[Hackathon 7th No.55] Add audiotools to PaddleSpeech (#3900)

* add AudioSignal && util

* fix codestyle

* add basemodel && decorator

* add util && add quality

* add acc && data && transforms

* add utils

* fix dir

* add *.py; wo unitest

* add unitest

* fix codestyle

* fix cuda error

* add readme && __all__

* add 2 file test

* change download dir

* fix CI download path

* add tar -zxvf

* change requirements path

* add audiotools path

* fix place error

* fix paddle2.5 verion Q

* FFTConv1d -> FFTConv1D

* FFTConv1d -> FFTConv1D

* mv unfold

* add _unfold1d 2 loudness

* fix stupid device variable

* bias -> bias_attr

* () -> []

* fix .to()

* rm ✅

* fix exp

* deepcopy -> clone

* fix dim error

* fix slice && tensor.to

* fix paddle2.5 index bug

* git rm std

* rm comment && ✅

* rm some useless comment

* add __all__

* fix codestyle

* fix soundfile.info error

* fix sth

* add License

* fix cycle import

* Adapt to paddle3.0 && update readme

* fix License

* fix License

* rm duplicate requirements

* fix trasform problems

* rm disp

* Update test_transforms.py

* change path

* rm notebook && add audio path

* rm import

* add comment

* fix cycle import && rm TYPE_CHECKING

* rm IPython

* rm sth useless

* rm uesless deps

* Update requirements.txt

Author: DrRyanHuang

audio/audiotools/README.md

@@ -0,0 +1,68 @@
+Audiotools is a comprehensive toolkit designed for audio processing and analysis, providing robust solutions for audio signal processing, data management, model training, and evaluation.
+
+### Directory Structure
+
+```
+.
+├── audiotools
+│   ├── README.md
+│   ├── __init__.py
+│   ├── core
+│   │   ├── __init__.py
+│   │   ├── _julius.py
+│   │   ├── audio_signal.py
+│   │   ├── display.py
+│   │   ├── dsp.py
+│   │   ├── effects.py
+│   │   ├── ffmpeg.py
+│   │   ├── loudness.py
+│   │   └── util.py
+│   ├── data
+│   │   ├── __init__.py
+│   │   ├── datasets.py
+│   │   ├── preprocess.py
+│   │   └── transforms.py
+│   ├── metrics
+│   │   ├── __init__.py
+│   │   └── quality.py
+│   ├── ml
+│   │   ├── __init__.py
+│   │   ├── accelerator.py
+│   │   ├── basemodel.py
+│   │   └── decorators.py
+│   ├── requirements.txt
+│   └── post.py
+├── tests
+│   └── audiotools
+│       ├── core
+│       │   ├── test_audio_signal.py
+│       │   ├── test_bands.py
+│       │   ├── test_display.py
+│       │   ├── test_dsp.py
+│       │   ├── test_effects.py
+│       │   ├── test_fftconv.py
+│       │   ├── test_grad.py
+│       │   ├── test_highpass.py
+│       │   ├── test_loudness.py
+│       │   ├── test_lowpass.py
+│       │   └── test_util.py
+│       ├── data
+│       │   ├── test_datasets.py
+│       │   ├── test_preprocess.py
+│       │   └── test_transforms.py
+│       ├── ml
+│       │   ├── test_decorators.py
+│       │   └── test_model.py
+│       └── test_post.py
+
+```
+
+- **core**: Contains the core class AudioSignal, which is responsible for the fundamental representation and manipulation of audio signals.
+
+- **data**: Primarily dedicated to storing and processing datasets, including classes and functions for data preprocessing, ensuring efficient loading and transformation of audio data.
+
+- **metrics**: Implements functions for various audio evaluation metrics, enabling precise assessment of the performance of audio models and processing algorithms.
+
+- **ml**: Comprises classes and methods related to model training, supporting the construction, training, and optimization of machine learning models in the context of audio.
+
+This project aims to provide developers and researchers with an efficient and flexible framework to foster innovation and exploration across various domains of audio technology.

audio/audiotools/__init__.py

@@ -0,0 +1,25 @@
+# Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+from . import metrics
+from . import ml
+from . import post
+from .core import AudioSignal
+from .core import highpass_filter
+from .core import highpass_filters
+from .core import Meter
+from .core import STFTParams
+from .core import util
+from .data import datasets
+from .data import preprocess
+from .data import transforms

audio/audiotools/core/__init__.py

@@ -0,0 +1,28 @@
+# Copyright (c) 2025 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+from . import util
+from ._julius import fft_conv1d
+from ._julius import FFTConv1D
+from ._julius import highpass_filter
+from ._julius import highpass_filters
+from ._julius import lowpass_filter
+from ._julius import LowPassFilter
+from ._julius import LowPassFilters
+from ._julius import pure_tone
+from ._julius import resample_frac
+from ._julius import split_bands
+from ._julius import SplitBands
+from .audio_signal import AudioSignal
+from .audio_signal import STFTParams
+from .loudness import Meter