nkululeko 0.61.0__tar.gz → 0.62.1__tar.gz

{nkululeko-0.61.0 → nkululeko-0.62.1}/CHANGELOG.md

@@ -1,6 +1,17 @@
 Changelog
 =========
 
+Version 0.62.0
+--------------
+* added min and max_length for silero segmenatation
+
+Version 0.62.0
+--------------
+* fixed segment silero bug
+* added all Wav2vec2 models
+* added resampler module
+* added error on file for embeddings
+
 Version 0.61.0
 --------------
 * added HUBERT embeddings

{nkululeko-0.61.0/nkululeko.egg-info → nkululeko-0.62.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: nkululeko
-Version: 0.61.0
+Version: 0.62.1
 Summary: Machine learning audio prediction experiments based on templates
 Home-page: https://github.com/felixbur/nkululeko
 Author: Felix Burkhardt
@@ -228,6 +228,17 @@ Nkululeko can be used under the [MIT license](https://choosealicense.com/license
 Changelog
 =========
 
+Version 0.62.0
+--------------
+* added min and max_length for silero segmenatation
+
+Version 0.62.0
+--------------
+* fixed segment silero bug
+* added all Wav2vec2 models
+* added resampler module
+* added error on file for embeddings
+
 Version 0.61.0
 --------------
 * added HUBERT embeddings

nkululeko-0.62.1/nkululeko/constants.py

@@ -0,0 +1 @@
+VERSION = '0.62.1'

{nkululeko-0.61.0 → nkululeko-0.62.1}/nkululeko/feature_extractor.py

@@ -50,13 +50,13 @@ class FeatureExtractor:
                 self.featExtractor = TRILLset(
                     f"{store_name}_{self.feats_designation}", self.data_df
                 )
-            elif feats_type == "wav2vec":
+            elif feats_type.startswith("wav2vec2"):
                 from nkululeko.feat_extract.feats_wav2vec2 import Wav2vec2
                 self.featExtractor = Wav2vec2(
-                    f"{store_name}_{self.feats_designation}", self.data_df
+                    f"{store_name}_{self.feats_designation}", self.data_df,
+                    feats_type
                 )
-            elif feats_type in ("hubert", "hubert_ft", "hubert_large", 
-                                "hubert_xlarge", "hubert_xlarge_ft"):
+            elif feats_type.startswith("hubert"):
                 from nkululeko.feat_extract.feats_hubert import Hubert
                 self.featExtractor = Hubert(
                     f"{store_name}_{self.feats_designation}", self.data_df,

nkululeko-0.62.1/nkululeko/resample.py

@@ -0,0 +1,63 @@
+# resample.py
+# change the sampling rate for train and test splits
+
+from nkululeko.experiment import Experiment
+import configparser
+from nkululeko.util import Util
+from  nkululeko.constants import VERSION
+import argparse
+import os
+import pandas as pd
+from nkululeko.augmenting .resampler import Resampler
+
+def main(src_dir):
+    parser = argparse.ArgumentParser(description='Call the nkululeko framework.')
+    parser.add_argument('--config', default='exp.ini', help='The base configuration')
+    args = parser.parse_args()
+    if args.config is not None:
+        config_file = args.config    
+    else:
+        config_file = f'{src_dir}/exp.ini'
+
+    # test if the configuration file exists
+    if not os.path.isfile(config_file):
+        print(f'ERROR: no such file: {config_file}')
+        exit()
+        
+    # load one configuration per experiment
+    config = configparser.ConfigParser()
+    config.read(config_file)
+    # create a new experiment
+    expr = Experiment(config)
+    util = Util('resample')
+    util.debug(f'running {expr.name} from config {config_file}, nkululeko version {VERSION}')
+
+    if util.config_val('EXP', 'no_warnings', False):
+        import warnings
+        warnings.filterwarnings("ignore")
+
+    # load the data
+    expr.load_datasets()
+
+    # split into train and test
+    expr.fill_train_and_tests()
+    util.debug(f'train shape : {expr.df_train.shape}, test shape:{expr.df_test.shape}')
+
+    sample_selection = util.config_val('RESAMPLE', 'sample_selection', 'all')
+    if sample_selection=='all':
+        df = pd.concat([expr.df_train, expr.df_test])
+    elif sample_selection=='train':
+        df = expr.df_train
+    elif sample_selection=='test':
+        df = expr.df_test
+    else:
+        util.error(f'unknown selection specifier {sample_selection}, should be [all | train | test]')
+    util.debug(f'resampling {sample_selection}: {df.shape[0]} samples') 
+    rs = Resampler(df)
+    rs.resample()
+    print('DONE')
+
+
+if __name__ == "__main__":
+    cwd = os.path.dirname(os.path.abspath(__file__))
+    main(cwd) # use this if you want to state the config file path on command line

{nkululeko-0.61.0 → nkululeko-0.62.1}/nkululeko/segment.py

@@ -44,16 +44,6 @@ def main(src_dir):
 
     # segment
     segment_target = util.config_val('SEGMENT', 'target', '_seg')
-    # this if a specific dataset is to be segmented
-    # segment_db = util.config_val('DATA', 'segment', False)
-    # if segment_db:
-    #     for dataset in expr.datasets.keys:
-    #         if segment_db == dataset:
-    #             df = expr.datasets[dataset].df
-    #             util.debug(f'segmenting {dataset}') 
-    #             df_seg = segment_dataframe(df) 
-    #             name = f'{dataset}{segment_target}'
-    #             df_seg.to_csv(f'{expr.data_dir}/{name}.csv')
 
     segmenter = util.config_val('SEGMENT', 'method', 'silero')
     sample_selection = util.config_val('SEGMENT', 'sample_selection', 'all')
@@ -65,8 +55,6 @@ def main(src_dir):
         df = expr.df_test
     else:
         util.error(f'unknown segmentation selection specifier {sample_selection}, should be [all | train | test]')
-    # if "duration" not in df.columns:
-    #     df = df.drop(columns=['duration'], inplace=True)
     util.debug(f'segmenting {sample_selection}: {df.shape[0]} samples with {segmenter}') 
     if segmenter=='silero':
         from nkululeko.segmenting.seg_silero import Silero_segmenter
@@ -86,14 +74,21 @@ def main(src_dir):
         df['duration'] = df.index.to_series().map(lambda x:calc_dur(x)) 
     num_before = df.shape[0]
     num_after = df_seg.shape[0]
-    dataname = '_'.join(expr.datasets.keys())
-    name = f'{dataname}{segment_target}'
-    df_seg.to_csv(f'{expr.data_dir}/{name}.csv')
+    # plot distributions
     from nkululeko.plots import Plots
     plots = Plots()
     plots.plot_durations(df, 'original_durations', sample_selection)
     plots.plot_durations(df_seg, 'segmented_durations', sample_selection)
     print('')
+    # remove encoded labels
+    target = util.config_val('DATA', 'target', 'emotion')
+    if 'class_label' in df_seg.columns:
+        df_seg = df_seg.drop(columns=[target])
+        df_seg = df_seg.rename(columns={'class_label':target})
+    # save file
+    dataname = '_'.join(expr.datasets.keys())
+    name = f'{dataname}{segment_target}'
+    df_seg.to_csv(f'{expr.data_dir}/{name}.csv')
     util.debug(f'saved {name}.csv to {expr.data_dir}, {num_after} samples (was {num_before})')
     print('DONE')
 

{nkululeko-0.61.0 → nkululeko-0.62.1}/nkululeko/util.py

@@ -2,7 +2,6 @@
 import audeer
 import ast 
 import sys
-import nkululeko.glob_conf as glob_conf
 import numpy as np
 import os.path
 import configparser
@@ -11,16 +10,20 @@ import pandas as pd
 
 class Util:
 
+    # a list of words that need not to be warned upon if default values are used
     stopvals = [False, 'False', 'classification', 'png']
 
-    def __init__(self, caller=None):
-        self.got_data_roots = self.config_val('DATA', 'root_folders', False)
-        if self.got_data_roots:
-            # if there is a global data rootfolder file, read from there
-            if not os.path.isfile(self.got_data_roots):
-                self.error(f'no such file: {self.got_data_roots}')
-            self.data_roots = configparser.ConfigParser()
-            self.data_roots.read(self.got_data_roots)
+    def __init__(self, caller=None, has_config=True):
+        if has_config:
+            import nkululeko.glob_conf as glob_conf
+            self.config = glob_conf.config
+            self.got_data_roots = self.config_val('DATA', 'root_folders', False)
+            if self.got_data_roots:
+                # if there is a global data rootfolder file, read from there
+                if not os.path.isfile(self.got_data_roots):
+                    self.error(f'no such file: {self.got_data_roots}')
+                self.data_roots = configparser.ConfigParser()
+                self.data_roots.read(self.got_data_roots)
         if caller is not None:
             self.caller = caller
         else:  
@@ -30,10 +33,10 @@ class Util:
         """
         This method allows the user to get the directory path for the given argument.
         """
-        root = glob_conf.config['EXP']['root']
-        name = glob_conf.config['EXP']['name']
+        root = self.config['EXP']['root']
+        name = self.config['EXP']['name']
         try:
-            entryn = glob_conf.config['EXP'][entry]
+            entryn = self.config['EXP'][entry]
         except KeyError:
             # some default values
             if entry == 'fig_dir':
@@ -61,7 +64,7 @@ class Util:
         If the value is present in the experiment configuration it will be used, else 
         we look in a global file specified by the root_folders value.
         """
-        configuration = glob_conf.config
+        configuration = self.config
         try:
             if len(key)>0:
                 return configuration['DATA'][dataset+'.'+key]
@@ -95,8 +98,8 @@ class Util:
         """
             Get the experiment directory
         """
-        root = glob_conf.config['EXP']['root']
-        name = glob_conf.config['EXP']['name']
+        root = self.config['EXP']['root']
+        name = self.config['EXP']['name']
         dir_name = f'{root}{name}'
         audeer.mkdir(dir_name)
         return dir_name
@@ -124,19 +127,19 @@ class Util:
         """
             Get a string as name from all databases that are useed
         """
-        return '_'.join(ast.literal_eval(glob_conf.config['DATA']['databases']))
+        return '_'.join(ast.literal_eval(self.config['DATA']['databases']))
     
     def get_exp_name(self, only_train = False, only_data = False):
         trains_val = self.config_val('DATA', 'trains', False)
         if only_train and trains_val:
             # try to get only the train tables
-            ds = '_'.join(ast.literal_eval(glob_conf.config['DATA']['trains']))
+            ds = '_'.join(ast.literal_eval(self.config['DATA']['trains']))
         else:
-            ds = '_'.join(ast.literal_eval(glob_conf.config['DATA']['databases']))
+            ds = '_'.join(ast.literal_eval(self.config['DATA']['databases']))
         mt = ''
         if not only_data:
-            mt = f'_{glob_conf.config["MODEL"]["type"]}'
-        ft = '_'.join(ast.literal_eval(glob_conf.config['FEATS']['type']))
+            mt = f'_{self.config["MODEL"]["type"]}'
+        ft = '_'.join(ast.literal_eval(self.config['FEATS']['type']))
         ft += '_'
         set = self.config_val('FEATS', 'set', False)
         set_string = ''
@@ -159,7 +162,7 @@ class Util:
 
     def get_plot_name(self):
         try:
-            plot_name = glob_conf.config['PLOT']['name']
+            plot_name = self.config['PLOT']['name']
         except KeyError:
             plot_name = self.get_exp_name()
         return plot_name
@@ -183,10 +186,10 @@ class Util:
     def set_config_val(self, section, key, value):
         try:
             # does the section already exists?
-            glob_conf.config[section][key] = str(value)
+            self.config[section][key] = str(value)
         except KeyError:
-            glob_conf.config.add_section(section)
-            glob_conf.config[section][key] = str(value)
+            self.config.add_section(section)
+            self.config[section][key] = str(value)
     
     def check_df(self, i, df):
         """Check a dataframe"""
@@ -195,7 +198,7 @@ class Util:
         )
     def config_val(self, section, key, default):
         try:
-            return glob_conf.config[section][key]
+            return self.config[section][key]
         except KeyError:
             if not default in self.stopvals:
                 self.debug(f'value for {key} not found, using default: {default}')
@@ -203,7 +206,7 @@ class Util:
             
     def config_val_list(self, section, key, default):
         try:
-            return ast.literal_eval(glob_conf.config[section][key])
+            return ast.literal_eval(self.config[section][key])
         except KeyError:
             if not default in self.stopvals:
                 self.debug(f'value for {key} not found, using default: {default}')
@@ -213,11 +216,11 @@ class Util:
         # try:
         #     labels = glob_conf.label_encoder.classes_
         # except AttributeError:
-        labels = ast.literal_eval(glob_conf.config['DATA']['labels'])
+        labels = ast.literal_eval(self.config['DATA']['labels'])
         return labels
 
     def continuous_to_categorical(self, array):
-        bins = ast.literal_eval(glob_conf.config['DATA']['bins'])
+        bins = ast.literal_eval(self.config['DATA']['bins'])
         result =  np.digitize(array, bins)-1
         return result
 

{nkululeko-0.61.0 → nkululeko-0.62.1/nkululeko.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: nkululeko
-Version: 0.61.0
+Version: 0.62.1
 Summary: Machine learning audio prediction experiments based on templates
 Home-page: https://github.com/felixbur/nkululeko
 Author: Felix Burkhardt
@@ -228,6 +228,17 @@ Nkululeko can be used under the [MIT license](https://choosealicense.com/license
 Changelog
 =========
 
+Version 0.62.0
+--------------
+* added min and max_length for silero segmenatation
+
+Version 0.62.0
+--------------
+* fixed segment silero bug
+* added all Wav2vec2 models
+* added resampler module
+* added error on file for embeddings
+
 Version 0.61.0
 --------------
 * added HUBERT embeddings

{nkululeko-0.61.0 → nkululeko-0.62.1}/nkululeko.egg-info/SOURCES.txt

@@ -22,6 +22,7 @@ nkululeko/nkululeko.py
 nkululeko/plots.py
 nkululeko/predict.py
 nkululeko/reporter.py
+nkululeko/resample.py
 nkululeko/result.py
 nkululeko/runmanager.py
 nkululeko/scaler.py

nkululeko-0.61.0/nkululeko/constants.py

@@ -1 +0,0 @@
-VERSION = '0.61.0'