nkululeko 0.88.9__tar.gz → 0.88.10__tar.gz

{nkululeko-0.88.9 → nkululeko-0.88.10}/CHANGELOG.md

@@ -1,6 +1,11 @@
 Changelog
 =========
 
+Version 0.88.10
+--------------
+* SVM C val defaults to 1
+* fixed agender_agender naming bug
+
 Version 0.88.9
 --------------
 * added performance_weighted ensemble

{nkululeko-0.88.9/nkululeko.egg-info → nkululeko-0.88.10}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: nkululeko
-Version: 0.88.9
+Version: 0.88.10
 Summary: Machine learning audio prediction experiments based on templates
 Home-page: https://github.com/felixbur/nkululeko
 Author: Felix Burkhardt
@@ -204,9 +204,10 @@ 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
-  * *--method* (optional): majority_voting, mean, max, sum, max_class, uncertainty_threshold, uncertainty_weighted, confidence_weighted
+  * *--method* (optional): majority_voting, mean (default), max, sum, uncertainty, uncertainty_weighted, confidence_weighted, performance_weighted  
   * *--threshold*: uncertainty threshold (1.0 means no threshold)
-  * *--outfile* (optional): name of CSV file for output
+  * *--weightes*: weights for performance_weighted method (could be from previous UAR, ACC)
+  * *--outfile* (optional): name of CSV file for output (default: ensemble_result.csv)
   * *--no_labels* (optional): indicate that no ground truth is given 
 * **nkululeko.multidb**: do [multiple experiments](http://blog.syntheticspeech.de/2024/01/02/nkululeko-compare-several-databases/), comparing several databases cross and in itself
 * **nkululeko.demo**: [demo the current best model](http://blog.syntheticspeech.de/2022/01/24/nkululeko-try-out-demo-a-trained-model/) on the command line
@@ -360,6 +361,11 @@ F. Burkhardt, Johannes Wagner, Hagen Wierstorf, Florian Eyben and Björn Schulle
 Changelog
 =========
 
+Version 0.88.10
+--------------
+* SVM C val defaults to 1
+* fixed agender_agender naming bug
+
 Version 0.88.9
 --------------
 * added performance_weighted ensemble

{nkululeko-0.88.9 → nkululeko-0.88.10}/README.md

@@ -160,9 +160,10 @@ 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
-  * *--method* (optional): majority_voting, mean, max, sum, max_class, uncertainty_threshold, uncertainty_weighted, confidence_weighted
+  * *--method* (optional): majority_voting, mean (default), max, sum, uncertainty, uncertainty_weighted, confidence_weighted, performance_weighted  
   * *--threshold*: uncertainty threshold (1.0 means no threshold)
-  * *--outfile* (optional): name of CSV file for output
+  * *--weightes*: weights for performance_weighted method (could be from previous UAR, ACC)
+  * *--outfile* (optional): name of CSV file for output (default: ensemble_result.csv)
   * *--no_labels* (optional): indicate that no ground truth is given 
 * **nkululeko.multidb**: do [multiple experiments](http://blog.syntheticspeech.de/2024/01/02/nkululeko-compare-several-databases/), comparing several databases cross and in itself
 * **nkululeko.demo**: [demo the current best model](http://blog.syntheticspeech.de/2022/01/24/nkululeko-try-out-demo-a-trained-model/) on the command line

{nkululeko-0.88.9 → nkululeko-0.88.10}/nkululeko/constants.py

@@ -1,2 +1,2 @@
-VERSION="0.88.9"
+VERSION="0.88.10"
 SAMPLING_RATE = 16000

{nkululeko-0.88.9 → nkululeko-0.88.10}/nkululeko/explore.py

@@ -1,3 +1,24 @@
+"""
+Explore the feature sets of a machine learning experiment.
+
+This script is the entry point for the 'explore' module of the nkululeko framework.
+It handles loading the experiment configuration, setting up the experiment, and
+running various feature exploration techniques based on the configuration.
+
+The script supports the following configuration options:
+- `no_warnings`: If set to `True`, it will ignore all warnings during the exploration.
+- `feature_distributions`: If set to `True`, it will generate plots of the feature distributions.
+- `tsne`: If set to `True`, it will generate a t-SNE plot of the feature space.
+- `scatter`: If set to `True`, it will generate a scatter plot of the feature space.
+- `spotlight`: If set to `True`, it will generate a 'spotlight' plot of the feature space.
+- `shap`: If set to `True`, it will generate SHAP feature importance plots.
+- `model`: The type of model to use for the feature exploration (e.g. 'SVM').
+- `plot_tree`: If set to `True`, it will generate a decision tree plot.
+
+The script can be run from the command line with the `--config` argument to specify
+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-0.88.9 → nkululeko-0.88.10}/nkululeko/feat_extract/feats_agender_agender.py

@@ -7,18 +7,19 @@ import nkululeko.glob_conf as glob_conf
 import audonnx
 import numpy as np
 import audinterface
+import torch
 
-
-class AgenderAgenderSet(Featureset):
+class Agender_agenderSet(Featureset):
     """
     Age and gender predictions from the wav2vec2. based model finetuned on agender, described in the paper
     "Speech-based Age and Gender Prediction with Transformers"
     https://arxiv.org/abs/2306.16962
     """
 
-    def __init__(self, name, data_df):
-        super().__init__(name, data_df)
+    def __init__(self, name, data_df, feats_type):
+        super().__init__(name, data_df, feats_type)
         self.model_loaded = False
+        self.feats_type = feats_type
 
     def _load_model(self):
         model_url = "https://zenodo.org/record/7761387/files/w2v2-L-robust-6-age-gender.25c844af-1.1.1.zip"

{nkululeko-0.88.9 → nkululeko-0.88.10}/nkululeko/feat_extract/feats_spkrec.py

@@ -12,7 +12,7 @@ import pandas as pd
 import torch
 import torchaudio
 from nkululeko.feat_extract.featureset import Featureset
-from speechbrain.pretrained import EncoderClassifier
+from speechbrain.inference import EncoderClassifier
 from tqdm import tqdm
 
 # from transformers import HubertModel, Wav2Vec2FeatureExtractor

nkululeko-0.88.10/nkululeko/models/model_gmm.py

@@ -0,0 +1,37 @@
+# model_gmm.py
+
+from sklearn import mixture
+from nkululeko.models.model import Model
+import pandas as pd
+
+class GMM_model(Model):
+    """An GMM model"""
+
+    is_classifier = True
+
+    def __init__(self, df_train, df_test, feats_train, feats_test):
+        super().__init__(df_train, df_test, feats_train, feats_test)
+        self.name = "gmm"
+        self.n_components = int(self.util.config_val("MODEL", "GMM_components", "4"))
+        covariance_type = self.util.config_val("MODEL", "GMM_covariance_type", "full")
+        self.clf = mixture.GaussianMixture(
+            n_components=self.n_components, 
+            covariance_type=covariance_type,
+            random_state = 42,
+        )
+        # set up the classifier
+
+    def get_predictions(self):
+        """Use the predict_proba method of the GaussianMixture model to get 
+        probabilities. Create a DataFrame with these probabilities and return 
+        it along with the predictions."""
+        probs = self.clf.predict_proba(self.feats_test)
+        preds = self.clf.predict(self.feats_test)
+        
+        # Convert predictions to a list
+        preds = preds.tolist()
+        
+        # Create a DataFrame for probabilities
+        proba_df = pd.DataFrame(probs, index=self.feats_test.index, columns=range(self.n_components))
+        
+        return preds, proba_df

{nkululeko-0.88.9 → nkululeko-0.88.10}/nkululeko/multidb.py

@@ -1,3 +1,10 @@
+"""
+Demonstrates the usage of the ML-experiment framework for the nkululeko MULTIDB project.
+
+The `main` function is the entry point of the script, which parses command-line arguments, reads a configuration file, and runs the nkululeko or aug_train functions based on the configuration.
+
+The `plot_heatmap` function generates a heatmap plot of the results and saves it to a file, along with some summary statistics.
+"""
 # main.py
 # Demonstration code to use the ML-experiment framework
 

{nkululeko-0.88.9 → nkululeko-0.88.10}/nkululeko/plots.py

@@ -4,14 +4,14 @@ import ast
 import matplotlib.pyplot as plt
 import numpy as np
 import pandas as pd
-from scipy import stats
 import seaborn as sns
+from scipy import stats
 from sklearn.manifold import TSNE
 
 import nkululeko.glob_conf as glob_conf
+import nkululeko.utils.stats as su
 from nkululeko.reporting.defines import Header
 from nkululeko.reporting.report_item import ReportItem
-import nkululeko.utils.stats as su
 from nkululeko.utils.util import Util
 
 
@@ -32,9 +32,9 @@ class Plots:
         # plot the distribution of samples per speaker
         # one up because of the runs
         fig_dir = self.util.get_path("fig_dir") + "../"
-        self.util.debug(f"plotting samples per speaker")
+        self.util.debug("plotting samples per speaker")
         if "gender" in df_speakers:
-            filename = f"samples_value_counts"
+            filename = "samples_value_counts"
             ax = (
                 df_speakers.groupby("samplenum")["gender"]
                 .value_counts()
@@ -46,7 +46,7 @@ class Plots:
                     rot=0,
                 )
             )
-            ax.set_ylabel(f"number of speakers")
+            ax.set_ylabel("number of speakers")
             ax.set_xlabel("number of samples")
             self.save_plot(
                 ax,
@@ -58,7 +58,7 @@ class Plots:
 
             # fig.clear()
         else:
-            filename = f"samples_value_counts"
+            filename = "samples_value_counts"
             ax = (
                 df_speakers["samplenum"]
                 .value_counts()
@@ -265,7 +265,8 @@ class Plots:
         """Plot relation of categorical distribution with continuous."""
         dist_type = self.util.config_val("EXPL", "dist_type", "hist")
         cats, cat_str, es = su.get_effect_size(df, cat_col, cont_col)
-        if dist_type == "hist":
+        model_type = self.util.get_model_type()
+        if dist_type == "hist" and model_type != "tree":
             ax = sns.histplot(df, x=cont_col, hue=cat_col, kde=True)
             caption = f"{ylab} {df.shape[0]}. {cat_str} ({cats}):" f" {es}"
             ax.set_title(caption)
@@ -489,7 +490,7 @@ class Plots:
         glob_conf.report.add_item(
             ReportItem(
                 Header.HEADER_EXPLORE,
-                f"Scatter plot",
+                "Scatter plot",
                 f"using {dimred_type}",
                 filename,
             )
@@ -561,8 +562,8 @@ class Plots:
         glob_conf.report.add_item(
             ReportItem(
                 Header.HEADER_EXPLORE,
-                f"Tree plot",
-                f"for feature importance",
+                "Tree plot",
+                "for feature importance",
                 filename,
             )
         )

{nkululeko-0.88.9 → nkululeko-0.88.10/nkululeko.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: nkululeko
-Version: 0.88.9
+Version: 0.88.10
 Summary: Machine learning audio prediction experiments based on templates
 Home-page: https://github.com/felixbur/nkululeko
 Author: Felix Burkhardt
@@ -204,9 +204,10 @@ 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
-  * *--method* (optional): majority_voting, mean, max, sum, max_class, uncertainty_threshold, uncertainty_weighted, confidence_weighted
+  * *--method* (optional): majority_voting, mean (default), max, sum, uncertainty, uncertainty_weighted, confidence_weighted, performance_weighted  
   * *--threshold*: uncertainty threshold (1.0 means no threshold)
-  * *--outfile* (optional): name of CSV file for output
+  * *--weightes*: weights for performance_weighted method (could be from previous UAR, ACC)
+  * *--outfile* (optional): name of CSV file for output (default: ensemble_result.csv)
   * *--no_labels* (optional): indicate that no ground truth is given 
 * **nkululeko.multidb**: do [multiple experiments](http://blog.syntheticspeech.de/2024/01/02/nkululeko-compare-several-databases/), comparing several databases cross and in itself
 * **nkululeko.demo**: [demo the current best model](http://blog.syntheticspeech.de/2022/01/24/nkululeko-try-out-demo-a-trained-model/) on the command line
@@ -360,6 +361,11 @@ F. Burkhardt, Johannes Wagner, Hagen Wierstorf, Florian Eyben and Björn Schulle
 Changelog
 =========
 
+Version 0.88.10
+--------------
+* SVM C val defaults to 1
+* fixed agender_agender naming bug
+
 Version 0.88.9
 --------------
 * added performance_weighted ensemble

nkululeko-0.88.9/nkululeko/models/model_gmm.py

@@ -1,20 +0,0 @@
-# model_gmm.py
-
-from sklearn import mixture
-from nkululeko.models.model import Model
-
-
-class GMM_model(Model):
-    """An GMM model"""
-
-    is_classifier = True
-
-    def __init__(self, df_train, df_test, feats_train, feats_test):
-        super().__init__(df_train, df_test, feats_train, feats_test)
-        self.name = "gmm"
-        n_components = int(self.util.config_val("MODEL", "GMM_components", "4"))
-        covariance_type = self.util.config_val("MODEL", "GMM_covariance_type", "full")
-        self.clf = mixture.GaussianMixture(
-            n_components=n_components, covariance_type=covariance_type
-        )
-        # set up the classifier