nkululeko 0.89.2__py3-none-any.whl → 0.90.1__py3-none-any.whl

nkululeko/explore.py

@@ -19,6 +19,7 @@ The script can be run from the command line with the `--config` argument to spec
 the configuration file to use. If no configuration file is provided, it will look
 for an `exp.ini` file in the same directory as the script.
 """
+
 # explore.py
 # explore the feature sets
 

nkululeko/export.py

@@ -1,16 +1,18 @@
 # export.py
 # export the loaded training and test sets to it's own folder
 
-import os
-import pandas as pd
+import argparse
 import configparser
+import os
+import shutil
+
 import audeer
-import argparse
 import audiofile
+import pandas as pd
+
+from nkululeko.constants import VERSION
 from nkululeko.experiment import Experiment
 from nkululeko.utils.util import Util
-from nkululeko.constants import VERSION
-import shutil
 
 
 def main(src_dir):

nkululeko/feat_extract/feats_agender.py

@@ -1,16 +1,17 @@
 # feats_agender.py
 
-from nkululeko.feat_extract.featureset import Featureset
 import os
 
 # import pandas as pd
 import audeer
-import nkululeko.glob_conf as glob_conf
+import audinterface
 import audonnx
 import numpy as np
-import audinterface
 import torch
 
+import nkululeko.glob_conf as glob_conf
+from nkululeko.feat_extract.featureset import Featureset
+
 
 class AgenderSet(Featureset):
     """
@@ -37,7 +38,7 @@ class AgenderSet(Featureset):
         cuda = "cuda" if torch.cuda.is_available() else "cpu"
         device = self.util.config_val("MODEL", "device", cuda)
         self.model = audonnx.load(model_root, device=device)
-        self.util.debug(f"initialized agender model")
+        self.util.debug("initialized agender model")
         self.model_loaded = True
 
     def extract(self):

nkululeko/feat_extract/feats_agender_agender.py

@@ -1,14 +1,16 @@
 # feats_audmodel_dim.py
-from nkululeko.feat_extract.featureset import Featureset
 import os
-import pandas as pd
+
 import audeer
-import nkululeko.glob_conf as glob_conf
+import audinterface
 import audonnx
 import numpy as np
-import audinterface
 import torch
 
+import nkululeko.glob_conf as glob_conf
+from nkululeko.feat_extract.featureset import Featureset
+
+
 class Agender_agenderSet(Featureset):
     """
     Age and gender predictions from the wav2vec2. based model finetuned on agender, described in the paper
@@ -29,8 +31,7 @@ class Agender_agenderSet(Featureset):
         if not os.path.isdir(model_root):
             cache_root = audeer.mkdir("cache")
             model_root = audeer.mkdir(model_root)
-            archive_path = audeer.download_url(
-                model_url, cache_root, verbose=True)
+            archive_path = audeer.download_url(model_url, cache_root, verbose=True)
             audeer.extract_archive(archive_path, model_root)
         cuda = "cuda" if torch.cuda.is_available() else "cpu"
         device = self.util.config_val("MODEL", "device", cuda)

nkululeko/feat_extract/feats_analyser.py

@@ -1,18 +1,18 @@
 # feats_analyser.py
 import ast
+
+import matplotlib.pyplot as plt
 import pandas as pd
 from sklearn.inspection import permutation_importance
-from sklearn.linear_model import LogisticRegression
-from sklearn.tree import DecisionTreeClassifier
-from sklearn.linear_model import LinearRegression
-from sklearn.tree import DecisionTreeRegressor
-import matplotlib.pyplot as plt
-from nkululeko.utils.util import Util
-from nkululeko.utils.stats import normalize
-from nkululeko.plots import Plots
+from sklearn.linear_model import LinearRegression, LogisticRegression
+from sklearn.tree import DecisionTreeClassifier, DecisionTreeRegressor
+
 import nkululeko.glob_conf as glob_conf
-from nkululeko.reporting.report_item import ReportItem
+from nkululeko.plots import Plots
 from nkululeko.reporting.defines import Header
+from nkululeko.reporting.report_item import ReportItem
+from nkululeko.utils.stats import normalize
+from nkululeko.utils.util import Util
 
 
 class FeatureAnalyser:
@@ -58,9 +58,9 @@ class FeatureAnalyser:
                 model_func = model.clf.predict
             else:
                 raise Exception("Model not supported for SHAP analysis")
-            
+
             self.util.debug(f"using SHAP explainer for {model_name} model")
-            
+
             explainer = shap.Explainer(
                 model_func,
                 self.features,
@@ -68,14 +68,14 @@ class FeatureAnalyser:
                 algorithm="permutation",
                 npermutations=5,
             )
- 
+
             self.util.debug("computing SHAP values...")
             shap_values = explainer(self.features)
             self.util.to_pickle(shap_values, name)
         else:
             shap_values = self.util.from_pickle(name)
         # plt.figure()
-        plt.close('all')
+        plt.close("all")
         plt.tight_layout()
         shap.plots.bar(shap_values)
         fig_dir = self.util.get_path("fig_dir") + "../"  # one up because of the runs
@@ -272,7 +272,7 @@ class FeatureAnalyser:
         fig = ax.figure
         fig.clear()
         plt.close(fig)
-        caption = f"Feature importance"
+        caption = "Feature importance"
         if permutation:
             caption += " based on permutation of features."
         glob_conf.report.add_item(
@@ -285,7 +285,9 @@ class FeatureAnalyser:
         )
 
         # print feature importance values to file and debug and save to result
-        self.util.debug(f"Importance features from {model_name}: features = \n{df_imp['feats'].values.tolist()}")
+        self.util.debug(
+            f"Importance features from {model_name}: features = \n{df_imp['feats'].values.tolist()}"
+        )
         # result file
         res_dir = self.util.get_path("res_dir")
         filename = f"_EXPL_{model_name}"
@@ -300,7 +302,7 @@ class FeatureAnalyser:
 
         df_imp.to_csv(filename, mode="a")
         self.util.debug(f"Saved feature importance values to {filename}")
-        
+
         # check if feature distributions should be plotted
         plot_feats = self.util.config_val("EXPL", "feature_distributions", False)
         if plot_feats:

nkululeko/feat_extract/feats_ast.py

@@ -4,10 +4,9 @@ import os
 import numpy as np
 import pandas as pd
 import torch
-import torch.nn.functional as F
 import torchaudio
 from tqdm import tqdm
-from transformers import AutoProcessor, ASTModel
+from transformers import ASTModel, AutoProcessor
 
 import nkululeko.glob_conf as glob_conf
 from nkululeko.feat_extract.featureset import Featureset
@@ -34,7 +33,6 @@ class Ast(Featureset):
         self.model.eval()
         self.model_initialized = True
 
-
     def extract(self):
         """Extract the features or load them from disk if present."""
         store = self.util.get_path("store")
@@ -80,11 +78,12 @@ class Ast(Featureset):
                     f"got nan: {self.df.shape} {self.df.isnull().sum().sum()}"
                 )
 
-
     def get_embeddings(self, signal, sampling_rate, file):
         """Extract embeddings from raw audio signal."""
         try:
-            inputs = self.processor(signal.numpy(), sampling_rate=sampling_rate, return_tensors="pt")
+            inputs = self.processor(
+                signal.numpy(), sampling_rate=sampling_rate, return_tensors="pt"
+            )
 
             inputs = {k: v.to(self.device) for k, v in inputs.items()}
 
@@ -102,14 +101,16 @@ class Ast(Featureset):
 
             # print(f"hs shape: {embeddings.shape}")
             # hs shape: (1, 768)
-        
+
         except Exception as e:
-            self.util.error(f"Error extracting embeddings for file {file}: {str(e)}, fill with")
+            self.util.error(
+                f"Error extracting embeddings for file {file}: {str(e)}, fill with"
+            )
             return np.zeros(
                 self.model.config.hidden_size
             )  # Return zero vector on error
         return embeddings.ravel()
-    
+
     def extract_sample(self, signal, sr):
         self.init_model()
         feats = self.get_embeddings(signal, sr, "no file")

nkululeko/feat_extract/feats_auddim.py

@@ -1,16 +1,14 @@
 # feats_audmodel_dim.py
 import os
 
-import numpy as np
-import pandas as pd
-import torch
-
 import audeer
 import audinterface
 import audonnx
+import numpy as np
+import torch
 
-from nkululeko.feat_extract.featureset import Featureset
 import nkululeko.glob_conf as glob_conf
+from nkululeko.feat_extract.featureset import Featureset
 
 
 class AuddimSet(Featureset):

nkululeko/feat_extract/feats_audmodel.py

@@ -4,10 +4,10 @@ import os
 import audeer
 import audinterface
 import audonnx
-import nkululeko.glob_conf as glob_conf
 import numpy as np
-import pandas as pd
 import torch
+
+import nkululeko.glob_conf as glob_conf
 from nkululeko.feat_extract.featureset import Featureset
 
 

nkululeko/feat_extract/feats_clap.py

@@ -1,14 +1,14 @@
 # feats_clap.py
 
-from nkululeko.utils.util import Util
-from nkululeko.feat_extract.featureset import Featureset
 import os
+
+import audiofile
+import laion_clap
 import pandas as pd
 from tqdm import tqdm
-import os
+
 import nkululeko.glob_conf as glob_conf
-import laion_clap
-import audiofile
+from nkululeko.feat_extract.featureset import Featureset
 
 
 class ClapSet(Featureset):
@@ -26,21 +26,19 @@ class ClapSet(Featureset):
         self.util.debug("loading clap model...")
         self.model = laion_clap.CLAP_Module(enable_fusion=False)
         self.model.load_ckpt()  # download the default pretrained checkpoint.
-        print(f"loaded clap model")
+        print("loaded clap model")
 
     def extract(self):
         """Extract the features or load them from disk if present."""
         store = self.util.get_path("store")
         store_format = self.util.config_val("FEATS", "store_format", "pkl")
         storage = f"{store}{self.name}.{store_format}"
-        extract = self.util.config_val(
-            "FEATS", "needs_feature_extraction", False)
+        extract = self.util.config_val("FEATS", "needs_feature_extraction", False)
         no_reuse = eval(self.util.config_val("FEATS", "no_reuse", "False"))
         if extract or no_reuse or not os.path.isfile(storage):
             if not self.model_initialized:
                 self.init_model()
-            self.util.debug(
-                "extracting clap embeddings, this might take a while...")
+            self.util.debug("extracting clap embeddings, this might take a while...")
             emb_series = pd.Series(index=self.data_df.index, dtype=object)
             length = len(self.data_df.index)
             for idx, (file, start, end) in enumerate(
@@ -54,8 +52,7 @@ class ClapSet(Featureset):
                 )
                 emb = self.get_embeddings(signal, sampling_rate)
                 emb_series[idx] = emb
-            self.df = pd.DataFrame(
-                emb_series.values.tolist(), index=self.data_df.index)
+            self.df = pd.DataFrame(emb_series.values.tolist(), index=self.data_df.index)
             self.util.write_store(self.df, storage, store_format)
             try:
                 glob_conf.config["DATA"]["needs_feature_extraction"] = "false"

nkululeko/feat_extract/feats_hubert.py

@@ -10,11 +10,10 @@ import pandas as pd
 import torch
 import torchaudio
 from tqdm import tqdm
-from transformers import HubertModel
-from transformers import Wav2Vec2FeatureExtractor
+from transformers import HubertModel, Wav2Vec2FeatureExtractor
 
-from nkululeko.feat_extract.featureset import Featureset
 import nkululeko.glob_conf as glob_conf
+from nkululeko.feat_extract.featureset import Featureset
 
 
 class Hubert(Featureset):

nkululeko/feat_extract/feats_import.py

@@ -1,10 +1,11 @@
 # feats_import.py
 
-import os
 import ast
+import os
+
 import audformat
 import pandas as pd
-from nkululeko.utils.util import Util
+
 from nkululeko.feat_extract.featureset import Featureset
 
 
@@ -20,11 +21,11 @@ class ImportSet(Featureset):
         try:
             feat_import_files = self.util.config_val("FEATS", "import_file", False)
             feat_import_files = ast.literal_eval(feat_import_files)
-        except ValueError as e:
+        except ValueError:
             self.util.error(
                 "feature type == import needs import_file = ['file1', 'filex']"
             )
-        except SyntaxError as se:
+        except SyntaxError:
             if type(feat_import_files) == str:
                 feat_import_files = [feat_import_files]
             else:

nkululeko/feat_extract/feats_mld.py

@@ -1,11 +1,9 @@
 # mld_fset.py
-from nkululeko.feat_extract.featureset import Featureset
-import sys
 import os
-import pandas as pd
-import numpy as np
-from nkululeko.utils.util import Util
+import sys
+
 import nkululeko.glob_conf as glob_conf
+from nkululeko.feat_extract.featureset import Featureset
 
 
 class MLD_set(Featureset):

nkululeko/feat_extract/feats_mos.py

@@ -12,15 +12,16 @@ pip install --pre torch torchvision torchaudio --extra-index-url https://downloa
 """
 
 import os
+
+import audiofile
 import pandas as pd
-from tqdm import tqdm
 import torch
 import torchaudio
 from torchaudio.pipelines import SQUIM_SUBJECTIVE
 from torchaudio.utils import download_asset
-import audiofile
+from tqdm import tqdm
+
 import nkululeko.glob_conf as glob_conf
-from nkululeko.utils.util import Util
 from nkululeko.feat_extract.featureset import Featureset
 
 

nkululeko/feat_extract/feats_opensmile.py

@@ -1,10 +1,11 @@
 # opensmileset.py
-from nkululeko.feat_extract.featureset import Featureset
 import os
+
+import opensmile
 import pandas as pd
+
 import nkululeko.glob_conf as glob_conf
-import ast
-import opensmile
+from nkululeko.feat_extract.featureset import Featureset
 
 
 class Opensmileset(Featureset):

nkululeko/feat_extract/feats_oxbow.py

@@ -1,10 +1,11 @@
 # feats_oxbow.py
 
-from nkululeko.utils.util import Util
-from nkululeko.feat_extract.featureset import Featureset
 import os
-import pandas as pd
+
 import opensmile
+import pandas as pd
+
+from nkululeko.feat_extract.featureset import Featureset
 
 
 class Openxbow(Featureset):
@@ -53,7 +54,7 @@ class Openxbow(Featureset):
             # check if JAR file exist
             if not os.path.isfile(f"{xbow_path}/openXBOW.jar"):
                 # download using wget if not exist and locate in xbow_path
-                os.system(f"git clone https://github.com/openXBOW/openXBOW")
+                os.system("git clone https://github.com/openXBOW/openXBOW")
             # get the size of the codebook
             size = self.util.config_val("FEATS", "size", 500)
             # get the number of assignements

nkululeko/feat_extract/feats_praat.py

@@ -1,13 +1,12 @@
 # feats_praat.py
-import ast
 import os
 
 import numpy as np
 import pandas as pd
 
+import nkululeko.glob_conf as glob_conf
 from nkululeko.feat_extract import feinberg_praat
 from nkululeko.feat_extract.featureset import Featureset
-import nkululeko.glob_conf as glob_conf
 
 
 class PraatSet(Featureset):
@@ -26,12 +25,10 @@ class PraatSet(Featureset):
         store = self.util.get_path("store")
         store_format = self.util.config_val("FEATS", "store_format", "pkl")
         storage = f"{store}{self.name}.{store_format}"
-        extract = self.util.config_val(
-            "FEATS", "needs_feature_extraction", False)
+        extract = self.util.config_val("FEATS", "needs_feature_extraction", False)
         no_reuse = eval(self.util.config_val("FEATS", "no_reuse", "False"))
         if extract or no_reuse or not os.path.isfile(storage):
-            self.util.debug(
-                "extracting Praat features, this might take a while...")
+            self.util.debug("extracting Praat features, this might take a while...")
             self.df = feinberg_praat.compute_features(self.data_df.index)
             self.df = self.df.set_index(self.data_df.index)
             for i, col in enumerate(self.df.columns):
@@ -54,8 +51,8 @@ class PraatSet(Featureset):
         self.df = self.df.astype(float)
 
     def extract_sample(self, signal, sr):
-        import audiofile
         import audformat
+        import audiofile
 
         tmp_audio_names = ["praat_audio_tmp.wav"]
         audiofile.write(tmp_audio_names[0], signal, sr)

nkululeko/feat_extract/feats_snr.py

@@ -2,6 +2,7 @@
 
 SNR (signal to noise ratio) is extracted as acoustic features.
 """
+
 import os
 
 import audiofile
@@ -11,7 +12,6 @@ from tqdm import tqdm
 import nkululeko.glob_conf as glob_conf
 from nkululeko.autopredict.estimate_snr import SNREstimator
 from nkululeko.feat_extract.featureset import Featureset
-from nkululeko.utils.util import Util
 
 
 class SNRSet(Featureset):
@@ -28,8 +28,7 @@ class SNRSet(Featureset):
         store = self.util.get_path("store")
         store_format = self.util.config_val("FEATS", "store_format", "pkl")
         storage = f"{store}{self.name}.{store_format}"
-        extract = self.util.config_val(
-            "FEATS", "needs_feature_extraction", False)
+        extract = self.util.config_val("FEATS", "needs_feature_extraction", False)
         no_reuse = eval(self.util.config_val("FEATS", "no_reuse", "False"))
         if extract or no_reuse or not os.path.isfile(storage):
             self.util.debug("estimating SNR, this might take a while...")
@@ -46,8 +45,7 @@ class SNRSet(Featureset):
                 snr = self.get_snr(signal[0], sampling_rate)
                 snr_series[idx] = snr
             print("")
-            self.df = pd.DataFrame(
-                snr_series.values.tolist(), index=self.data_df.index)
+            self.df = pd.DataFrame(snr_series.values.tolist(), index=self.data_df.index)
             self.df.columns = ["snr"]
             self.util.write_store(self.df, storage, store_format)
             try:

nkululeko/feat_extract/feats_spectra.py

@@ -6,21 +6,20 @@ Inspired by code from Su Lei
 """
 
 import os
+import pathlib
+
+import audeer
+import matplotlib.pyplot as plt
+import numpy as np
+import pandas as pd
 import torchaudio
 import torchaudio.transforms as T
-import torch
-from torch.utils.data import Dataset
 from PIL import Image, ImageOps
-import pandas as pd
-import numpy as np
-import matplotlib.pyplot as plt
 from tqdm import tqdm
-import pathlib
-import audeer
 
-from nkululeko.feat_extract.featureset import Featureset
-from nkululeko.constants import SAMPLING_RATE
 import nkululeko.glob_conf as glob_conf
+from nkululeko.constants import SAMPLING_RATE
+from nkululeko.feat_extract.featureset import Featureset
 
 
 class Spectraloader(Featureset):

nkululeko/feat_extract/feats_spkrec.py

@@ -7,14 +7,15 @@
 
 import os
 
-import nkululeko.glob_conf as glob_conf
 import pandas as pd
 import torch
 import torchaudio
-from nkululeko.feat_extract.featureset import Featureset
 from speechbrain.inference import EncoderClassifier
 from tqdm import tqdm
 
+import nkululeko.glob_conf as glob_conf
+from nkululeko.feat_extract.featureset import Featureset
+
 # from transformers import HubertModel, Wav2Vec2FeatureExtractor
 
 
@@ -49,16 +50,12 @@ class Spkrec(Featureset):
         """Extract the features or load them from disk if present."""
         store = self.util.get_path("store")
         storage = f"{store}{self.name}.pkl"
-        extract = self.util.config_val(
-            "FEATS", "needs_feature_extraction", False
-        )
+        extract = self.util.config_val("FEATS", "needs_feature_extraction", False)
         no_reuse = eval(self.util.config_val("FEATS", "no_reuse", "False"))
         if extract or no_reuse or not os.path.isfile(storage):
             if not self.classifier_initialized:
                 self.init_model()
-            self.util.debug(
-                "extracting Spkrec embeddings, this might take a while..."
-            )
+            self.util.debug("extracting Spkrec embeddings, this might take a while...")
             emb_series = pd.Series(index=self.data_df.index, dtype=object)
             length = len(self.data_df.index)
             for idx, (file, start, end) in enumerate(
@@ -77,9 +74,7 @@ class Spkrec(Featureset):
                 # fill series with embeddings
                 emb_series.iloc[idx] = emb
             # print(f"emb_series shape: {emb_series.shape}")
-            self.df = pd.DataFrame(
-                emb_series.values.tolist(), index=self.data_df.index
-            )
+            self.df = pd.DataFrame(emb_series.values.tolist(), index=self.data_df.index)
             print(f"df shape: {self.df.shape}")
             self.df.to_pickle(storage)
             try:

nkululeko/feat_extract/feats_squim.py

@@ -14,17 +14,15 @@ pip install --pre torch torchvision torchaudio --extra-index-url https://downloa
 
 import os
 
+import audiofile
 import pandas as pd
 import torch
 import torchaudio
 from torchaudio.pipelines import SQUIM_OBJECTIVE
 from tqdm import tqdm
 
-import audiofile
-
-from nkululeko.feat_extract.featureset import Featureset
 import nkululeko.glob_conf as glob_conf
-from nkululeko.utils.util import Util
+from nkululeko.feat_extract.featureset import Featureset
 
 
 class SquimSet(Featureset):

nkululeko/feat_extract/feats_trill.py

@@ -1,17 +1,14 @@
 # feats_trill.py
 import os
 
+import audiofile as af
 import pandas as pd
 import tensorflow as tf
 import tensorflow_hub as hub
 from tqdm import tqdm
 
-import audiofile as af
-
-from nkululeko.feat_extract.featureset import Featureset
 import nkululeko.glob_conf as glob_conf
-from nkululeko.utils.util import Util
-
+from nkululeko.feat_extract.featureset import Featureset
 
 # Import TF 2.X and make sure we're running eager.
 assert tf.executing_eagerly()

nkululeko/feat_extract/feats_wav2vec2.py

@@ -7,14 +7,16 @@ Currently only supports wav2vec2
 """
 
 import os
-from tqdm import tqdm
+
 import pandas as pd
 import torch
 import torchaudio
-from transformers import Wav2Vec2FeatureExtractor, Wav2Vec2Model
 import transformers
-from nkululeko.feat_extract.featureset import Featureset
+from tqdm import tqdm
+from transformers import Wav2Vec2FeatureExtractor, Wav2Vec2Model
+
 import nkululeko.glob_conf as glob_conf
+from nkululeko.feat_extract.featureset import Featureset
 
 
 class Wav2vec2(Featureset):
@@ -47,7 +49,9 @@ class Wav2vec2(Featureset):
         config.num_hidden_layers = layer_num - hidden_layer
         self.util.debug(f"using hidden layer #{config.num_hidden_layers}")
         self.processor = Wav2Vec2FeatureExtractor.from_pretrained(model_path)
-        self.model = Wav2Vec2Model.from_pretrained(model_path, config=config).to(self.device)
+        self.model = Wav2Vec2Model.from_pretrained(model_path, config=config).to(
+            self.device
+        )
         print(f"intialized Wav2vec model on {self.device}")
         self.model.eval()
         self.model_initialized = True