def build_corpus_input_file(directory_root, corpus_output_file_path):
'''
directory_root: the root of files, if there are sub directories under the root, the name of sub directories will be category name
'''
f = TextFileUtil.open_file_write(corpus_output_file_path)
for input_file in list_files(directory_root):
TextFileUtil.write_file(f, "UNK\t" + input_file)
for sub_directory_name, sub_directory_path in list_sub_directories(directory_root):
Corpus._build_corpus_input_file(sub_directory_path, sub_directory_name, f)
f.close()
def build_corpus_input_file(directory_root, corpus_output_file_path):
'''
directory_root: the root of files, if there are sub directories under the root, the name of sub directories will be category name
'''
f = TextFileUtil.open_file_write(corpus_output_file_path)
for input_file in list_files(directory_root):
TextFileUtil.write_file(f, "UNK\t" + input_file)
for sub_directory_name, sub_directory_path in list_sub_directories(
directory_root):
Corpus._build_corpus_input_file(sub_directory_path,
sub_directory_name, f)
f.close()
def GetResults(self):
# get list of files wich starts with 'results'
inputDir = self.main_config.get('Results', 'inputDir')
file_list = utilities.list_files(inputDir, 'results')
print 'found ',len(file_list), 'files'
if self.verbosity:
print ' * staring list', file_list
file_list = [i for i in file_list if i.endswith('.res') and '_new' in i]
print 'reduced to ',len(file_list), 'files'
# print 'updated file list ', self.file_list
print ' * get result config'
results = self.read_result_config(file_list)
return results
def create_tables(self):
if self.update_crosstalk_factors:
print 'update correction factors'
correctionFactor.update_crosstalk_factors(self.config.get('Results', 'inputDir'))
print 'read dictionaries'
mapper = dictCreater.dictCreater(self.config_dir)
self.map = mapper.get_combined_list()
print 'get file list'
# get list of files wich starts with 'results'
file_list = utilities.list_files(self.config.get('Results', 'inputDir'), 'results')
if self.verbosity:
print 'staring list', file_list
self.file_list = [i for i in file_list if i.endswith('.res') and '_new' in i]
# print 'updated file list ', self.file_list
self.results = self.read_result_config()
self.set_csv_mapping()
ascii.write(t, output=fOut, delimiter=' ',
format='fixed_width_two_line')
fOut.close()
if __name__ == '__main__':
indent = " "
lambda_obs = [479.66, 638.26, 776.90, 910.82]
limMagDict = {
'g': 25.2,
'r': 25.4,
'i': 25.1,
'z': 24.9
}
lsDirData = util.list_files("*SN*.DAT", path=args.dirData+os.sep)
"""
KERNEL SPECIFICATION
"""
kern = GPy.kern.RBF(1)
# kern = GPy.kern.RatQuad(1)
"""
----------------------
"""
if args.band not in ['g', 'r', 'i', 'z']:
print 'Band {:<} not recognised! Changing to r'.format(args.band)
def _build_corpus_input_file(cur_directory, category, output_file):
for input_file in list_files(cur_directory):
TextFileUtil.write_file(output_file, category + "\t" + input_file)
def _build_corpus_input_file(cur_directory, category, output_file):
for input_file in list_files(cur_directory):
TextFileUtil.write_file(output_file, category + "\t" + input_file)
