nkululeko 0.94.2__tar.gz → 0.94.3__tar.gz

{nkululeko-0.94.2 → nkululeko-0.94.3}/CHANGELOG.md

@@ -1,6 +1,12 @@
 Changelog
 =========
 
+Version 0.94.3 (25-07-22)
+--------------------------
+* adding the following features (related to dementia/alzheimer):
+* pause_lognorm_mu, pause_lognorm_sigma, pause_lognorm_ks_pvalue
+* pause_mean_duration, pause_std_duration, pause_cv, proportion_pause_duration (
+
 Version 0.94.2 (25-06-02)
 --------------------------
 * added better error message: util.py might not have a logger 

{nkululeko-0.94.2/nkululeko.egg-info → nkululeko-0.94.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nkululeko
-Version: 0.94.2
+Version: 0.94.3
 Summary: Machine learning audio prediction experiments based on templates
 Home-page: https://github.com/felixbur/nkululeko
 Author: Felix Burkhardt

{nkululeko-0.94.2 → nkululeko-0.94.3}/nkululeko/augmenting/resampler.py

@@ -17,7 +17,7 @@ class Resampler:
     def __init__(self, df, replace, not_testing=True):
         self.SAMPLING_RATE = 16000
         self.df = df
-        self.util = Util("resampler", has_config=not_testing)
+        self.util = Util("resampler", has_config=not not_testing)
         self.util.warn(f"all files might be resampled to {self.SAMPLING_RATE}")
         self.not_testing = not_testing
         self.replace = (
@@ -30,7 +30,7 @@ class Resampler:
         files = self.df.index.get_level_values(0).values
         # replace = eval(self.util.config_val("RESAMPLE", "replace", "False"))
         replace = self.replace
-        if self.not_testing:
+        if not self.not_testing:
             store = self.util.get_path("store")
         else:
             store = "./"
@@ -67,17 +67,25 @@ class Resampler:
             self.df = self.df.set_index(
                 self.df.index.set_levels(new_files, level="file")
             )
-            target_file = self.util.config_val("RESAMPLE", "target", "resampled.csv")
-            # remove encoded labels
-            target = self.util.config_val("DATA", "target", "emotion")
-            if "class_label" in self.df.columns:
-                self.df = self.df.drop(columns=[target])
-                self.df = self.df.rename(columns={"class_label": target})
-            # save file
-            self.df.to_csv(target_file)
-            self.util.debug(
-                "saved resampled list of files to" f" {os.path.abspath(target_file)}"
-            )
+            if not self.not_testing:
+                target_file = self.util.config_val("RESAMPLE", "target", "resampled.csv")
+                # remove encoded labels
+                target = self.util.config_val("DATA", "target", "emotion")
+                if "class_label" in self.df.columns:
+                    self.df = self.df.drop(columns=[target])
+                    self.df = self.df.rename(columns={"class_label": target})
+                # save file
+                self.df.to_csv(target_file)
+                self.util.debug(
+                    "saved resampled list of files to" f" {os.path.abspath(target_file)}"
+                )
+            else:
+                # When running from command line, save to simple resampled.csv
+                target_file = "resampled.csv"
+                self.df.to_csv(target_file)
+                self.util.debug(
+                    f"saved resampled list of files to {os.path.abspath(target_file)}"
+                )
         self.util.debug(f"resampled {succes} files, {error} errors")
 
 
@@ -91,7 +99,7 @@ def main():
         df_sample.index, allow_nat=False
     )
     df_sample.head(10)
-    resampler = Resampler(df_sample, not_testing=False)
+    resampler = Resampler(df_sample, False, not_testing=False)
     resampler.resample()
     shutil.copyfile(testfile, "tmp.resample_result.wav")
     shutil.copyfile("tmp.wav", testfile)

{nkululeko-0.94.2 → nkululeko-0.94.3}/nkululeko/constants.py

@@ -1,2 +1,2 @@
-VERSION="0.94.2"
+VERSION="0.94.3"
 SAMPLING_RATE = 16000

{nkululeko-0.94.2 → nkululeko-0.94.3}/nkululeko/models/model.py

@@ -3,11 +3,15 @@ import ast
 import pickle
 import random
 
+from joblib import parallel_backend
 import numpy as np
 import pandas as pd
+from sklearn.model_selection import GridSearchCV
+from sklearn.model_selection import LeaveOneGroupOut
+from sklearn.model_selection import StratifiedKFold
 import sklearn.utils
-from joblib import parallel_backend
-from sklearn.model_selection import GridSearchCV, LeaveOneGroupOut, StratifiedKFold
+
+import audeer
 
 import nkululeko.glob_conf as glob_conf
 from nkululeko.reporting.reporter import Reporter
@@ -301,8 +305,15 @@ class Model:
     def get_type(self):
         return "generic"
 
-    def predict_sample(self, features):
-        """Predict one sample"""
+    def predict_sample(self, features: np.ndarray) -> dict | float:
+        """Predict a single sample using the trained model.
+
+        Args:
+            features (np.ndarray): The feature vector of the sample to predict.
+
+        Returns:
+            dict: A dictionary containing the predicted class probabilities or value.
+        """
         prediction = {}
         if self.util.exp_is_classification():
             # get the class probabilities
@@ -336,3 +347,30 @@ class Model:
         self.set_id(run, epoch)
         with open(path, "rb") as handle:
             self.clf = pickle.load(handle)
+
+    # next function exports the model to onnx
+    def export_onnx(self, onnx_path, input_shape=None):
+        """Export the trained sklearn model to ONNX format.
+
+        Args:
+            onnx_path (str): Path to save the ONNX model.
+            input_shape (tuple, optional): Shape of the input features. If None, inferred from feats_train.
+        """
+        import skl2onnx
+        from skl2onnx import convert_sklearn
+        from skl2onnx.common.data_types import FloatTensorType
+
+        if not hasattr(self, "clf"):
+            self.util.error("No trained model found to export.")
+            return
+
+        if input_shape is None:
+            n_features = self.feats_train.shape[1]
+            initial_type = [("input", FloatTensorType([None, n_features]))]
+        else:
+            initial_type = [("input", FloatTensorType(input_shape))]
+
+        onnx_model = convert_sklearn(self.clf, initial_types=initial_type)
+        with open(audeer.path(onnx_path), "wb") as f:
+            f.write(onnx_model.SerializeToString())
+        self.util.debug(f"Model exported to ONNX at {onnx_path}")

{nkululeko-0.94.2 → nkululeko-0.94.3}/nkululeko/models/model_xgb.py

@@ -1,4 +1,4 @@
-# xgbmodel.py
+# model_xgb.py
 
 from xgboost import XGBClassifier
 

{nkululeko-0.94.2 → nkululeko-0.94.3}/nkululeko/nkululeko.py

@@ -54,6 +54,15 @@ def doit(config_file):
     reports, last_epochs = expr.run()
     result = expr.get_best_report(reports).result.test
     expr.store_report()
+
+    # check if we want to export the model
+    o_path = util.config_val("EXP", "export_onnx", "False")
+    if eval(o_path):
+        print(f"Exporting ONNX model to {o_path}")
+        o_path = o_path.replace('"', '')
+        expr.runmgr.get_best_model().export_onnx(str(o_path))
+
+
     print("DONE")
     return result, int(np.asarray(last_epochs).min())
 

{nkululeko-0.94.2 → nkululeko-0.94.3}/nkululeko/runmanager.py

@@ -181,7 +181,7 @@ class Runmanager:
         """
         # self.load_model(report)
         # report = self.model.predict()
-        self.util.debug(f"plotting conf matrix to {plot_name}")
+        self.util.debug(f"plotting conf matrix as {plot_name}")
         report.plot_confmatrix(plot_name, epoch=report.epoch)
         report.print_results(report.epoch, file_name=plot_name)
         report.print_probabilities(file_name=plot_name)

{nkululeko-0.94.2 → nkululeko-0.94.3/nkululeko.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nkululeko
-Version: 0.94.2
+Version: 0.94.3
 Summary: Machine learning audio prediction experiments based on templates
 Home-page: https://github.com/felixbur/nkululeko
 Author: Felix Burkhardt