nkululeko 0.91.1__tar.gz → 0.91.3__tar.gz

{nkululeko-0.91.1 → nkululeko-0.91.3}/CHANGELOG.md

@@ -1,6 +1,15 @@
 Changelog
 =========
 
+Version 0.91.3
+--------------
+* some additions for robustness
+
+Version 0.91.2
+--------------
+* making lint work by excluding constants from check
+
+Version 0.91.1
 --------------
 * minor refactoring in ensemble module
 

{nkululeko-0.91.1/nkululeko.egg-info → nkululeko-0.91.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: nkululeko
-Version: 0.91.1
+Version: 0.91.3
 Summary: Machine learning audio prediction experiments based on templates
 Home-page: https://github.com/felixbur/nkululeko
 Author: Felix Burkhardt
@@ -200,7 +200,7 @@ All of them take *--config <my_config.ini>* as an argument.
 
 * **nkululeko.nkululeko**: do machine learning experiments combining features and learners
 * **nkululeko.ensemble**: [combine several nkululeko experiments](http://blog.syntheticspeech.de/2024/06/25/nkululeko-ensemble-classifiers-with-late-fusion/) and report on late fusion results
-  * *configurations*: which experiments to combine
+  * *--config*: which experiments (INI files) to combine
   * *--method* (optional): majority_voting, mean (default), max, sum, uncertainty, uncertainty_weighted, confidence_weighted, performance_weighted  
   * *--threshold*: uncertainty threshold (1.0 means no threshold)
   * *--weights*: weights for performance_weighted method (could be from previous UAR, ACC)
@@ -261,7 +261,7 @@ There's my [blog](http://blog.syntheticspeech.de/?s=nkululeko) with tutorials:
 * [Predict new labels for your data from public models and check bias](http://blog.syntheticspeech.de/2023/08/16/nkululeko-how-to-predict-labels-for-your-data-from-existing-models-and-check-them/)
 * [Resample](http://blog.syntheticspeech.de/2023/08/31/how-to-fix-different-sampling-rates-in-a-dataset-with-nkululeko/)
 * [Get some statistics on correlation and effect-size](http://blog.syntheticspeech.de/2023/09/05/nkululeko-get-some-statistics-on-correlation-and-effect-size/)
-* [Automatic generation  of a latex / pdf report](http://blog.syntheticspeech.de/2023/09/26/nkululeko-generate-a-latex-pdf-report/) 
+* [Automatic generation of a latex / pdf report](http://blog.syntheticspeech.de/2023/09/26/nkululeko-generate-a-latex-pdf-report/) 
 * [Inspect your data with Spotlight](http://blog.syntheticspeech.de/2023/10/31/nkululeko-inspect-your-data-with-spotlight/)
 * [Automatically stratify your split sets](http://blog.syntheticspeech.de/2023/11/07/nkululeko-automatically-stratify-your-split-sets/)
 * [re-name data column names](http://blog.syntheticspeech.de/2023/11/16/nkululeko-re-name-data-column-names/)
@@ -355,6 +355,15 @@ F. Burkhardt, Johannes Wagner, Hagen Wierstorf, Florian Eyben and Björn Schulle
 Changelog
 =========
 
+Version 0.91.3
+--------------
+* some additions for robustness
+
+Version 0.91.2
+--------------
+* making lint work by excluding constants from check
+
+Version 0.91.1
 --------------
 * minor refactoring in ensemble module
 

{nkululeko-0.91.1 → nkululeko-0.91.3}/README.md

@@ -157,7 +157,7 @@ All of them take *--config <my_config.ini>* as an argument.
 
 * **nkululeko.nkululeko**: do machine learning experiments combining features and learners
 * **nkululeko.ensemble**: [combine several nkululeko experiments](http://blog.syntheticspeech.de/2024/06/25/nkululeko-ensemble-classifiers-with-late-fusion/) and report on late fusion results
-  * *configurations*: which experiments to combine
+  * *--config*: which experiments (INI files) to combine
   * *--method* (optional): majority_voting, mean (default), max, sum, uncertainty, uncertainty_weighted, confidence_weighted, performance_weighted  
   * *--threshold*: uncertainty threshold (1.0 means no threshold)
   * *--weights*: weights for performance_weighted method (could be from previous UAR, ACC)
@@ -218,7 +218,7 @@ There's my [blog](http://blog.syntheticspeech.de/?s=nkululeko) with tutorials:
 * [Predict new labels for your data from public models and check bias](http://blog.syntheticspeech.de/2023/08/16/nkululeko-how-to-predict-labels-for-your-data-from-existing-models-and-check-them/)
 * [Resample](http://blog.syntheticspeech.de/2023/08/31/how-to-fix-different-sampling-rates-in-a-dataset-with-nkululeko/)
 * [Get some statistics on correlation and effect-size](http://blog.syntheticspeech.de/2023/09/05/nkululeko-get-some-statistics-on-correlation-and-effect-size/)
-* [Automatic generation  of a latex / pdf report](http://blog.syntheticspeech.de/2023/09/26/nkululeko-generate-a-latex-pdf-report/) 
+* [Automatic generation of a latex / pdf report](http://blog.syntheticspeech.de/2023/09/26/nkululeko-generate-a-latex-pdf-report/) 
 * [Inspect your data with Spotlight](http://blog.syntheticspeech.de/2023/10/31/nkululeko-inspect-your-data-with-spotlight/)
 * [Automatically stratify your split sets](http://blog.syntheticspeech.de/2023/11/07/nkululeko-automatically-stratify-your-split-sets/)
 * [re-name data column names](http://blog.syntheticspeech.de/2023/11/16/nkululeko-re-name-data-column-names/)

{nkululeko-0.91.1 → nkululeko-0.91.3}/data/ravdess/process_database.py

@@ -9,14 +9,9 @@ Download and unzip the file Audio_Speech_Actors_01-24.zip e.g., `ravdess_speech`
 
 adapted from https://www.kaggle.com/code/shivamburnwal/speech-emotion-recognition
 
-<<<<<<< HEAD
-Usage: `python3 process_database.py /data/ravdess_speech/`
-
-=======
 Usage: `python3 process_database.py -d /data/ravdess_speech
 OR
 `python3 process_database.py /data/ravdess_speech`
->>>>>>> e639b1104ec3f50b3944440119f2416aea6e31ac
 """
 
  
@@ -25,12 +20,6 @@ import sys
 
 import pandas as pd
 
-<<<<<<< HEAD
-
-# ravdess source directory as argument
-# source_dir = './'
-source_dir = sys.argv[1]
-=======
 import argparse
 from pathlib import Path, PurePath
 
@@ -40,7 +29,6 @@ parser = argparse.ArgumentParser()
 parser.add_argument('-d', '--dir', type=str, default='./', help='path to RAVDESS speech directory')
 args = parser.parse_args()
 source_dir = str(Path(args.dir))
->>>>>>> e639b1104ec3f50b3944440119f2416aea6e31ac
 database_name = 'ravdess'
 
 # check if directory (e.g., Actor_01) exists

nkululeko-0.91.3/nkululeko/autopredict/ap_sid.py

@@ -0,0 +1,43 @@
+""""
+A predictor for sid - Speaker ID.
+"""
+
+from pyannote.audio import Pipeline
+
+
+import numpy as np
+
+import nkululeko.glob_conf as glob_conf
+from nkululeko.feature_extractor import FeatureExtractor
+from nkululeko.utils.util import Util
+
+
+class SIDPredictor:
+    """SIDPredictor.
+
+    predicting speaker id.
+    """
+
+    def __init__(self, df):
+        self.df = df
+        self.util = Util("sidPredictor")
+        self.pipeline = Pipeline.from_pretrained(
+            "pyannote/speaker-diarization-3.1",
+            use_auth_token="HUGGINGFACE_ACCESS_TOKEN_GOES_HERE",
+        )
+
+    def predict(self, split_selection):
+        self.util.debug(f"estimating PESQ for {split_selection} samples")
+        return_df = self.df.copy()
+        feats_name = "_".join(ast.literal_eval(glob_conf.config["DATA"]["databases"]))
+        self.feature_extractor = FeatureExtractor(
+            self.df, ["squim"], feats_name, split_selection
+        )
+        result_df = self.feature_extractor.extract()
+        # replace missing values by 0
+        result_df = result_df.fillna(0)
+        result_df = result_df.replace(np.nan, 0)
+        result_df.replace([np.inf, -np.inf], 0, inplace=True)
+        pred_vals = result_df.pesq * 100
+        return_df["pesq_pred"] = pred_vals.astype("int") / 100
+        return return_df

{nkululeko-0.91.1 → nkululeko-0.91.3}/nkululeko/constants.py

@@ -1,2 +1,2 @@
-VERSION="0.91.1"
+VERSION="0.91.3"
 SAMPLING_RATE = 16000

{nkululeko-0.91.1 → nkululeko-0.91.3}/nkululeko/data/dataset_csv.py

@@ -39,6 +39,8 @@ class Dataset_CSV(Dataset):
             df = audformat.utils.read_csv(data_file)
         if isinstance(df, pd.Series):
             df = df.to_frame()
+        elif isinstance(df, pd.Index):
+            df = pd.DataFrame(index=df)
         rename_cols = self.util.config_val_data(self.name, "colnames", False)
         if rename_cols:
             col_dict = ast.literal_eval(rename_cols)
@@ -78,7 +80,11 @@ class Dataset_CSV(Dataset):
 
         self.df = df
         self.db = None
-        self.got_target = True
+        target = self.util.config_val("DATA", "target", None)
+        if target is not None:
+            self.got_target = True
+        else:
+            self.got_target = False
         self.is_labeled = self.got_target
         self.start_fresh = eval(self.util.config_val("DATA", "no_reuse", "False"))
         is_index = False

{nkululeko-0.91.1 → nkululeko-0.91.3}/nkululeko/experiment.py

@@ -125,7 +125,15 @@ class Experiment:
         # df = pd.read_csv(storage, header=0, index_col=[0,1,2])
         # df.index.set_levels(pd.to_timedelta(df.index.levels[1]), level=1)
         # df.index.set_levels(pd.to_timedelta(df.index.levels[2]), level=2)
-        df = audformat.utils.read_csv(storage)
+        try:
+            df = audformat.utils.read_csv(storage)
+        except ValueError:
+            # split might be empty
+            return pd.DataFrame()
+        if isinstance(df, pd.Series):
+            df = df.to_frame()
+        elif isinstance(df, pd.Index):
+            df = pd.DataFrame(index=df)
         df.is_labeled = True if self.target in df else False
         # print(df.head())
         return df

{nkululeko-0.91.1 → nkululeko-0.91.3}/nkululeko/plots.py

@@ -305,6 +305,9 @@ class Plots:
         except AttributeError as ae:
             self.util.warn(ae)
             ax = sns.histplot(df, x="duration", kde=True)
+        except ValueError as error:
+            self.util.warn(error)
+            ax = sns.histplot(df, x="duration", kde=True)
         min = self.util.to_3_digits(df.duration.min())
         max = self.util.to_3_digits(df.duration.max())
         title = f"Duration distr. for {sample_selection} {df.shape[0]}. min={min}, max={max}"

{nkululeko-0.91.1 → nkululeko-0.91.3}/nkululeko/segment.py

@@ -110,7 +110,7 @@ def main():
     )
     print("")
     # remove encoded labels
-    target = util.config_val("DATA", "target", "emotion")
+    target = util.config_val("DATA", "target", None)
     if "class_label" in df_seg.columns:
         df_seg = df_seg.drop(columns=[target])
         df_seg = df_seg.rename(columns={"class_label": target})

{nkululeko-0.91.1 → nkululeko-0.91.3}/nkululeko/test_pretrain.py

@@ -216,7 +216,6 @@ def doit(config_file):
     )
 
     class Trainer(transformers.Trainer):
-
         def compute_loss(
             self,
             model,

{nkululeko-0.91.1 → nkululeko-0.91.3/nkululeko.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: nkululeko
-Version: 0.91.1
+Version: 0.91.3
 Summary: Machine learning audio prediction experiments based on templates
 Home-page: https://github.com/felixbur/nkululeko
 Author: Felix Burkhardt
@@ -200,7 +200,7 @@ All of them take *--config <my_config.ini>* as an argument.
 
 * **nkululeko.nkululeko**: do machine learning experiments combining features and learners
 * **nkululeko.ensemble**: [combine several nkululeko experiments](http://blog.syntheticspeech.de/2024/06/25/nkululeko-ensemble-classifiers-with-late-fusion/) and report on late fusion results
-  * *configurations*: which experiments to combine
+  * *--config*: which experiments (INI files) to combine
   * *--method* (optional): majority_voting, mean (default), max, sum, uncertainty, uncertainty_weighted, confidence_weighted, performance_weighted  
   * *--threshold*: uncertainty threshold (1.0 means no threshold)
   * *--weights*: weights for performance_weighted method (could be from previous UAR, ACC)
@@ -261,7 +261,7 @@ There's my [blog](http://blog.syntheticspeech.de/?s=nkululeko) with tutorials:
 * [Predict new labels for your data from public models and check bias](http://blog.syntheticspeech.de/2023/08/16/nkululeko-how-to-predict-labels-for-your-data-from-existing-models-and-check-them/)
 * [Resample](http://blog.syntheticspeech.de/2023/08/31/how-to-fix-different-sampling-rates-in-a-dataset-with-nkululeko/)
 * [Get some statistics on correlation and effect-size](http://blog.syntheticspeech.de/2023/09/05/nkululeko-get-some-statistics-on-correlation-and-effect-size/)
-* [Automatic generation  of a latex / pdf report](http://blog.syntheticspeech.de/2023/09/26/nkululeko-generate-a-latex-pdf-report/) 
+* [Automatic generation of a latex / pdf report](http://blog.syntheticspeech.de/2023/09/26/nkululeko-generate-a-latex-pdf-report/) 
 * [Inspect your data with Spotlight](http://blog.syntheticspeech.de/2023/10/31/nkululeko-inspect-your-data-with-spotlight/)
 * [Automatically stratify your split sets](http://blog.syntheticspeech.de/2023/11/07/nkululeko-automatically-stratify-your-split-sets/)
 * [re-name data column names](http://blog.syntheticspeech.de/2023/11/16/nkululeko-re-name-data-column-names/)
@@ -355,6 +355,15 @@ F. Burkhardt, Johannes Wagner, Hagen Wierstorf, Florian Eyben and Björn Schulle
 Changelog
 =========
 
+Version 0.91.3
+--------------
+* some additions for robustness
+
+Version 0.91.2
+--------------
+* making lint work by excluding constants from check
+
+Version 0.91.1
 --------------
 * minor refactoring in ensemble module
 

{nkululeko-0.91.1 → nkululeko-0.91.3}/nkululeko.egg-info/SOURCES.txt

@@ -100,6 +100,7 @@ nkululeko/autopredict/ap_gender.py
 nkululeko/autopredict/ap_mos.py
 nkululeko/autopredict/ap_pesq.py
 nkululeko/autopredict/ap_sdr.py
+nkululeko/autopredict/ap_sid.py
 nkululeko/autopredict/ap_snr.py
 nkululeko/autopredict/ap_stoi.py
 nkululeko/autopredict/ap_valence.py