sonusai 0.15.6__py3-none-any.whl → 0.15.9__py3-none-any.whl

sonusai/torchl_predict.py

@@ -43,15 +43,38 @@ Outputs the following to tpredict-<TIMESTAMP> directory:
     torch_predict.log
 
 """
+from os import makedirs
+from os.path import basename
+from os.path import isdir
 from os.path import join
+from os.path import normpath
+from os.path import splitext
 from typing import Any
 
 import h5py
 import torch
+from docopt import docopt
+from lightning.pytorch import Trainer
 from lightning.pytorch.callbacks import BasePredictionWriter
+from pyaaware import FeatureGenerator
+from pyaaware import TorchInverseTransform
+from torchinfo import summary
 
+import sonusai
+from sonusai import create_file_handler
+from sonusai import initial_log_messages
 from sonusai import logger
+from sonusai import update_console_handler
+from sonusai.data_generator import TorchFromMixtureDatabase
 from sonusai.mixture import Feature
+from sonusai.mixture import MixtureDatabase
+from sonusai.mixture import get_audio_from_feature
+from sonusai.mixture import get_feature_from_audio
+from sonusai.mixture import read_audio
+from sonusai.utils import create_ts_name
+from sonusai.utils import import_keras_model
+from sonusai.utils import trim_docstring
+from sonusai.utils import write_wav
 
 
 class CustomWriter(BasePredictionWriter):
@@ -61,7 +84,7 @@ class CustomWriter(BasePredictionWriter):
 
     def write_on_epoch_end(self, trainer, pl_module, predictions, batch_indices):
         # this will create N (num processes) files in `output_dir` each containing
-        # the predictions of it's respective rank
+        # the predictions of its respective rank
         # torch.save(predictions, os.path.join(self.output_dir, f"predictions_{trainer.global_rank}.pt"))
 
         # optionally, you can also save `batch_indices` to get the information about the data index
@@ -119,11 +142,6 @@ def power_uncompress(real, imag):
 
 
 def main() -> None:
-    from docopt import docopt
-
-    import sonusai
-    from sonusai.utils import trim_docstring
-
     args = docopt(trim_docstring(__doc__), version=sonusai.__version__, options_first=True)
 
     verbose = args['--verbose']
@@ -139,27 +157,6 @@ def main() -> None:
     wavdbg = args['--wavdbg']  # write .wav if true
     input_name = args['INPUT']
 
-    from os import makedirs
-    from os.path import basename
-    from os.path import isdir
-    from os.path import isfile
-    from os.path import join
-    from os.path import splitext
-    from os.path import normpath
-    import h5py
-    # from sonusai.utils import float_to_int16
-
-    from torchinfo import summary
-    from sonusai import create_file_handler
-    from sonusai import initial_log_messages
-    from sonusai import update_console_handler
-    from sonusai.mixture import MixtureDatabase
-    from sonusai.mixture import get_feature_from_audio
-    from sonusai.utils import import_keras_model
-    from sonusai.mixture import read_audio
-    from sonusai.utils import create_ts_name
-    from sonusai.data_generator import TorchFromMixtureDatabase
-
     if batch_size is not None:
         batch_size = int(batch_size)
     if batch_size != 1:
@@ -222,6 +219,8 @@ def main() -> None:
                 hparams['timesteps'] = timesteps
 
         logger.info(f'Building model with hparams and batch_size={batch_size}, timesteps={timesteps}')
+        # hparams['cl_per_wght'] = 0.0
+        # hparams['feature'] = 'hum00ns1'
         try:
             model = litemodule.MyHyperModel(**hparams)  # use hparams
             # litemodule.MyHyperModel.load_from_checkpoint(ckpt_name, **hparams)
@@ -303,33 +302,25 @@ def main() -> None:
                                              drop_last=False,
                                              num_workers=dlcpu)
 
-        if wavdbg:  # setup for wav write if enabled
-            # Info needed to setup inverse transform
-            from pyaaware import FeatureGenerator
-            from pyaaware import TorchInverseTransform
-            from torchaudio import save
-            # from sonusai.utils import write_wav
-
-            half = model.num_classes // 2
-            fg = FeatureGenerator(feature_mode=model.hparams.feature,
-                                  num_classes=model.num_classes,
-                                  truth_mutex=model.truth_mutex)
-            itf = TorchInverseTransform(N=fg.itransform_N,
-                                        R=fg.itransform_R,
-                                        bin_start=fg.bin_start,
-                                        bin_end=fg.bin_end,
-                                        ttype=fg.itransform_ttype)
-
-            if mixdb.target_files[0].truth_settings[0].function == 'target_f' or \
-                    mixdb.target_files[0].truth_settings[0].function == 'target_mixture_f':
-                enable_truth_wav = True
-            else:
-                enable_truth_wav = False
-
+        # Info needed to set up inverse transform
+        half = model.num_classes // 2
+        fg = FeatureGenerator(feature_mode=model.hparams.feature,
+                              num_classes=model.num_classes,
+                              truth_mutex=model.truth_mutex)
+        itf = TorchInverseTransform(N=fg.itransform_N,
+                                    R=fg.itransform_R,
+                                    bin_start=fg.bin_start,
+                                    bin_end=fg.bin_end,
+                                    ttype=fg.itransform_ttype)
+
+        enable_truth_wav = False
+        enable_mix_wav = False
+        if wavdbg:
             if mixdb.target_files[0].truth_settings[0].function == 'target_mixture_f':
                 enable_mix_wav = True
-            else:
-                enable_mix_wav = False
+                enable_truth_wav = True
+            elif mixdb.target_files[0].truth_settings[0].function == 'target_f':
+                enable_truth_wav = True
 
         if reset:
             logger.info(f'Running {mixdb.num_mixtures} mixtures individually with model reset ...')
@@ -351,26 +342,25 @@ def main() -> None:
                 if wavdbg:
                     owav_base = splitext(output_name)[0]
                     tmp = torch.complex(ypred[..., :half], ypred[..., half:]).permute(2, 0, 1).detach()
+                    itf.reset()
                     predwav, _ = itf.execute_all(tmp)
-                    # predwav, _ = calculate_audio_from_transform(tmp, itf, trim=True)
-                    save(owav_base + '.wav', predwav.permute([1, 0]), 16000, encoding='PCM_S', bits_per_sample=16)
+                    # predwav, _ = calculate_audio_from_transform(tmp.numpy(), itf, trim=True)
+                    write_wav(owav_base + '.wav', predwav.permute([1, 0]).numpy(), 16000)
                     if enable_truth_wav:
                         # Note this support truth type target_f and target_mixture_f
                         tmp = torch.complex(val[0][..., :half], val[0][..., half:2 * half]).permute(2, 0, 1).detach()
+                        itf.reset()
                         truthwav, _ = itf.execute_all(tmp)
-                        save(owav_base + '_truth.wav', truthwav.permute([1, 0]), 16000, encoding='PCM_S',
-                             bits_per_sample=16)
+                        write_wav(owav_base + '_truth.wav', truthwav.permute([1, 0]).numpy(), 16000)
 
                     if enable_mix_wav:
                         tmp = torch.complex(val[0][..., 2 * half:3 * half], val[0][..., 3 * half:]).permute(2, 0, 1)
+                        itf.reset()
                         mixwav, _ = itf.execute_all(tmp.detach())
-                        save(owav_base + "_mix.wav", mixwav.permute([1, 0]), 16000, encoding='PCM_S',
-                             bits_per_sample=16)
-                        # write_wav(owav_base + "_truth.wav", truthwav, 16000)
+                        write_wav(owav_base + '_mix.wav', mixwav.permute([1, 0]).numpy(), 16000)
 
         else:
             logger.info(f'Running {mixdb.num_mixtures} mixtures with model builtin prediction loop ...')
-            from lightning.pytorch import Trainer
             pred_writer = CustomWriter(output_dir=output_dir, write_interval="epoch")
             trainer = Trainer(default_root_dir=output_dir,
                               callbacks=[pred_writer],
@@ -489,32 +479,37 @@ def main() -> None:
         # logger.info(f'Saved results to {output_dir}')
         # return
 
-    if not all(isfile(file) and splitext(file)[1] == '.wav' for file in input_name):
-        logger.exception(f'Do not know how to process input from {input_name}')
-        raise SystemExit(1)
-
-    logger.info(f'Run prediction on {len(input_name):,} WAV files')
+    logger.info(f'Run prediction on {len(input_name):,} audio files')
     for file in input_name:
-        # Convert WAV to feature data
-        audio = read_audio(file)
-        feature = get_feature_from_audio(audio=audio, feature=model.feature)
+        # Convert audio to feature data
+        audio_in = read_audio(file)
+        feature = get_feature_from_audio(audio=audio_in, feature_mode=model.hparams.feature)
 
-        # feature, predict = _pad_and_predict(hypermodel=hypermodel,
-        #                                     built_model=built_model,
-        #                                     feature=feature,
-        #                                     frames_per_batch=frames_per_batch)
+        with torch.no_grad():
+            predict = model(torch.tensor(feature))
 
-        # clean = torch_istft_olsa_hanns(clean_spec_cmplx, mixdb.ift_config.N, mixdb.ift_config.R)
+        audio_out = get_audio_from_feature(feature=predict.numpy(), feature_mode=model.hparams.feature)
 
         output_name = join(output_dir, splitext(basename(file))[0] + '.h5')
         with h5py.File(output_name, 'a') as f:
+            if 'audio_in' in f:
+                del f['audio_in']
+            f.create_dataset(name='audio_in', data=audio_in)
+
             if 'feature' in f:
                 del f['feature']
             f.create_dataset(name='feature', data=feature)
 
-            # if 'predict' in f:
-            #     del f['predict']
-            # f.create_dataset(name='predict', data=predict)
+            if 'predict' in f:
+                del f['predict']
+            f.create_dataset(name='predict', data=predict)
+
+            if 'audio_out' in f:
+                del f['audio_out']
+            f.create_dataset(name='audio_out', data=audio_out)
+
+        output_name = join(output_dir, splitext(basename(file))[0] + '_predict.wav')
+        write_wav(output_name, audio_out, 16000)
 
     logger.info(f'Saved results to {output_dir}')
     del model

sonusai/utils/__init__.py

@@ -2,10 +2,14 @@
 from .asl_p56 import asl_p56
 from .asr import ASRResult
 from .asr import calc_asr
+from .audio_devices import get_default_input_device
+from .audio_devices import get_input_device_index_by_name
+from .audio_devices import get_input_devices
 from .braced_glob import braced_glob
 from .braced_glob import braced_iglob
 from .calculate_input_shape import calculate_input_shape
 from .convert_string_to_number import convert_string_to_number
+from .create_timestamp import create_timestamp
 from .create_ts_name import create_ts_name
 from .dataclass_from_dict import dataclass_from_dict
 from .db import db_to_linear

sonusai/utils/audio_devices.py

@@ -0,0 +1,41 @@
+import pyaudio
+
+
+def get_input_device_index_by_name(p: pyaudio.PyAudio, name: str = None) -> int:
+    info = p.get_host_api_info_by_index(0)
+    device_count = info.get('deviceCount')
+    for i in range(0, device_count):
+        device_info = p.get_device_info_by_host_api_device_index(0, i)
+        if name is None:
+            device_name = None
+        else:
+            device_name = device_info.get('name')
+        if name == device_name and device_info.get('maxInputChannels') > 0:
+            return i
+
+    raise ValueError(f'Could not find {name}')
+
+
+def get_input_devices(p: pyaudio.PyAudio) -> list[str]:
+    names = []
+    info = p.get_host_api_info_by_index(0)
+    device_count = info.get('deviceCount')
+    for i in range(0, device_count):
+        device_info = p.get_device_info_by_host_api_device_index(0, i)
+        device_name = device_info.get('name')
+        if device_info.get('maxInputChannels') > 0:
+            names.append(device_name)
+
+    return names
+
+
+def get_default_input_device(p: pyaudio.PyAudio) -> str:
+    info = p.get_host_api_info_by_index(0)
+    device_count = info.get('deviceCount')
+    for i in range(0, device_count):
+        device_info = p.get_device_info_by_host_api_device_index(0, i)
+        device_name = device_info.get('name')
+        if device_info.get('maxInputChannels') > 0:
+            return device_name
+
+    raise ValueError('No input audio devices found')

sonusai/utils/calculate_input_shape.py

@@ -13,13 +13,12 @@ def calculate_input_shape(feature: str,
     """
     from pyaaware import FeatureGenerator
 
-    # num_classes is irrelevant, set to 2
-    fg = FeatureGenerator(feature_mode=feature, num_classes=2)
+    fg = FeatureGenerator(feature_mode=feature)
 
     if flatten:
-        in_shape = [fg.stride * fg.num_bands]
+        in_shape = [fg.stride * fg.feature_parameters]
     else:
-        in_shape = [fg.stride, fg.num_bands]
+        in_shape = [fg.stride, fg.feature_parameters]
 
     if timesteps > 0:
         in_shape.insert(0, timesteps)

sonusai/utils/create_timestamp.py

@@ -0,0 +1,5 @@
+def create_timestamp() -> str:
+    """Create a timestamp."""
+    from datetime import datetime
+
+    return datetime.now().strftime('%Y%m%d-%H%M%S')

sonusai/utils/reshape.py

@@ -17,14 +17,14 @@ def reshape_inputs(feature: Feature,
                    timesteps: int = 0,
                    flatten: bool = False,
                    add1ch: bool = False) -> tuple[Feature, Optional[Truth]]:
-    """Check SonusAI feature and truth data and reshape feature of size [frames, strides, bands] into
+    """Check SonusAI feature and truth data and reshape feature of size [frames, strides, feature_parameters] into
     one of several options:
 
     If timesteps > 0: (i.e., for recurrent NNs):
-      no-flatten, no-channel:   [sequences, timesteps, strides, bands]      (4-dim)
-      flatten, no-channel:      [sequences, timesteps, strides*bands]       (3-dim)
-      no-flatten, add-1channel: [sequences, timesteps, strides, bands, 1]   (5-dim)
-      flatten, add-1channel:    [sequences, timesteps, strides*bands, 1]    (4-dim)
+      no-flatten, no-channel:   [sequences, timesteps, strides, feature_parameters]      (4-dim)
+      flatten, no-channel:      [sequences, timesteps, strides*feature_parameters]       (3-dim)
+      no-flatten, add-1channel: [sequences, timesteps, strides, feature_parameters, 1]   (5-dim)
+      flatten, add-1channel:    [sequences, timesteps, strides*feature_parameters, 1]    (4-dim)
 
     If batch_size is None, then do not reshape; just calculate new input shape and return.
 
@@ -40,7 +40,7 @@ def reshape_inputs(feature: Feature,
     """
     from sonusai import SonusAIError
 
-    frames, strides, bands = feature.shape
+    frames, strides, feature_parameters = feature.shape
     if truth is not None:
         truth_frames, num_classes = truth.shape
         # Double-check correctness of inputs
@@ -50,7 +50,7 @@ def reshape_inputs(feature: Feature,
         num_classes = None
 
     if flatten:
-        feature = np.reshape(feature, (frames, strides * bands))
+        feature = np.reshape(feature, (frames, strides * feature_parameters))
 
     # Reshape for Keras/TF recurrent models that require timesteps/sequence length dimension
     if timesteps > 0:
@@ -73,14 +73,14 @@ def reshape_inputs(feature: Feature,
 
         # Reshape
         if feature.ndim == 2:  # flattened input
-            # was [frames, bands*timesteps]
-            feature = np.reshape(feature, (sequences, timesteps, strides * bands))
+            # was [frames, feature_parameters*timesteps]
+            feature = np.reshape(feature, (sequences, timesteps, strides * feature_parameters))
             if truth is not None:
                 # was [frames, num_classes]
                 truth = np.reshape(truth, (sequences, timesteps, num_classes))
         elif feature.ndim == 3:  # un-flattened input
-            # was [frames, bands, timesteps]
-            feature = np.reshape(feature, (sequences, timesteps, strides, bands))
+            # was [frames, feature_parameters, timesteps]
+            feature = np.reshape(feature, (sequences, timesteps, strides, feature_parameters))
             if truth is not None:
                 # was [frames, num_classes]
                 truth = np.reshape(truth, (sequences, timesteps, num_classes))

sonusai/utils/wave.py

@@ -5,15 +5,22 @@ from sonusai.mixture.datatypes import AudioT
 def write_wav(name: str, audio: AudioT, sample_rate: int = SAMPLE_RATE) -> None:
     """ Write a simple, uncompressed WAV file.
 
-    To write multiple channels, use a 2D array of shape [samples, channels].
+    To write multiple channels, use a 2D array of shape [channels, samples].
     The bits per sample and PCM/float are determined by the data type.
 
     """
-    import numpy as np
     import torch
     import torchaudio
 
-    if audio.ndim == 1:
-        audio = np.reshape(audio, (1, audio.shape[0]))
+    data = torch.tensor(audio)
 
-    torchaudio.save(name, torch.tensor(audio), sample_rate)
+    if data.dim() == 1:
+        data = torch.reshape(data, (1, data.shape[0]))
+    if data.dim() != 2:
+        raise ValueError(f'audio must be a 1D or 2D array')
+
+    # Assuming data has more samples than channels, check if array needs to be transposed
+    if data.shape[1] < data.shape[0]:
+        data = torch.transpose(data, 0, 1)
+
+    torchaudio.save(uri=name, src=data, sample_rate=sample_rate)

{sonusai-0.15.6.dist-info → sonusai-0.15.9.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sonusai
-Version: 0.15.6
+Version: 0.15.9
 Summary: Framework for building deep neural network models for sound, speech, and voice AI
 Home-page: https://aaware.com
 License: GPL-3.0-only
@@ -16,40 +16,43 @@ Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Requires-Dist: PyYAML (>=6.0.1,<7.0.0)
 Requires-Dist: aixplain (>=0.2.6,<0.3.0)
+Requires-Dist: bitarray (>=2.9.2,<3.0.0)
+Requires-Dist: ctranslate2 (==4.1.0)
 Requires-Dist: dataclasses-json (>=0.6.1,<0.7.0)
 Requires-Dist: deepgram-sdk (>=3.0.0,<4.0.0)
 Requires-Dist: docopt (>=0.6.2,<0.7.0)
-Requires-Dist: faster-whisper (>=0.10.0,<0.11.0)
-Requires-Dist: ffmpeg (>=1.4,<2.0)
-Requires-Dist: greenlet (>=3.0.1,<4.0.0)
-Requires-Dist: grpcio (==1.60.0)
+Requires-Dist: einops (>=0.7.0,<0.8.0)
+Requires-Dist: faster-whisper (>=1.0.1,<2.0.0)
+Requires-Dist: geomloss (>=0.2.6,<0.3.0)
+Requires-Dist: h5py (>=3.11.0,<4.0.0)
+Requires-Dist: hydra-core (>=1.3.2,<2.0.0)
 Requires-Dist: jiwer (>=3.0.3,<4.0.0)
-Requires-Dist: keras-tuner (>=1.4.5,<2.0.0)
+Requires-Dist: keras (>=3.1.1,<4.0.0)
+Requires-Dist: keras-tuner (>=1.4.7,<2.0.0)
 Requires-Dist: librosa (>=0.10.1,<0.11.0)
-Requires-Dist: lightning (>=2.1.0,<3.0.0)
+Requires-Dist: lightning (>=2.2,<2.3)
 Requires-Dist: matplotlib (>=3.8.0,<4.0.0)
-Requires-Dist: onnx (==1.14.1)
+Requires-Dist: omegaconf (>=2.3.0,<3.0.0)
+Requires-Dist: onnx (>=1.14.1,<2.0.0)
 Requires-Dist: onnxruntime (>=1.16.1,<2.0.0)
-Requires-Dist: paho-mqtt (>=1.6.1,<2.0.0)
+Requires-Dist: paho-mqtt (>=2.0.0,<3.0.0)
 Requires-Dist: pandas (>=2.1.1,<3.0.0)
 Requires-Dist: pesq (>=0.0.4,<0.0.5)
 Requires-Dist: pyaaware (>=1.5.3,<2.0.0)
+Requires-Dist: pyaudio (>=0.2.14,<0.3.0)
 Requires-Dist: pydub (>=0.25.1,<0.26.0)
 Requires-Dist: pystoi (>=0.4.0,<0.5.0)
 Requires-Dist: python-magic (>=0.4.27,<0.5.0)
 Requires-Dist: requests (>=2.31.0,<3.0.0)
-Requires-Dist: resampy (>=0.4.2,<0.5.0)
-Requires-Dist: samplerate (>=0.1.0,<0.2.0)
-Requires-Dist: scikit-learn (>=1.3.1,<2.0.0)
-Requires-Dist: sh (>=2.0.6,<3.0.0)
+Requires-Dist: sacrebleu (>=2.4.2,<3.0.0)
+Requires-Dist: samplerate (>=0.2.1,<0.3.0)
 Requires-Dist: soundfile (>=0.12.1,<0.13.0)
 Requires-Dist: sox (>=1.4.1,<2.0.0)
-Requires-Dist: speechrecognition (>=3.10.0,<4.0.0)
+Requires-Dist: speechrecognition (>=3.10.1,<4.0.0)
 Requires-Dist: tensorflow (>=2.15.0,<3.0.0)
-Requires-Dist: tensorflow-addons (>=0.23.0,<0.24.0)
 Requires-Dist: tf2onnx (>=1.15.1,<2.0.0)
-Requires-Dist: torch (>=2.1.0,<3.0.0)
-Requires-Dist: torchaudio (>=2.1.0,<3.0.0)
+Requires-Dist: torch (>=2.2,<2.3)
+Requires-Dist: torchaudio (>=2.2,<2.3)
 Requires-Dist: torchinfo (>=1.8.0,<2.0.0)
 Requires-Dist: tqdm (>=4.66.1,<5.0.0)
 Description-Content-Type: text/x-rst
@@ -59,7 +62,7 @@ Sonus AI: Framework for simplified creation of deep NN models for sound, speech,
 Sonus AI includes functions for pre-processing training and validation data and
 creating performance metrics reports for key types of Keras models:
 - recurrent, convolutional, or a combination (i.e. RCNNs)
-- binary, multiclass single-label, multiclass multi-label, and regresssion
+- binary, multiclass single-label, multiclass multi-label, and regression
 - training with data augmentations:  noise mixing, pitch and time stretch, etc.
 
 Sonus AI python functions are used by:

{sonusai-0.15.6.dist-info → sonusai-0.15.9.dist-info}/RECORD

@@ -1,27 +1,27 @@
 sonusai/__init__.py,sha256=KmIJ9wni9d9v5pyu0pUxbacZIHGkAywB9CJwl7JME28,1526
 sonusai/aawscd_probwrite.py,sha256=GukR5owp_0A3DrqSl9fHWULYgclNft4D5OkHIwfxxkc,3698
-sonusai/calc_metric_spenh.py,sha256=Cf02uYB6fzDX5anQuBTR_aD7mYt1WgxkEIzZA08uFvs,60825
+sonusai/audiofe.py,sha256=XE_cgOhhTryjPUePxW_8NY1TwrnRZ6BHCsH-gp8PmYw,11471
+sonusai/calc_metric_spenh.py,sha256=D8iQVSIhFhrsUwKuIP-S38NBnyfAOZlsOIIgOZwGOOI,60852
 sonusai/data/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 sonusai/data/genmixdb.yml,sha256=-XSs_hUR6wHJVoTPmSewzXL7u61X-xmHY46lNPatxSE,1025
 sonusai/data/speech_ma01_01.wav,sha256=PK0vMKg-NR6rPE3KouxHGF6PKXnJCr7AwjMqfu98LUA,76644
 sonusai/data/whitenoise.wav,sha256=I2umov0m34y56F9IsIBi1XtE76ZeZaSKDf70cJRe3pI,1920044
 sonusai/data_generator/__init__.py,sha256=ouCpY5EDV35fKFeKGQfIcU8uE-c3QcuNerTxUA1X5L8,232
-sonusai/data_generator/dataset_from_mixdb.py,sha256=Fe_7xxCYURvbZjuIZrvzozdkrx7ZO7gxGg8c5ob2dys,5478
-sonusai/data_generator/keras_from_mixdb.py,sha256=V5CUsGz-akIYdgQy9ABxwNKMYKv01klA4xtMDveF6uI,6167
+sonusai/data_generator/dataset_from_mixdb.py,sha256=D14L8BL7a0WgkF8a8eogQ9Hk9ow4_RK3QBGsZ-HDAog,5493
+sonusai/data_generator/keras_from_mixdb.py,sha256=14r89aX6Dr9ZKsmMRC7HDXbJrPrCZC1liwwLmZUKj0w,6182
 sonusai/data_generator/torch_from_mixdb.py,sha256=lvEe9DDu_rIaoyhv9PW4UAnAWp5N74L8kRfxUhsh7oo,4279
 sonusai/doc/__init__.py,sha256=rP5Hgn0Iys_xkuv4caxngdqehuU4zLZsiKuv8Nde67M,19
 sonusai/doc/doc.py,sha256=3z210v6ZckuOlsGZ3ySQBdlCNmBp2M1ahqhqG_eUN58,22664
 sonusai/doc.py,sha256=l8CaFgLI8mqx4tn0aXfxKqa2dy9GgC0zjYxZAkpmi1E,878
-sonusai/evaluate.py,sha256=1BX9dCXMHg6LefbvkevxYpjM3JR-i0jR3Eob2oNh_hQ,9983
-sonusai/genft.py,sha256=CeQN8Sxb_NKeXWJxN9HtzUu687eXl97nHBxzzCzQdLg,5557
+sonusai/genft.py,sha256=6hOds8d-pYRawesLYh7XLrQh4PweWUj8U5Cbzx45bNQ,5572
 sonusai/genmix.py,sha256=0AiUfF7n0CGOp5v-woNfeP3-QaVQUb0CJZc0oXkvqpk,7016
-sonusai/genmixdb.py,sha256=onWkaCPzmUMvDtiGf3A7UdGkOA8xe9zuJTsuLfMdb_s,19597
+sonusai/genmixdb.py,sha256=rAxCKNPkOXaAugEfp9pTcHCQONapdTnxMlBsIPIoizE,19639
 sonusai/gentcst.py,sha256=8jeXirgJe0OGgknC8A-rIudjHeH8UTYPpuh71Ha-I3w,20165
 sonusai/keras_onnx.py,sha256=WHcPIcff4VPdiXqGX-TU-_x-UuKUD3nNpQtCX-2NEVQ,2658
-sonusai/keras_predict.py,sha256=ffrtI12T8_Rm26KsBee0dfEcnEuhOxvtV80YmbJR0m0,9085
+sonusai/keras_predict.py,sha256=_83EtPtnfrqwUzC2H2tk4LI90RiQdyEEBxFGTgFPl3M,9090
 sonusai/keras_train.py,sha256=8_M5vY-CkonPzbOtOF3Vk-wox-42o8fkaOKLjk7Oc2k,13226
-sonusai/lsdb.py,sha256=qTEHQ5X9Ruc0ph4OUhYKFZ6xGC3gyhcNq7t25oyLhc8,5929
-sonusai/main.py,sha256=_s3WCFP_zQ7pD0OgdSVRPQRE38V5Qp6bG56JX-Y79ME,3120
+sonusai/lsdb.py,sha256=TTMQ-0H8fFzUSczt6yjy-9xUjZSdIGQzTVH5Xr6XPSA,5941
+sonusai/main.py,sha256=KjN0dCI6rWare4wo_ACzTlURW7pvTw03n51pH7EyLAU,3108
 sonusai/metrics/__init__.py,sha256=56itZW3S1I7ZYvbxPmFIVPAh1AIJZdljByz1uCrHqFE,635
 sonusai/metrics/calc_class_weights.py,sha256=dyY7daEIf5Ms5tfTf6wF0fkx_GnMADHOZR_rtsfGoVM,3933
 sonusai/metrics/calc_optimal_thresholds.py,sha256=9fRfwl-aKAbzHJyqGHv4o8BpZXG9HHB7zUJObHXfYM4,3522
@@ -35,24 +35,24 @@ sonusai/metrics/class_summary.py,sha256=4Mb25nuk6eqotnQSFMuOQL3zofGcpNXDfDlPa513
 sonusai/metrics/confusion_matrix_summary.py,sha256=3qg6TMKjJeHtNjj2YnNjPFSlMrQXt0Zcu1dLkGB_aPU,4001
 sonusai/metrics/one_hot.py,sha256=QSeH_GdqBpOAKLrNnQ8gjcPC-vSdUqC0yPEQueTA6VI,13548
 sonusai/metrics/snr_summary.py,sha256=P4U5_Xr7v9F8kF-rZBnpsVNt3p42rIVS6zmch8yfVfg,5575
-sonusai/mixture/__init__.py,sha256=WzPHGSWz6v64HCSFTGboG5o-xBy_0II4i4tkf1UL1Vw,5251
-sonusai/mixture/audio.py,sha256=6W2ihjJGBy7Xggx1imF7bzkymDuipPOgC63j2J7Wf-E,3456
+sonusai/mixture/__init__.py,sha256=fCVSlizYxUUQQD9nSZ8bEbfc_TB2yiOC14HPOB4KFz4,5287
+sonusai/mixture/audio.py,sha256=S-ZROf5rVvwv1TCEuwJHz1FfX4oVubb4QhbybUMMqtM,2150
 sonusai/mixture/augmentation.py,sha256=Blb90tdTwBOj5w9tRcYyS5H67YJuFiXsGqwZWd7ON4g,10468
 sonusai/mixture/class_count.py,sha256=_wFnVl2yEOnbor7pLg7cYOUeX6nioov-03Cv3SEbh2k,996
 sonusai/mixture/config.py,sha256=CXIkVRJmaW2QW_sGl0aIqPf7I_TesyGhUYzxouw5UX4,22266
 sonusai/mixture/constants.py,sha256=xjCskcQi6khqYZDf7j6z1OkeN1C6wE06kBBapcJiNI4,1428
-sonusai/mixture/datatypes.py,sha256=xN-GdPCEHGE2Ak_TdFbjuSyMs4x7TLRp59trbMTiYLg,8164
+sonusai/mixture/datatypes.py,sha256=zaxfOHw8ddt-i8JPYOPnlqWz_EHBEDoO4q2VAqJViHM,8173
 sonusai/mixture/eq_rule_is_valid.py,sha256=MpQwRA5M76wSiQWEI1lW2cLFdPaMttBLcQp3tWD8efM,1243
-sonusai/mixture/feature.py,sha256=io6OiJAJ3GYvPChiUmPQuP3h0OB2onjYF8o9-AWkmqM,1996
-sonusai/mixture/generation.py,sha256=5SXV1R9os4XPN0pVm8ribG23vsXkYzsegoctRZdXQT4,39039
-sonusai/mixture/helpers.py,sha256=XqpcB15MezEMVJwf3jxzATDJSpj_27b8Cru1TDIFD7w,21326
-sonusai/mixture/log_duration_and_sizes.py,sha256=r-wVjrLW1XBciOL4pkZSYMR7ZNADbojE95TPSQkp3kc,1329
+sonusai/mixture/feature.py,sha256=Rwuf82IoXzhHPGbKYVGcatImF_ssBf_FfvbqghVPXtg,4116
+sonusai/mixture/generation.py,sha256=miUrc3QOSUNIG6mDkiMCZ6M2ulivUZxlYUAJUOVomWc,39039
+sonusai/mixture/helpers.py,sha256=GSGSD2KnvOeEIB6IwNTxyaQNjghTSBMB729kUEd_RiM,22403
+sonusai/mixture/log_duration_and_sizes.py,sha256=baTUpqyM15wA125jo9E3posmVJUe3WlpksyO6v9Jul0,1347
 sonusai/mixture/mapped_snr_f.py,sha256=mlbYM1t14OXe_Zg4CjpWTuA_Zun4W0O3bSUXeodRBQs,1845
-sonusai/mixture/mixdb.py,sha256=FQ5hirb2zR8Aj1UNtz89qJQ8wlE0ELC80IxQDmyhsKk,45188
+sonusai/mixture/mixdb.py,sha256=9Pe0mEG8pnEf9NZynTIldc05GfdOrgmcVoIt63RG5DA,45279
 sonusai/mixture/soundfile_audio.py,sha256=Ow_IWIMz4pMsLxMP_JsQ8AuHLCWlYQinLa58CFW97f8,2804
 sonusai/mixture/sox_audio.py,sha256=HT3kYA9TP5QPCuoOJdUMnGVN-qY6q96DGL8zxuog76o,12277
 sonusai/mixture/sox_augmentation.py,sha256=F9tBdNvX2guCn7gRppAFrxRnBtjw9q6qAq2_v_A4hh0,4490
-sonusai/mixture/spectral_mask.py,sha256=LKFnrqZryPHT6FBNiT7yFxOeXc6-AUg6X54N26d8ctg,2107
+sonusai/mixture/spectral_mask.py,sha256=8AkCwhy-PSdP1Uri9miKZP-bXFYnFcH_c9xZCGrHavU,2071
 sonusai/mixture/target_class_balancing.py,sha256=NTNiKZH0_PWLooeow0l41CjJKK8ZTMVbUqz9ZkaNtWk,4900
 sonusai/mixture/targets.py,sha256=wyy5vhLhuN-hqBMBGoziVvEJg3FKFvJFgmEE7_LaV2M,7908
 sonusai/mixture/tokenized_shell_vars.py,sha256=gCxw8SQUcal6mqWKF7hOBTgSQmbJUk1nT0Gn3H8GA0U,4705
@@ -61,24 +61,24 @@ sonusai/mixture/torchaudio_augmentation.py,sha256=1vEDHI0caL1vrgoY2lAWe4CiHE2jKR
 sonusai/mixture/truth.py,sha256=Y41pZ52Xkols9LUler0NlgnilUOscBIucmw4GcxXNzU,1612
 sonusai/mixture/truth_functions/__init__.py,sha256=82lKYHhLy8KW3gHngrocoqwupGVLVsWdIXdYs3vhjOc,359
 sonusai/mixture/truth_functions/crm.py,sha256=_Vy8UMrOUQXsrM3nutvUMWCpvI8GePr01QFlyqLFd4k,2626
-sonusai/mixture/truth_functions/data.py,sha256=wUlZAAeE4AXg4Egtcvs7Qhbf5dnWur-1Qr_Dz6GoJiQ,2778
+sonusai/mixture/truth_functions/data.py,sha256=okFJeOf43NxfdLqWFCBA2pOGqujRlNDYdAcwwR_m8z8,2875
 sonusai/mixture/truth_functions/energy.py,sha256=ydMtMLjMloG76DB30ZHQ5tkBVh4dkMJ82XEhKBokmIk,4281
 sonusai/mixture/truth_functions/file.py,sha256=jOJuC_3y9BH6GGOp9eKcbVrHLVRzUA80BJq59LhcBUM,1539
 sonusai/mixture/truth_functions/phoneme.py,sha256=stYdlPuNytQK_LLT61OJLfYSqKd-sDjQZdtJKGzt5wA,479
 sonusai/mixture/truth_functions/sed.py,sha256=8cHjEFjZaH_0hIOHhPmj4AJz2GpEADM6Ys2x4NoiWSY,2469
-sonusai/mixture/truth_functions/target.py,sha256=V-bXcJZ3P83NHKqYs4RcsxxuZdO8kqzSEN5CNRa5e0k,5394
+sonusai/mixture/truth_functions/target.py,sha256=KAsjugDRooOA5BRcHVAbZRgV7l8S5CFg7CZ0XtKZaQ0,5764
 sonusai/mkmanifest.py,sha256=dIPVFKKhnhHdq63OGr6p__pK7fyx3OdKVtbmGUJxsR8,7078
 sonusai/mkwav.py,sha256=LZNyhq4gJEs_NtGvRsYHA2qfgkkODpt6HoH1b-Tjjuw,5266
-sonusai/onnx_predict.py,sha256=RhQbbNG3w6rCXuSFUWCaQmUH5JzSP2hmu6TG5_81IVA,9055
-sonusai/plot.py,sha256=GPrbk7SxwojtjR9KLE9jQ6ywYWXca296W0iaRPdnBHI,16982
-sonusai/post_spenh_targetf.py,sha256=Y9J5JbFnCFKCednNiYvVXaC5Z3lF5KxifFJ2RCy4jmg,4975
+sonusai/onnx_predict.py,sha256=Bz_pR28oAZBarNajlKwyzBxmW7ktum77SmxDN2onKPM,9060
+sonusai/plot.py,sha256=u-PvF8guNcm0b-GN99xfEkrcAAtidAEY3RLDzNvcyYk,17014
+sonusai/post_spenh_targetf.py,sha256=NIMhDXeDuUqeWukNaAUMvDw9JpEVCauwjrL2F4M9nrI,4927
 sonusai/queries/__init__.py,sha256=oKY5JeqZ4Cz7DwCwPc1_ydB8bUs6KaMcWFp_w02TjOs,255
 sonusai/queries/queries.py,sha256=FNMUKnoY_Ya9S5sNhsB8ppwy0B7V55ilbbjhQRv_UN8,7552
 sonusai/torchl_onnx.py,sha256=5JYow3XpBaUdtuyAW0mOZyCKL_4FrHvEekYBRdDT6KA,8967
-sonusai/torchl_predict.py,sha256=-wlUdRGPjOvGjCQZY277D8tGeZp1KyDU7TpYH7ovC0c,26657
+sonusai/torchl_predict.py,sha256=P1ySDH_ITOPefZ2xZqyxyIrsNDqblKTBLZqFApgo5EU,26238
 sonusai/torchl_train.py,sha256=NPCRB0gwTvabivmOz78gjUreDeO1z16PYuw7L1-pIRQ,9680
 sonusai/tplot.py,sha256=yFyyyg9ymp2Eh-64Muu0EFFEY61MoJSV0a_fy9OWaCk,14485
-sonusai/utils/__init__.py,sha256=ynmdoPJVdbk5eYq9TA9t4QpGEkdiGATDx1pFu43u7YQ,2180
+sonusai/utils/__init__.py,sha256=tVSmxinSo0Enexpol6wCzz6tU7WrueC-YslFgQr-o7M,2382
 sonusai/utils/asl_p56.py,sha256=GCKlz-NLInQ0z41XBi0mOvGdSfRZf3WI53necVNDo80,3837
 sonusai/utils/asr.py,sha256=QN1wdO9-EqD72-ixr4lnzsPfT8i0syhTGj1evKNJWe4,2021
 sonusai/utils/asr_functions/__init__.py,sha256=4boXXOXlQHTt8K2DWOwFXSlc8D2NLFd8QTc68yL2ejU,214
@@ -93,9 +93,11 @@ sonusai/utils/asr_manifest_functions/__init__.py,sha256=Lz12aCGvfngZkLoUxHSqFjHc
 sonusai/utils/asr_manifest_functions/data.py,sha256=mJsaHccBReguOJu9qsshRhL-3GbeyqM0-PXMseFnZbE,151
 sonusai/utils/asr_manifest_functions/librispeech.py,sha256=HIaytcYmjRUkuR6fCQlv3Jh3IDWSox_A6WFcFFAHN9M,1635
 sonusai/utils/asr_manifest_functions/vctk_noisy_speech.py,sha256=-69lM0dz18KbU5_-dmSeqDoNNwgJj4UlxgGkNBEi3wM,2169
+sonusai/utils/audio_devices.py,sha256=LgaXTln1oRArBzaet3rZiIO2plgtaThuGBc3sJ_sLlo,1414
 sonusai/utils/braced_glob.py,sha256=h4hab7YDbM4CjLg9iSzyHZrkd22IPUOY5zZqHdifkh8,1510
-sonusai/utils/calculate_input_shape.py,sha256=fyf8Ggxn9xljJ87BNwIzOh2KD4pKUGFZ2RNzE3195NI,1023
+sonusai/utils/calculate_input_shape.py,sha256=63ILxibYKuTQozY83QN8Y2OOhBEbW_1X47Q0askcHDM,984
 sonusai/utils/convert_string_to_number.py,sha256=i17yIxurp8Iz6NPE-imTRlARrXWqadwm8qbOTuzHZvE,236
+sonusai/utils/create_timestamp.py,sha256=TxoQXWZ3SFdBEHLOv-ujeIsTEJuiFnKOGRy-FQq45YU,148
 sonusai/utils/create_ts_name.py,sha256=8RLKmgXwuGcbDMGgtTuc0MvGFfA7IOVqfjkE2T18GOo,405
 sonusai/utils/dataclass_from_dict.py,sha256=vAGnuMjhy0W9bxZ5usrH7mbQsFog3n0__IC4xyJyVUc,390
 sonusai/utils/db.py,sha256=lI77MJJLs4CTYxhjFUvBom2Kk2imAP34okOeO4irbDc,371
@@ -114,15 +116,15 @@ sonusai/utils/print_mixture_details.py,sha256=BzYM4-wHHNa6zxPzBMUJxwKt0gKHmvbwdd
 sonusai/utils/ranges.py,sha256=NPBZOVzMb95GTOIxltVO-wSzgcXqZ14wbdV46JDLKrw,1222
 sonusai/utils/read_mixture_data.py,sha256=Sb30RgSpw6DnH_iD81O7G_KOsdfjQWWLk3euEkxfMa8,453
 sonusai/utils/read_predict_data.py,sha256=5rR_ijrrcS2cKO1Sea2M2QEicokTtW5XtAo6jT5YSX8,1064
-sonusai/utils/reshape.py,sha256=sz-V4Za3DIvqFztTh0yYhQRcofnjc9XBW0CBhxg18lo,5854
+sonusai/utils/reshape.py,sha256=E8Eu6grynaeWwVO6peIR0BF22SrVaJSa1Rkl109lq6Y,5997
 sonusai/utils/seconds_to_hms.py,sha256=oxLuZhTJJr9swj-fOSOrZJ5vBNM7_BrOMQhX1pYpiv0,260
 sonusai/utils/stacked_complex.py,sha256=feLhz3GC1ILxBGMHOj3sJK--sidsXKbfwkalwAVwizc,2950
 sonusai/utils/stratified_shuffle_split.py,sha256=rJNXvBp-GxoKzH3OpL7k0ANSu5xMP2zJ7K1fm_33UzE,7022
 sonusai/utils/trim_docstring.py,sha256=dSrtiRsEN4wkkvKBp6WDr13RUypfqZzgH_jOBLs1ouY,881
-sonusai/utils/wave.py,sha256=TKE-CNPGFXNXUW626CBPzCTNgWJut8I0ZEUsgG9q4Po,586
+sonusai/utils/wave.py,sha256=O4ZXkZ6wjrKGa99wBCdFd8G6bp91MXXDnmGihpaEMh0,856
 sonusai/utils/yes_or_no.py,sha256=eMLXBVH0cEahiXY4W2KNORmwNQ-ba10eRtldh0y4NYg,263
 sonusai/vars.py,sha256=m2AefF0m5bXWGXpJj8Pi42zWL2ydeEj7bkak3GrtMyM,940
-sonusai-0.15.6.dist-info/METADATA,sha256=bUbyCYkQJaHaxBfqMyieWqqGRewQJ3Vu0LFzGpOR3Lg,3096
-sonusai-0.15.6.dist-info/WHEEL,sha256=FMvqSimYX_P7y0a7UY-_Mc83r5zkBZsCYPm7Lr0Bsq4,88
-sonusai-0.15.6.dist-info/entry_points.txt,sha256=zMNjEphEPO6B3cD1GNpit7z-yA9tUU5-j3W2v-UWstU,92
-sonusai-0.15.6.dist-info/RECORD,,
+sonusai-0.15.9.dist-info/METADATA,sha256=DudNQlTEQpWpzqyzyowz_V-J9epd7mrKgAYM6rFxaPo,3209
+sonusai-0.15.9.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+sonusai-0.15.9.dist-info/entry_points.txt,sha256=zMNjEphEPO6B3cD1GNpit7z-yA9tUU5-j3W2v-UWstU,92
+sonusai-0.15.9.dist-info/RECORD,,

{sonusai-0.15.6.dist-info → sonusai-0.15.9.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: poetry-core 1.8.1
+Generator: poetry-core 1.9.0
 Root-Is-Purelib: true
 Tag: py3-none-any