nkululeko 0.90.4__py3-none-any.whl → 0.91.1__py3-none-any.whl

nkululeko/constants.py

@@ -1,2 +1,2 @@
-VERSION="0.90.4"
+VERSION="0.91.1"
 SAMPLING_RATE = 16000

nkululeko/ensemble.py

@@ -1,3 +1,4 @@
+#!/usr/bin/env python3
 """
 Ensemble predictions from multiple experiments.
 
@@ -16,9 +17,6 @@ Raises:
     AssertionError: If the number of config files is less than 2 for majority voting.
 """
 
-#!/usr/bin/env python
-# -*- coding: utf-8 -*-
-
 
 import configparser
 import time
@@ -300,10 +298,10 @@ def ensemble_predictions(
     return ensemble_preds
 
 
-def main(src_dir: Path) -> None:
+def main():
     parser = ArgumentParser()
     parser.add_argument(
-        "configs",
+        "--config",
         nargs="+",
         help="Paths to the configuration files of the experiments to ensemble. \
              Can be INI files for Nkululeko.nkululeko or CSV files from Nkululeko.demo.",
@@ -356,7 +354,7 @@ def main(src_dir: Path) -> None:
     start = time.time()
 
     ensemble_preds = ensemble_predictions(
-        args.configs, args.method, args.threshold, args.weights, args.no_labels
+        args.config, args.method, args.threshold, args.weights, args.no_labels
     )
 
     # save to csv
@@ -368,5 +366,4 @@ def main(src_dir: Path) -> None:
 
 
 if __name__ == "__main__":
-    cwd = Path(__file__).parent
-    main(cwd)
+    main()

nkululeko/segment.py

@@ -1,5 +1,22 @@
-# segment.py
-# segment data splits
+"""
+Segments the samples in the dataset into chunks based on voice activity detection using SILERO VAD [1].
+
+The segmentation results are saved to a file, and the distributions of the original and
+segmented durations are plotted.
+
+The module also handles configuration options, such as the segmentation method and sample
+selection, and reports the segmentation results.
+
+Usage:
+    python3 -m nkululeko.segment [--config CONFIG_FILE]
+
+Example:
+    nkululeko.segment --config tests/exp_androids_segment.ini
+
+References:
+    [1] https://github.com/snakers4/silero-vad
+    
+"""
 
 import argparse
 import configparser
@@ -7,9 +24,9 @@ import os
 
 import pandas as pd
 
-import nkululeko.glob_conf as glob_conf
 from nkululeko.constants import VERSION
 from nkululeko.experiment import Experiment
+import nkululeko.glob_conf as glob_conf
 from nkululeko.reporting.report_item import ReportItem
 from nkululeko.utils.util import Util
 
@@ -78,6 +95,7 @@ def main():
 
     if "duration" not in df.columns:
         df["duration"] = df.index.to_series().map(lambda x: calc_dur(x))
+    df_seg["duration"] = df_seg.index.to_series().map(lambda x: calc_dur(x))
     num_before = df.shape[0]
     num_after = df_seg.shape[0]
     # plot distributions
@@ -115,36 +133,5 @@ def main():
     print("DONE")
 
 
-def get_segmentation(file):
-    #    print(f'segmenting {file[0]}')
-    print(".", end="")
-    wav = read_audio(file[0], sampling_rate=SAMPLING_RATE)
-    speech_timestamps = get_speech_timestamps(
-        wav, vad_model, sampling_rate=SAMPLING_RATE
-    )
-    files, starts, ends = [], [], []
-    for entry in speech_timestamps:
-        start = float(entry["start"] / 1000.0)
-        end = float(entry["end"] / 1000.0)
-        files.append(file[0])
-        starts.append(start)
-        ends.append(end)
-    seg_index = segmented_index(files, starts, ends)
-    return seg_index
-
-
-def segment_dataframe(df):
-    dfs = []
-    for file, values in df.iterrows():
-        index = get_segmentation(file)
-        dfs.append(
-            pd.DataFrame(
-                values.to_dict(),
-                index,
-            )
-        )
-    return audformat.utils.concat(dfs)
-
-
 if __name__ == "__main__":
     main()  # use this if you want to state the config file path on command line

nkululeko/segmenting/seg_silero.py

@@ -32,8 +32,30 @@ class Silero_segmenter:
         self.no_testing = not_testing
         self.util = Util(has_config=not_testing)
 
-    def get_segmentation(self, file):
-        #    print(f'segmenting {file[0]}')
+    def get_segmentation_simple(self, file):
+        (
+            get_speech_timestamps,
+            save_audio,
+            read_audio,
+            VADIterator,
+            collect_chunks,
+        ) = vad_utils
+        SAMPLING_RATE = 16000
+        wav = read_audio(file[0], sampling_rate=SAMPLING_RATE)
+        speech_timestamps = get_speech_timestamps(
+            wav, vad_model, sampling_rate=SAMPLING_RATE
+        )
+        files, starts, ends = [], [], []
+        for entry in speech_timestamps:
+            start = float(entry["start"] / SAMPLING_RATE)
+            end = float(entry["end"] / SAMPLING_RATE)
+            files.append(file[0])
+            starts.append(start)
+            ends.append(end)
+        seg_index = segmented_index(files, starts, ends)
+        return seg_index
+
+    def get_segmentation(self, file, min_length, max_length):
         (
             get_speech_timestamps,
             save_audio,
@@ -42,12 +64,6 @@ class Silero_segmenter:
             collect_chunks,
         ) = vad_utils
         SAMPLING_RATE = 16000
-        if self.no_testing:
-            min_length = float(self.util.config_val("SEGMENT", "min_length", 2))
-            max_length = float(self.util.config_val("SEGMENT", "max_length", 10))
-        else:
-            min_length = 2
-            max_length = 10
         wav = read_audio(file[0], sampling_rate=SAMPLING_RATE)
         speech_timestamps = get_speech_timestamps(
             wav, vad_model, sampling_rate=SAMPLING_RATE
@@ -76,8 +92,18 @@ class Silero_segmenter:
 
     def segment_dataframe(self, df):
         dfs = []
+        max_length = eval(self.util.config_val("SEGMENT", "max_length", "False"))
+        if max_length:
+            if self.no_testing:
+                min_length = float(self.util.config_val("SEGMENT", "min_length", 2))
+            else:
+                min_length = 2
+            self.util.debug(f"segmenting with max length: {max_length+min_length}")
         for file, values in tqdm(df.iterrows()):
-            index = self.get_segmentation(file)
+            if max_length:
+                index = self.get_segmentation(file, min_length, max_length)
+            else:
+                index = self.get_segmentation_simple(file)
             dfs.append(
                 pd.DataFrame(
                     values.to_dict(),

{nkululeko-0.90.4.dist-info → nkululeko-0.91.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: nkululeko
-Version: 0.90.4
+Version: 0.91.1
 Summary: Machine learning audio prediction experiments based on templates
 Home-page: https://github.com/felixbur/nkululeko
 Author: Felix Burkhardt
@@ -10,7 +10,7 @@ Classifier: License :: OSI Approved :: MIT License
 Classifier: Operating System :: OS Independent
 Classifier: Development Status :: 3 - Alpha
 Classifier: Topic :: Scientific/Engineering
-Requires-Python: >=3.6
+Requires-Python: >=3.9
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: audeer
@@ -23,7 +23,6 @@ Requires-Dist: audonnx
 Requires-Dist: confidence-intervals
 Requires-Dist: datasets
 Requires-Dist: imageio
-Requires-Dist: laion-clap
 Requires-Dist: matplotlib
 Requires-Dist: numpy
 Requires-Dist: opensmile
@@ -356,6 +355,14 @@ F. Burkhardt, Johannes Wagner, Hagen Wierstorf, Florian Eyben and Björn Schulle
 Changelog
 =========
 
+--------------
+* minor refactoring in ensemble module
+
+Version 0.91.0
+--------------
+* fixed duration display in segmentation
+* added possibility to use original segmentations (without max. duration)
+
 Version 0.90.4
 --------------
 * added plot format for multidb

{nkululeko-0.90.4.dist-info → nkululeko-0.91.1.dist-info}/RECORD

@@ -2,12 +2,12 @@ nkululeko/__init__.py,sha256=62f8HiEzJ8rG2QlTFJXUCMpvuH3fKI33DoJSj33mscc,63
 nkululeko/aug_train.py,sha256=FoMbBrfyOZd4QAw7oIHl3X6-UpsqAKWVDIolCA7qOWs,3196
 nkululeko/augment.py,sha256=3RzaxB3gRxovgJVjHXi0glprW01J7RaHhUkqotW2T3U,2955
 nkululeko/cacheddataset.py,sha256=XFpWZmbJRg0pvhnIgYf0TkclxllD-Fctu-Ol0PF_00c,969
-nkululeko/constants.py,sha256=jZ8xPXzwC4olxRWBxh7QNAfDpWxH99Bim1eoRIcVwtY,39
+nkululeko/constants.py,sha256=iYlIApfoK9ylYhWWe4mizk150FEYR7mV6iaGEZJ9NNU,39
 nkululeko/demo-ft.py,sha256=iD9Pzp9QjyAv31q1cDZ75vPez7Ve8A4Cfukv5yfZdrQ,770
 nkululeko/demo.py,sha256=4Yzhg6pCPBYPGJrP7JX2TysVosl_R1llpVDKc2P_gUA,4955
 nkululeko/demo_feats.py,sha256=BvZjeNFTlERIRlq34OHM4Z96jdDQAhB01BGQAUcX9dM,2026
 nkululeko/demo_predictor.py,sha256=lDF-xOxRdEAclOmbepAYg-BQXQdGkHfq2n74PTIoop8,4872
-nkululeko/ensemble.py,sha256=QONr-1VwMr2D0I7wjWxwGjtYzWf4v9DoI3C-fFnar7E,12862
+nkululeko/ensemble.py,sha256=71V-rre61H3J4sh7lu-OTo4I2_g7mm_rQxwW1ARDHgY,12782
 nkululeko/experiment.py,sha256=octx5S4Y8-gAD0dXCRb6DFZwsXTYgzk06RBA3LX2SN0,31388
 nkululeko/explore.py,sha256=Y5lPPychnI-7fyP8zvwVb9P09fvprbUPOofOppuABYQ,3658
 nkululeko/export.py,sha256=U-V4acxtuL6qKt6oAsVcM5TTeWogYUJ3GU-lA6rq6d4,4336
@@ -25,7 +25,7 @@ nkululeko/predict.py,sha256=b35YOqovGb5PLDz0nDuhJGykEAPq2Y45R9lzxJZMuMU,2083
 nkululeko/resample.py,sha256=akSAjJ3qn-O5NAyLJHVHdsK7MUZPGaZUvM2TwMSmj2M,5194
 nkululeko/runmanager.py,sha256=AswmORVUkCIH0gTx6zEyufvFATQBS8C5TXo2erSNdVg,7611
 nkululeko/scaler.py,sha256=7VOZ4sREMoQtahfETt9RyuR29Fb7PCwxlYVjBbdCVFc,4101
-nkululeko/segment.py,sha256=lSeI1i96HZTloSqdH75FhD7VyDQ16Do99-5mhI30To8,4571
+nkululeko/segment.py,sha256=S8TZt728CADXBEVw7GTWQq42vdXkRxmL738C8V_iy3k,4324
 nkululeko/syllable_nuclei.py,sha256=5w_naKxNxz66a_qLkraemi2fggM-gWesiiBPS47iFcE,9931
 nkululeko/test.py,sha256=1w624vo5KTzmFC8BUStGlLDmIEAFuJUz7J0W-gp7AxI,1677
 nkululeko/test_predictor.py,sha256=DEHE_D3A6m6KJTrpDKceA1n655t_UZV3WQd57K4a3Ho,2863
@@ -105,14 +105,14 @@ nkululeko/reporting/reporter.py,sha256=4OlYZAParkfJKO_aAyxqVpLc21zxZ-jDhtJKIMeUs
 nkululeko/reporting/result.py,sha256=G63a2tHCwHhM6NBJgYzsWKWJm4Yu3r4hsCHA2Km7eHU,1073
 nkululeko/segmenting/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nkululeko/segmenting/seg_inaspeechsegmenter.py,sha256=b3t0zdpJYofKWMyKRMtMMX91xeR-k8d5pbnNaQHcsOE,1902
-nkululeko/segmenting/seg_silero.py,sha256=CnhjKGTW5OXf-bmw4YsSJeN2yUwkY5m3xnulM_PYCW0,3256
+nkululeko/segmenting/seg_silero.py,sha256=ulodnvtRq5MLHDxy_RmAK4tJg6h1d-mPq-uCPFkGVKg,4258
 nkululeko/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nkululeko/utils/files.py,sha256=UiGAtZRWYjHSvlmPaTMtzyNNGE6qaLaxQkybctS7iRM,4021
 nkululeko/utils/stats.py,sha256=vCRzhCR0Gx5SiJyAGbj1TIto8ocGz58CM5Pr3LltagA,2948
 nkululeko/utils/util.py,sha256=XFZdhCc_LM4EmoZ5tKKaBCQLXclcNmvHwhfT_CXB98c,16723
-nkululeko-0.90.4.dist-info/LICENSE,sha256=0zGP5B_W35yAcGfHPS18Q2B8UhvLRY3dQq1MhpsJU_U,1076
-nkululeko-0.90.4.dist-info/METADATA,sha256=t64nFqxKkX3gaQ8J0PjpiRxc03LBS0yGO3i5wTR1bxc,41242
-nkululeko-0.90.4.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
-nkululeko-0.90.4.dist-info/entry_points.txt,sha256=KpQhz4HKBvYLrNooqLIc83hub76axRbYUgWzYkH3GnU,397
-nkululeko-0.90.4.dist-info/top_level.txt,sha256=DPFNNSHPjUeVKj44dVANAjuVGRCC3MusJ08lc2a8xFA,10
-nkululeko-0.90.4.dist-info/RECORD,,
+nkululeko-0.91.1.dist-info/LICENSE,sha256=0zGP5B_W35yAcGfHPS18Q2B8UhvLRY3dQq1MhpsJU_U,1076
+nkululeko-0.91.1.dist-info/METADATA,sha256=F-icPu_THEFHxMm-uAV5MALbVMLrXcO3ZuoPeo6bPwk,41417
+nkululeko-0.91.1.dist-info/WHEEL,sha256=OVMc5UfuAQiSplgO0_WdW7vXVGAt9Hdd6qtN4HotdyA,91
+nkululeko-0.91.1.dist-info/entry_points.txt,sha256=lNTkFEdh6Kjo5o95ZAWf_0Lq-4ztGoAoMVSDuPtuyS0,442
+nkululeko-0.91.1.dist-info/top_level.txt,sha256=DPFNNSHPjUeVKj44dVANAjuVGRCC3MusJ08lc2a8xFA,10
+nkululeko-0.91.1.dist-info/RECORD,,

{nkululeko-0.90.4.dist-info → nkululeko-0.91.1.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (75.1.0)
+Generator: setuptools (75.2.0)
 Root-Is-Purelib: true
 Tag: py3-none-any
 

{nkululeko-0.90.4.dist-info → nkululeko-0.91.1.dist-info}/entry_points.txt

@@ -1,6 +1,7 @@
 [console_scripts]
 nkululeko.augment = nkululeko.augment:main
 nkululeko.demo = nkululeko.demo:main
+nkululeko.ensemble = nkululeko.ensemble:main
 nkululeko.explore = nkululeko.explore:main
 nkululeko.export = nkululeko.export:main
 nkululeko.nkululeko = nkululeko.nkululeko:main