def process_directory(self, input_directory: str, output_directory: str):
"""
Run process to the input directory where audio files exist and save processed audio files
:param input_directory: Path to input audio file
:param output_directory: Output directory
"""
# Make directory if it does not exist
if not os.path.isdir(output_directory):
os.mkdir(output_directory)
# Extract file names in the input directory
file_names = FileUtil.get_file_names(input_directory)
# Apply process to audio files in the input directory
for audio_file in file_names:
# Process one audio file
try:
processed_audio = self.process_audio_file(
os.path.join(input_directory, audio_file),
self.audio_length, self.sampling_rate)
# Rename file
file_path, extension = os.path.splitext(audio_file)
# Save it
self.save_audio_file(
processed_audio,
os.path.join(output_directory, file_path + ".wav"),
self.sampling_rate, self.normalize)
except Exception as err:
print(f"Skipping too short audio: {err}")
def extract_dataset(self, dataset_path: str):
"""
Feature extraction to dataset
Extract time series feature as 2D pandas dataframe and 3D numpy array, as well as label vector as list
:param dataset_path: path to dataset
:return directory_files_feature_dict: dictionary of extracted features from all audio files in dataset folder
{key: name of directory, value: list of file names {key: file name, value: list of extracted features}}
:return label_list: list of numerical label vector
"""
# Make label
label_list = self.make_label_from_directory(dataset_path)
# Get file names and store them into a dictionary
directory_files_dict = {}
for directory in FileUtil.get_folder_names(dataset_path, sort=True):
directory_files_dict[directory] = FileUtil.get_file_names(
os.path.join(dataset_path, directory))
# Extract all features and store them into list
directory_files_feature_dict = {}
for directory, audio_files in tqdm(directory_files_dict.items()):
# Apply feature extraction to one directory
directory_files_feature_dict[directory] = self.extract_directory(
os.path.join(dataset_path, directory))
return directory_files_feature_dict, label_list
def folder2label(input_data_directory: str) -> list:
"""
# Get unique values in one column
:param input_data_directory: input data directory where data folder with the label name exist
:return list of folder names
"""
return FileUtil.get_file_names(input_data_directory)
def make_label_from_directory(dataset_path: str):
# Init parameter
dir_num = 0
label_list = []
# Iterate over directories
for directory in FileUtil.get_folder_names(dataset_path, sort=True):
# Make label as list
label_list.extend([dir_num] * len(FileUtil.get_file_names(os.path.join(dataset_path, directory))))
dir_num += 1
return label_list
def extract_dataset(self, dataset_path: str, stats_type: str):
"""
Feature extraction to dataset
Extract time series feature as 2D pandas dataframe and 3D numpy array, as well as label vector as list
:param dataset_path: path to dataset
:param stats_type: type of statistics for 2D feature
:return all_2d_dataframe: 2D feature pandas dataframe across all frames
:return all_3d_array: 3D feature numpy array across all frames
:return label_list: list of numerical label vector
"""
# Get folder names under data set path
directory_names = FileUtil.get_folder_names(dataset_path, sort=True)
# Get file names and store them into a dictionary
directory_files_dict = {}
for directory in directory_names:
directory_files_dict[directory] = FileUtil.get_file_names(os.path.join(dataset_path, directory))
# Extract all features and store them into list
all_2d_dataframe = pd.DataFrame()
dir_num = 0
label_list = []
for directory, audio_files in directory_files_dict.items():
# Apply feature extraction to a directory
file_feature_stat_dict, class_3d_feature = self.extract_directory(os.path.join(dataset_path, directory), stats_type)
# Convert dictionary to data frame
class_2d_dataframe = DataProcess.dict2dataframe(file_feature_stat_dict, segment_feature=True)
# Add label to 2D feature data frame
class_2d_dataframe_with_label = DataProcess.add_label(class_2d_dataframe, directory)
# Combine 2D feature data frame
all_2d_dataframe = all_2d_dataframe.append(class_2d_dataframe_with_label)
# Append 3D arrays
if dir_num == 0:
all_3d_array = class_3d_feature
else:
all_3d_array = np.dstack((all_3d_array, class_3d_feature))
# Make label as list
a = [dir_num] * len(audio_files)
label_list.extend(a)
dir_num += 1
# Transpose 3D array
all_3d_array = all_3d_array.T
return all_2d_dataframe, all_3d_array, label_list
def process_directory(self, input_directory: str, output_directory: str):
"""
Run process to the input directory where audio files exist and save processed audio files
:param input_directory: Path to input audio file
:param output_directory: Output directory
"""
# Make directory if it does not exist
if not os.path.isdir(output_directory):
os.mkdir(output_directory)
# Extract file names in the input directory
file_names = FileUtil.get_file_names(input_directory)
# Apply process to audio files in the input directory
for audio_file in file_names:
# Process one audio file
processed_audio = self.process_audio_file(os.path.join(input_directory, audio_file), self.audio_length, self.sampling_rate)
# Save it
self.save_audio_file(processed_audio, os.path.join(output_directory, audio_file), self.sampling_rate, self.normalize)
def extract_directory(self, input_directory: str):
"""
Feature extraction to a folder which contains audio files
:param input_directory: folder name which has audio files
:return dictionary of extracted features from audio file
{key: name of files, value: list of extracted features}
"""
# Extract file names in the input directory
file_names = FileUtil.get_file_names(input_directory)
# Extract features from audio files in a directory
# file_feature_stat_dict = {}
file_feature_dict = {}
start = time.time()
# Extract each audio file
for count, audio_file in tqdm(enumerate(file_names)):
# Extract features from one audio file
file_feature_dict[audio_file] = self.extract_file(os.path.join(input_directory, audio_file))
print("Extracted {0} with {1} \n".format(input_directory, time.time() - start))
return file_feature_dict
def extract_directory(self, input_directory: str, stats_type: str):
"""
Feature extraction to a folder which contains audio files
:param input_directory: folder name which has audio files
:param stats_type: type of statistics
:return dictionary of extracted features from audio file
{key: name of feature, value: list of array(number of frames)}
:return directory_short_feature_list: short-term features extracted from all audio files from one directory
"""
# Extract file names in the input directory
file_names = FileUtil.get_file_names(input_directory)
# Extract features from audio files in a directory
file_feature_stat_dict = {}
start = time.time()
# Extract each audio file
for count, audio_file in enumerate(file_names):
# Extract features from one audio file
frame_extracted_feature, file_short_feature_list = self.extract_file(os.path.join(input_directory, audio_file))
# Append short-term feature to 3D array
if count == 0:
directory_3d_feature = np.array(file_short_feature_list).T
print("Data structure is {}".format(directory_3d_feature.shape))
else:
directory_3d_feature = np.dstack((directory_3d_feature, np.array(file_short_feature_list).T))
print("Data structure is {}".format(directory_3d_feature.shape))
# Take stats across frames
file_feature_stat_dict[audio_file] = self.get_feature_stats(frame_extracted_feature, stats_type)
end = time.time()
print("Extracted {0} with {1} \n".format(input_directory, end - start))
return file_feature_stat_dict, directory_3d_feature