neverlib 0.2.8__py3-none-any.whl → 0.3.0__py3-none-any.whl

neverlib/utils/utils.py

@@ -16,14 +16,14 @@ import numpy as np
 EPS = np.finfo(float).eps
 
 
-def get_path_list(source_path, end="*.wav", shuffle=False):
+def get_path_list(source_dir, end="*.wav", shuffle=False):
     wav_list = []
-    for root, dirnames, filenames in os.walk(source_path):
+    for root, dirnames, filenames in os.walk(source_dir):
         # 实现列表特殊字符的过滤或筛选,返回符合匹配“.wav”字符列表
         for filename in fnmatch.filter(filenames, end):
             wav_list.append(os.path.join(root, filename))
     if os.environ.get("LOCAL_RANK", "0") == "0":
-        print(source_path, len(wav_list))
+        print(source_dir, len(wav_list))
     if shuffle:
         random.shuffle(wav_list)
     return wav_list
@@ -63,26 +63,6 @@ def get_audio_segments(sentence_len,
     return wav
 
 
-def rename_files_and_folders(directory, replace='_-', replacement='_'):
-    # 将路径的指定字符替换为指定字符
-    for root, dirs, files in os.walk(directory):
-        for filename in files:
-            if replace in filename:
-                new_filename = filename.replace(replace, replacement)
-                old_path = os.path.join(root, filename)
-                new_path = os.path.join(root, new_filename)
-                os.rename(old_path, new_path)
-                print(f'Renamed file: {old_path} -> {new_path}')
-
-        for folder in dirs:
-            if replace in folder:
-                new_folder = folder.replace(replace, replacement)
-                old_path = os.path.join(root, folder)
-                new_path = os.path.join(root, new_folder)
-                os.rename(old_path, new_path)
-                print(f'Renamed folder: {old_path} -> {new_path}')
-
-
 def get_file_time(file_path):
     # 获取最后修改时间
     mod_time = os.path.getmtime(file_path)
@@ -174,43 +154,13 @@ def TrainValTestSplit(dataset_dir,
     )
 
 
-def get_leaf_folders(directory):
-    # 获取最底层的文件夹路径
-    leaf_folders = []
-    for root, dirs, _ in os.walk(directory):
-        if not dirs:  # 如果当前文件夹没有子文件夹
-            leaf_folders.append(root)
-    return leaf_folders
-
-
-def del_empty_folders(path):
-    """递归删除空文件夹(先删除子文件夹, 再删除父文件夹)"""
-    if not os.path.isdir(path):
-        return
-
-    # 获取子文件夹
-    subfolders = [
-        os.path.join(path, d) for d in os.listdir(path)
-        if os.path.isdir(os.path.join(path, d))
-    ]
-
-    # 递归处理子文件夹
-    for subfolder in subfolders:
-        del_empty_folders(subfolder)
-
-    # 如果文件夹为空，则删除
-    if not os.listdir(path):
-        os.rmdir(path)
-        print(f"删除空文件夹: {path}")
-
-
 def DatasetSubfloderSplit(source_dir, split_dirs, percentage=None):
     """
     将一个数据集按照子文件夹数量分割成train/val/test数据集
     Args:
         source_dir (str): 源数据集目录
         split_dirs (list): 目标目录列表, 如 [train_dir, val_dir] 或 [train_dir, val_dir, test_dir]
-        percentage (list, optional): 分割比例, 如 [0.9, 0.1] 或 [0.8, 0.1, 0.1]。默认为 None, 此时: 
+        percentage (list, optional): 分割比例, 如 [0.9, 0.1] 或 [0.8, 0.1, 0.1]. 默认为 None, 此时: 
             - 如果是两路分割, 默认为 [0.9, 0.1]
             - 如果是三路分割, 默认为 [0.8, 0.1, 0.1]
     Example:
@@ -239,7 +189,7 @@ def DatasetSubfloderSplit(source_dir, split_dirs, percentage=None):
 
     # 验证输入参数
     if len(split_dirs) not in [2, 3]:
-        raise ValueError("只支持2路或3路分割（训练集/验证集 或 训练集/验证集/测试集）")
+        raise ValueError("只支持2路或3路分割(训练集/验证集 或 训练集/验证集/测试集)")
     if len(percentage) != len(split_dirs):
         raise ValueError("分割比例数量必须与目标目录数量相同")
     if sum(percentage) != 1.0:
@@ -282,30 +232,6 @@ def DatasetSubfloderSplit(source_dir, split_dirs, percentage=None):
         print(f"{split_name.capitalize()} folders: {len(folders)}")
 
 
-def pcm2wav(pcm_path, wav_path, sr=16000, channels=1, subtype='PCM_16'):
-    """
-    将pcm文件转换为wav文件
-    :param pcm_path: pcm文件路径
-    :param wav_path: wav文件路径
-    :param sr: 采样率
-    :param channels: 声道数
-    :param subtype: 子类型
-    """
-    pcm_data = np.fromfile(pcm_path, dtype=np.int16)
-    pcm_data = pcm_data.reshape(-1, channels)  # 支持多通道
-    sf.write(wav_path, pcm_data, sr, subtype=subtype)
-
-
-def wav2pcm(wav_path, pcm_path):
-    """
-    将wav文件转换为pcm文件
-    :param wav_path: wav文件路径
-    :param pcm_path: pcm文件路径
-    """
-    data, _ = sf.read(wav_path, dtype='int16')
-    data.tofile(pcm_path)
-
-
 def save_weight_histogram(model,
                           save_dir,
                           mode=["params", "buffers"],
@@ -387,16 +313,107 @@ def save_weight_histogram(model,
             plt.close()
 
 
-def read_pcm(file_path, channels=5, sample_rate=16000, sample_width=2):
-    # Read raw binary data from the PCM file
-    with open(file_path, 'rb') as f:
-        raw_data = f.read()
-
-    # Convert binary data to numpy array of the correct dtype
-    audio_data = np.frombuffer(raw_data, dtype=np.int16)
-
-    # Reshape the data into a 2D array with shape (num_frames, channels)
-    num_frames = len(audio_data) // channels
-    audio_data = audio_data.reshape((num_frames, channels))
+def merge_tensorboard(event_dir, start_step: int = None, skip_item: str = None):
+    try:
+        from torch.utils.tensorboard import SummaryWriter
+        from tensorboard.backend.event_processing import event_accumulator
+        from collections import defaultdict
+    except ImportError:
+        raise ImportError("请安装torch和tensorboard")
+
+    event_path_list = get_path_list(event_dir, "*events.out.tfevents.*")
+
+    # 1. 汇总所有数据
+    all_data = defaultdict(list)
+    for event_path in event_path_list:
+        print(f"Loading {os.path.basename(event_path)}")
+        ea = event_accumulator.EventAccumulator(event_path)
+        ea.Reload()
+        tags = ea.scalars.Keys()
+        for tag in tags:
+            all_data[tag].extend(ea.scalars.Items(tag))
+
+    # 2. 写入数据
+    writer = SummaryWriter(log_dir=event_dir)
+    for tag, items in all_data.items():
+        unique_items = {item.step: item for item in sorted(items, key=lambda x: x.wall_time)}   # 按wall_time排序
+        sorted_items = sorted(unique_items.values(), key=lambda x: x.step)  # 按step排序
+
+        if skip_item is not None and skip_item in tag:
+            continue
+        for item in sorted_items:
+            if start_step is not None and item.step < start_step:
+                continue
+            writer.add_scalar(tag, item.value, item.step, walltime=item.wall_time)
+
+    # 确保数据写入磁盘
+    writer.flush()
+    writer.close()
+    print("所有数据已写入TensorBoard, 请刷新TensorBoard页面查看")
+    for event_path in event_path_list:
+        os.remove(event_path)
+
+
+def log_time(log_path):
+    import re
+    from datetime import datetime, timedelta
+    
+    with open(log_path, "r", encoding="utf-8") as f:
+        log_data = f.read()
+
+    # 使用正则表达式提取 "train" 前面的时间戳
+    train_times = re.findall(r'(\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2},\d{3}) - train', log_data)
+
+    # 将字符串转换为 datetime 对象
+    train_times_dt = [datetime.strptime(time, '%Y-%m-%d %H:%M:%S,%f') for time in train_times]
+    # train_times_dt = train_times_dt[10:]
+
+    # 计算每段时间之间的时间差
+    for i in range(1, len(train_times_dt)):
+        time_diff = train_times_dt[i] - train_times_dt[i - 1]
+        print(f"时间间隔 {i}: {time_diff}")
+    # train_times_dt=  train_times_dt[-10:]
+    # 计算每段时间之间的时间差, 并存储在列表中
+    time_diffs = [train_times_dt[i] - train_times_dt[i - 1] for i in range(1, len(train_times_dt))]
+
+    # 计算时间间隔的总和
+    total_time_diff = sum(time_diffs, timedelta())
+
+    # 计算平均时间间隔
+    average_time_diff = total_time_diff / len(time_diffs)
+
+    # 输出平均时间间隔
+    print(f"平均时间间隔: {average_time_diff}")
+
+
+# 删除重复的路径
+def del_duplicate_path(file_path_list):
+    for file_path in file_path_list:
+        file_dir, file_name = os.path.dirname(file_path), os.path.basename(file_path)
+        itemID = file_dir.split("/")[-1]
+        if itemID in file_name:
+            file_name_new = file_name.replace(itemID + "_", "")
+            os.rename(file_path, os.path.join(file_dir, file_name_new))
+
+
+def remove_parent_dir_prefix(wav_path_list):
+    for wav_path in wav_path_list:
+        path = Path(wav_path)
+        
+        if not path.exists():
+            continue
+
+        # path.parent.name 自动获取父目录名
+        parent_name = path.parent.name
+        prefix = parent_name + "_"
+        
+        if path.name.startswith(prefix):
+            # 移除前缀
+            new_name = path.name.replace(prefix, "", 1) # 只替换第一个匹配项
+            target_path = path.with_name(new_name)
+            
+            if not target_path.exists():
+                path.rename(target_path)
+            else:
+                print(f"Target exists: {target_path}")
 
-    return audio_data

neverlib/vad/README.md

@@ -40,16 +40,16 @@ mode 0~3
 
 0: 最低的语音检测敏感度, 
 
-- 认为背景噪声不是语音, 适合环境较安静, 背景噪声少的情况。
-- 适合环境较安静, 背景噪声少的情况。
+- 认为背景噪声不是语音, 适合环境较安静, 背景噪声少的情况. 
+- 适合环境较安静, 背景噪声少的情况. 
 
 3: 最高的语音检测敏感度, 
 
-- VAD 会非常积极地尝试将任何噪声过滤掉, 只有明确的语音才会被认为是语音。
-- 适合环境较吵, 背景噪声多的情况。
+- VAD 会非常积极地尝试将任何噪声过滤掉, 只有明确的语音才会被认为是语音. 
+- 适合环境较吵, 背景噪声多的情况. 
 
 ## whisper
 
-whisper 检测的 词与词之间的VAD 都是连着的。但其实音频不是
+whisper 检测的 词与词之间的VAD 都是连着的. 但其实音频不是
 
 而且Whisper的VAD并没有直接提供调参接口, 所以无法调整VAD的参数

neverlib/vad/__init__.py

@@ -4,7 +4,7 @@
 from typing import TYPE_CHECKING, Any
 
 if TYPE_CHECKING:
-    # 仅在类型检查时导入，提供IDE补全支持
+    # 仅在类型检查时导入, 提供IDE补全支持
     from .PreProcess import NS, NS_test, pre_emphasis
     from .VAD_Energy import EnergyVad_C
     from .VAD_Silero import Silero_VAD_C
@@ -68,4 +68,4 @@ if TYPE_CHECKING:
         'from_vadArray_to_vadEndpoint',
         'vad2nad',
         'vad_smooth',
-    ]
+    ]

neverlib/vad/utils.py

@@ -38,7 +38,7 @@ def vad2nad(vad, total_length):
     """根据语音时间戳, 提取噪声时间戳 (优化版)
     Args:
         vad: [{start:xxx, end:xxx}, ...]
-        total_length: 音频总长度（样本数）
+        total_length: 音频总长度(样本数)
     Returns:
         nad: [{start:xxx, end:xxx}, ...] 噪声时间戳列表
     """

{neverlib-0.2.8.dist-info → neverlib-0.3.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: neverlib
-Version: 0.2.8
+Version: 0.3.0
 Summary: A successful sign for python setup
 Author-email: "Never.Ling" <1786088386@qq.com>
 License: MIT
@@ -47,9 +47,17 @@ Requires-Dist: neverlib[vad]; extra == "all"
 Requires-Dist: neverlib[gpu]; extra == "all"
 Dynamic: license-file
 
+<!--
+ * @Author: never.ling never.ling@anker-in.com
+ * @Date: 2025-11-04 17:10:57
+ * @LastEditors: never.ling never.ling@anker-in.com
+ * @LastEditTime: 2025-12-15 11:47:03
+ * @FilePath: /neverlib/README.md
+ * @Description: 这是默认设置,请设置`customMade`, 打开koroFileHeader查看配置 进行设置: https://github.com/OBKoro1/koro1FileHeader/wiki/%E9%85%8D%E7%BD%AE
+-->
 # NeverLib
 
-一个用于音频处理和VAD（语音活动检测）的Python工具库。
+一个用于音频处理和VAD(语音活动检测)的Python工具库. 
 
 ## 安装
 
@@ -113,9 +121,15 @@ from neverlib import utils
 from neverlib.message import seed_QQEmail
 ```
 
+## TODO List
+- [ ]  客观指标添加 [DNSMOSPro ](https://github.com/fcumlin/DNSMOSPro)
+- [ ]  github界面的自我介绍
+- [ ]  https://github.com/nathanhaigh/parallel-rsync
+
+
 ## 许可证
 
-本项目采用MIT许可证。详情请参阅LICENSE文件。
+本项目采用MIT许可证. 详情请参阅LICENSE文件. 
 
 ## 作者
 

neverlib-0.3.0.dist-info/RECORD

@@ -0,0 +1,120 @@
+neverlib/__init__.py,sha256=hPhtK8OM3gp9Jlu8VRP1etTi4-iOCCf3jlmQ4gS85wE,2041
+neverlib/Docs/audio_aug/test_volume.ipynb,sha256=mSTB_1R2b8t5IEiifpTuyV-HKOmW4Q2Z7QMgeRkp2B8,18366
+neverlib/Docs/audio_aug_test/test_snr.py,sha256=1BRUzTYL37pblvhHNB0ORtrbNyxTFWytAxlJ8CDSwfM,1317
+neverlib/Docs/audio_aug_test/test_volume.ipynb,sha256=ZKDPm7Dr8ooZvO6nItshiVIY7jkh2a5AcmUw4wqsB0k,1843460
+neverlib/Docs/audio_aug_test/test_volume.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+neverlib/Docs/filter/biquad.ipynb,sha256=vSAbUSvPUuCEtrO1tHzEyrtuRW0s_bCG38ZUornOVck,3087
+neverlib/Docs/filter/filter_family.ipynb,sha256=cisPF1iMUAArSpUxuzzqsuCywCslLMozFeZtDb5VUfI,1135210
+neverlib/Docs/filter/highpass.ipynb,sha256=-sT6-xH9zV7mR6wWZ9Jnqt9lbUEajGafbhGvUVY98IY,40181
+neverlib/Docs/filter/scipy_filter_family.ipynb,sha256=dLhf4BG3KF6jlfiPKeq6Rl5DrUnx0jDx74an4J7ktGk,79092
+neverlib/Docs/vad/VAD_Energy.ipynb,sha256=qqw6_YwRGAbMcU2_F-1XXomU-y42HDY9WYCMxQVvD58,317091
+neverlib/Docs/vad/VAD_Silero.ipynb,sha256=LREAdVkgdKW_XB5N6SFkHHCRqTkkQxigWOFH9Vhxops,303837
+neverlib/Docs/vad/VAD_WebRTC.ipynb,sha256=jK_L2_JuKGoVX0xQL5UKJzxHnc9ndLYwA6DHJSO-egs,324759
+neverlib/Docs/vad/VAD_funasr.ipynb,sha256=-4KVReX7IqKfufBHk0oTEYNfzaqlJE5mDuYC8TGhv7U,251035
+neverlib/Docs/vad/VAD_rvADfast.ipynb,sha256=YszYIEG_qVUU2hzvbraY4y5fdMToT--50Za1A3IBozc,252432
+neverlib/Docs/vad/VAD_statistics.ipynb,sha256=wT0k_91YUpLLIt6LZMKrjvJLk4kz8g-fZdBvVH07O2k,346246
+neverlib/Docs/vad/VAD_tenVAD.ipynb,sha256=w8oOBVeabW44Fw7TF6CSxIqa3u1gGydoMuWSxE7slaA,351893
+neverlib/Docs/vad/VAD_vadlib.ipynb,sha256=m55oKqOGeMOSgmQlbMkOR6zVh5_Hn6SIocHi6mOtg1k,300998
+neverlib/Docs/vad/VAD_whisper.ipynb,sha256=MsHzjFiD6sLGiyOQeVM4ec9HHfUwdMN-HP4c08-Q7xo,364205
+neverlib/LLM/__init__.py,sha256=jXxWbqG9AXcBaSMj1m29FkDgfzMX55JcJD2xsiCMXe0,909
+neverlib/LLM/bailian.py,sha256=GrUix-OyM1HqVQOQx4HMYuXwEdEd8znUucXJROwQSms,11673
+neverlib/LLM/image.py,sha256=kI1hzGBToLygbsfWnMJjSnrEEONKjV_qU-WBswZHoFk,2501
+neverlib/LLM/text.py,sha256=2G4NT4sbbcwyVCyHQ7jvAy9EbToDZUaUlpogvX5Kd7c,1003
+neverlib/QA/ImpactNoiseRejection.py,sha256=8kCNR-O8_2YtLBAAuh01SFfoptJtHFuliHxhcTBe73Q,3877
+neverlib/QA/gen_init.py,sha256=pVXJoThVMIygijkPIYg6wHcDiCwWAkoioFP7qzY7Flc,7993
+neverlib/QA/get_fun.py,sha256=-UP2FOm3jB_138ntpbsW0eG1S0tJ0d90NqI6f95ryYg,541
+neverlib/QA/html2markdown.py,sha256=ZiXvpwcJ_QBhIR6aetaMOmktWHtd_F7Ed07Deo3qM4c,669
+neverlib/audio_aug/HarmonicDistortion.py,sha256=6VlM7YHwM-Bg-f6bNLsuIIXELRrbbfu3TEySorM5KPo,2684
+neverlib/audio_aug/README.md,sha256=hoYtGNUSMXrPA3GUs1TZVwELxTf7nuQRyt_bpW6F95I,130
+neverlib/audio_aug/TFDrop.py,sha256=HDLf5qfqMT73-dxKder0v0AT7b0FYsSlexbbh3sL2iM,987
+neverlib/audio_aug/TFMask.py,sha256=Fakx5YTj88bZEdj_w3aH9Z9AhEGzGq01TZcnjF4v4BU,1769
+neverlib/audio_aug/__init__.py,sha256=t-m5D4u7gZMEE7xNGDi4hZMi7HINCk7DQ23c2JRToNY,3311
+neverlib/audio_aug/audio_aug.py,sha256=h_UjBnTvX7hVeRMDsQbyAIOyMLft-GWMRenGPfjFmdU,16607
+neverlib/audio_aug/clip_aug.py,sha256=xZi89AtdOgUpX7W4x4_UiZFeK-oCWktueoWSEAY19hU,1041
+neverlib/audio_aug/coder_aug.py,sha256=BPoEg96EDRMzK6cvkYrWTL6dOyX6YKpjbx5nURZcOWg,8314
+neverlib/audio_aug/coder_aug2.py,sha256=EVhzzYPSEUq-XYiu_jGaDLK2PtxgK0jFXsPPSzk2-ac,4713
+neverlib/audio_aug/loss_packet_aug.py,sha256=pQu3PIJVpFJVyyxs7dmLG0N3GBMiQDAAtjmIdl7CaTw,4824
+neverlib/audio_aug/quant_aug.py,sha256=q9BehCXz7HZhiZFW35iZ0fPChEdo-dE8ArJ-DJMDtA8,2601
+neverlib/data/000_short_enhance.wav,sha256=TUZSjHareIY3BFlgZiDuffzirpxnEKagnQDVwDYLdgk,216356
+neverlib/data/3956_speech.wav,sha256=wehCr85ITMOmkQWi1DnjRzb2ggC-XJUftZ3_9sARvS8,2977476
+neverlib/data/3956_sweep.wav,sha256=DoLnlIkYGSuraIE7nWK4nAvmgGWWW4VuhI7xDoMrLys,813832
+neverlib/data/vad_example.wav,sha256=xvcEvjmdjeZO-1fT7u0m0EuiJAqeXUWjBWewDWd8guY,216812
+neverlib/data/white.wav,sha256=JJ3aaPx8v7WakYU7U--m0kO1qUkDSMOrOrPKi6RqAA0,325086
+neverlib/data/white_EQ.wav,sha256=UN1W0hH_4JInnye9vIsQlagVXM-rnJPd1j7A_RIia8c,325086
+neverlib/data/white_matched.wav,sha256=h9vD-2q1eKw9s8pVUoIeXXjeKEC-VRiNABCY-CYcUQ0,115514
+neverlib/data_analyze/README.md,sha256=UoqrTIaMQiPQvAYrOk7Trw53YQNQXn2Vl6DSTEp3mos,6050
+neverlib/data_analyze/__init__.py,sha256=f992HttF5K3CActULLsJPlEIQKkzEOWnCllOIhHezw0,2779
+neverlib/data_analyze/dataset_analyzer.py,sha256=spB6zWru0oW8qP_zDGxkuj43cdf12izfN_BCSTS5CmU,21894
+neverlib/data_analyze/quality_metrics.py,sha256=Rv4xmxdQZDc0oUAPg8ec8a2RoobPa7DcqHnad7PdcFc,11678
+neverlib/data_analyze/rms_distrubution.py,sha256=EpK9s7WJLtJHXd1xD8SvuiyPVTguouYVHyS7Lha2Qag,1526
+neverlib/data_analyze/spectral_analysis.py,sha256=-QZGSvopovLaSh7OcJLYKFH9YLIp_THXUEV3uOreHz0,5894
+neverlib/data_analyze/statistics.py,sha256=2yCUZ5-B1vzocl-X3NYC5lBk3kshNp9jHyRrktYaIls,12953
+neverlib/data_analyze/temporal_features.py,sha256=qrxA0Uh-DxRPbuILQle66AAgv4SgwiLoPTIR9Q2_Jz4,3065
+neverlib/data_analyze/visualization.py,sha256=uvHO_PnzlaKtSFoUhDhRZyqR9uxSBM71PnBqBBPcGCA,16435
+neverlib/filter/README.md,sha256=AYbW2v5Z4WVA6CFwQ11EuR9GYjzoentQPWGh48htK7M,3792
+neverlib/filter/__init__.py,sha256=5vVwbjUitVG9aBZ13qWsXVt7QcG8d0k4uWz6fmUPdvc,1017
+neverlib/filter/biquad.py,sha256=_9hJWI3TY9zEdMkkOWAFJlbyJ4Rm46teezVt90o9fGM,1434
+neverlib/filter/common.py,sha256=9FD3VKwRpe_IdQujt2AJjlMJ_WR-YNEY8OmhOQUDDO8,1953
+neverlib/filter/core.py,sha256=JGNZugcMu9y1VLt5o3a7E-UDiI4PMUi9-NikNHR7i1s,13152
+neverlib/filter/auto_eq/README.md,sha256=Bhu9QthugB-RvEFZKYnv2C6_BYg43o1t1Yp_WBBWmg8,4966
+neverlib/filter/auto_eq/__init__.py,sha256=KOaiZjGti3SQsS0ogKqlxNR4jNACHICeFBVT_8ZBIU0,2188
+neverlib/filter/auto_eq/de_eq.py,sha256=AhAvSb8mMy2IyYDquERkphTzGRVvPv8bNluybTN427Y,16410
+neverlib/filter/auto_eq/freq_eq.py,sha256=1b5hnoqUy1zxVd2WSNqhpokAgJo0xTRwR5AyVPaMr3E,3494
+neverlib/filter/auto_eq/ga_eq_advanced.py,sha256=MCEujrRwPXn9Heu6RYTUBJuEYFDYNrkarvBJTo_WHiw,23450
+neverlib/filter/auto_eq/ga_eq_basic.py,sha256=QiA0Tp_rjD3hgA8oaUh2u8o8Co1FobKs_5ojgEuL2BE,15427
+neverlib/metrics/README.md,sha256=n6Z4AgrRVOhQfxz8wT1i-oHWFuSvUCwzCcZI6FXO4HI,621
+neverlib/metrics/__init__.py,sha256=-547hL2cr5d-EyLhbKBom0fH_ONdfKOF8XGMOaAoiwA,1626
+neverlib/metrics/dnsmos.py,sha256=BGCNoWI3gCNPPPyMWSVyWqqGYQk-6agJA7CVnWyZJ_k,4582
+neverlib/metrics/lpc_lsp.py,sha256=Bd8ngYbdiPekkSnuc57wRF8bPZLRFhvzsxXdsQshyT0,3686
+neverlib/metrics/snr.py,sha256=Oopu9jKEAeoBL7v0ClAPD20XTnAoM8UG9ey33NVhjkk,6355
+neverlib/metrics/spec.py,sha256=rvr7zXi5JnU0rKZn49Pa_M84atIaYlDLSzTgZOQ7YsA,5415
+neverlib/metrics/test_pesq.py,sha256=NldkQ_s99L-p747046gMtIUqc8fdYYhhH8jmvLmRFus,1207
+neverlib/metrics/time.py,sha256=MSUdaT3WvFRnL_wvjZEGjwcVrsy7KwArz3hLkYcXF7s,1292
+neverlib/metrics/DNSMOS/bak_ovr.onnx,sha256=XzNckJlGGBUBkqZWpHS8-KnLztvEeWVJS6jaeWBdEwg,742375
+neverlib/metrics/DNSMOS/model_v8.onnx,sha256=kkZIDFhWe8av_UIAk4537vSUaMi8ftN3bRCcB0VvbpE,224860
+neverlib/metrics/DNSMOS/sig.onnx,sha256=0vvbKTvCNm37rit0d8SQ-YHSSotEBe_TwReHVpxlSdc,742203
+neverlib/metrics/DNSMOS/sig_bak_ovr.onnx,sha256=Jp--vbUTqiPN37tZNULsxUAoSpGEmsUFFocOGsePbt0,1157965
+neverlib/metrics/pDNSMOS/sig_bak_ovr.onnx,sha256=njoZdEnKIXfwmXr-w71riQEXzi8XuJ1u6n-g1HJyyBw,1157962
+neverlib/metrics/pesq_c/PESQ,sha256=f8qbES1y_94XvHgddOpVqOu0aJBym0z2FXvwq1egBGQ,89400
+neverlib/metrics/pesq_c/dsp.c,sha256=N81r8Hdt-xCUW69ALvECknZHThEhah4okjnxIi1T8Ug,15233
+neverlib/metrics/pesq_c/dsp.h,sha256=I1mOP3JsTNqyEdhI3W3FRBAOltDJnJpXG0ShXbVXM2E,5457
+neverlib/metrics/pesq_c/pesq.h,sha256=wkqE2Vf6EB4QnP16c8jXM-cTBO2SVvmc9JiRNvbHLNs,9430
+neverlib/metrics/pesq_c/pesqdsp.c,sha256=YlGSRes0qJgCyPwe1Zb5bShjGQnDzRE08Ecpv5y0A1E,33879
+neverlib/metrics/pesq_c/pesqio.c,sha256=S8PLwBfx0oHHVLfZ1jUsGbwwMniFhHNNV8X6bcUxFYU,12930
+neverlib/metrics/pesq_c/pesqmain.c,sha256=NaANjEDJrCFwJBbXErX51ZKtE86SHKdrXLhS0P3TbT0,27322
+neverlib/metrics/pesq_c/pesqmod.c,sha256=9stYhCLUShFAd6Zc85nQn57RUKDbS2Q8MrhQFB_-ySA,52999
+neverlib/metrics/pesq_c/pesqpar.h,sha256=XZkwXcwycL8ekK6LQy_PT9Ih2K0zirSPHCMhGKyJExs,20986
+neverlib/signal_gen/babble_noise_generate.py,sha256=GFPAFyGHRAZiXzIvWEVlbPJZHfdbQBt1wg6dX7cPYzM,4077
+neverlib/tests/__init__.py,sha256=dQb_TDs8wv81tk9Xctx6V8a9QPRVdBPjgz3aDjD1Za4,809
+neverlib/tests/test_imports.py,sha256=Pqkywv553AdpVDPbtNZAZhdW0GyaZgLo54xtE2XGNGQ,546
+neverlib/tests/test_preprocess.py,sha256=GHj2PvlG8chJlV-zpRfy8IYlTXzGkn5heO9ipy6Xe20,1003
+neverlib/tests/test_vad.py,sha256=TjBuB_qPooflrjkkvoy19HySdBo-MiVeb5yTmLGBJvM,415
+neverlib/utils/README.md,sha256=dQWGObpcfC4cDpM0b4zpTY8dC1JwUJQ06VSr-CRrFuI,934
+neverlib/utils/__init__.py,sha256=Gr4WQip3tXuMz9ePEp6C3ug3AWntP2h-GDcIBFHOQ24,3134
+neverlib/utils/audio_split.py,sha256=ww_xpHagTO3vBVgHqTYGwBVWLnOl_4tP--2w97uFXXI,12339
+neverlib/utils/checkGPU.py,sha256=zwROBGeVFrnsXQTcX9iYgd-Zbj0XbRwVqTZei8oznzY,3790
+neverlib/utils/floder.py,sha256=vF44wOfWvGK8-NDCs-CCnzgASCHcxksToIN5K0HDdjo,3805
+neverlib/utils/lazy_expose.py,sha256=IZmexcrg6Bu0RvzZREj-al7TXVl6qEvLDYS4A2NX6sg,1026
+neverlib/utils/lazy_module.py,sha256=RmaBa5dzUsGOQWushSDHtPUZ2IyWXeXi5RsESiFRZXM,3987
+neverlib/utils/message.py,sha256=0pjThmNA8gDnMFf-viGboiSwE-2p5XBZxM3tkmoOxE0,4072
+neverlib/utils/pcm.py,sha256=N50JboAKJK7hIRW7SiXK0MLDsfTun8Xpee7JPJO9BEQ,1240
+neverlib/utils/utils.py,sha256=-EuEHlNHK_IIlXidbzy3Giz7AbD2xfvkJ3CxQXPy1TU,16136
+neverlib/vad/PreProcess.py,sha256=oib4iRfsUXp5U67pKzJF8k53rqjo8NopK_9LW4O8Mqk,2045
+neverlib/vad/README.md,sha256=rhEyxiuM-i78_eTfsbCZGOEoIltWE7TcempePHfp7oQ,1485
+neverlib/vad/VAD_Energy.py,sha256=ErsiDxn4mH_EOiFBRcoBBRqTUsFfEsvEvRELrdZflkw,2135
+neverlib/vad/VAD_Silero.py,sha256=L9_eO94VQcO3TBdOWypJxnjMl5ZZBYTkL5DR_6APfs4,2324
+neverlib/vad/VAD_WebRTC.py,sha256=3A2GY6AoJhdPeXKrYXFpDXF3Ln__CK0yYFKDtMTU6ps,1919
+neverlib/vad/VAD_funasr.py,sha256=oJqU3VOFT-KXmFV2g8kvT64nHIo5-PxcB4tdb-Ap3Bw,1599
+neverlib/vad/VAD_statistics.py,sha256=o310LKoMECBHO6LcqU3vb-uirF6PM2fHbaRautT2w-w,16663
+neverlib/vad/VAD_vadlib.py,sha256=5reZ7blu0BsqqO9WJn8YPIvPQVn8VcFPcJnH1scgEzQ,2228
+neverlib/vad/VAD_whisper.py,sha256=GhBFV-EIS865c-rj2RvM4IUKKG8ZsoD9LCTwrx1Fx2c,1792
+neverlib/vad/__init__.py,sha256=H2Kn5HDwAv4A4QwmOsvQzhOFvcMP3XBGU4OuCBGcGDU,2033
+neverlib/vad/class_get_speech.py,sha256=s1n24I4P4aTFaQSiyXrh_xAmOChuFZ-wWgrYEE6eCHs,2074
+neverlib/vad/class_vad.py,sha256=ao_20-vyzcI8CxEcvcA_-HoRPTiPEtfuKor2_YWEbz0,5212
+neverlib/vad/img.png,sha256=eLlaUgDnJNyleWZwW8sNv3eXaeaGoXwhFUZCCDJh9u8,179778
+neverlib/vad/utils.py,sha256=sEwZtl2Q9sR1FajumpL4XOPAR6A5KT9qH_9F-tnzpYE,3045
+neverlib-0.3.0.dist-info/licenses/LICENSE,sha256=h-U7vhFdQhieyaNCiA_utrRNMk1rjfrBAUJVodgkfsw,1096
+neverlib-0.3.0.dist-info/METADATA,sha256=QkWsOJuRGJAopvbjc3MG1MesXTeDTMlmu1jDPH2Ssag,3367
+neverlib-0.3.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+neverlib-0.3.0.dist-info/top_level.txt,sha256=QqwYFuDiY_iFTz0Kx8av6zKCP_s4M5NiMTNsIXOy8Po,9
+neverlib-0.3.0.dist-info/RECORD,,

neverlib/.claude/settings.local.json

@@ -1,9 +0,0 @@
-{
-  "permissions": {
-    "allow": [
-      "Read(/data03/never/Desktop/neverlib/**)"
-    ],
-    "deny": [],
-    "ask": []
-  }
-}

neverlib/.history/Docs/audio_aug/del_20250827162540.py

@@ -1,2 +0,0 @@
-
-from neverlib.audio_aug.audio_aug import volume_aug_linmax, volume_aug_dbrms, volume_aug_lufs, measure_loudness

neverlib/.history/Docs/audio_aug/del_20250827162541.py

@@ -1,7 +0,0 @@
-'''
-Author: 凌逆战 | Never
-Date: 2025-08-27 16:25:31
-Description: 
-'''
-
-from neverlib.audio_aug.audio_aug import volume_aug_linmax, volume_aug_dbrms, volume_aug_lufs, measure_loudness

neverlib/.history/Docs/audio_aug/del_20250827162606.py

@@ -1,7 +0,0 @@
-'''
-Author: 凌逆战 | Never
-Date: 2025-08-27 16:25:31
-Description: 
-'''
-
-from neverlib.audio_aug import volume_aug_linmax, volume_aug_dbrms, volume_aug_lufs, measure_loudness

neverlib/.history/Docs/audio_aug/del_20250827162637.py

@@ -1,8 +0,0 @@
-'''
-Author: 凌逆战 | Never
-Date: 2025-08-27 16:25:31
-Description: 
-'''
-
-from neverlib.audio_aug import volume_norm
-# volume_aug_linmax, volume_aug_dbrms, volume_aug_lufs, measure_loudness

neverlib/.history/Docs/audio_aug/del_20250827162645.py

@@ -1,8 +0,0 @@
-'''
-Author: 凌逆战 | Never
-Date: 2025-08-27 16:25:31
-Description: 
-'''
-
-from audio_aug import volume_norm
-# volume_aug_linmax, volume_aug_dbrms, volume_aug_lufs, measure_loudness

neverlib/.history/Docs/audio_aug/del_20250827162723.py

@@ -1,9 +0,0 @@
-'''
-Author: 凌逆战 | Never
-Date: 2025-08-27 16:25:31
-Description: 
-'''
-
-from audio_aug import volume_norm
-# volume_aug_linmax, volume_aug_dbrms, volume_aug_lufs, measure_loudness
-from audio_aug import volume_aug_linmax, volume_aug_dbrms, volume_aug_lufs, measure_loudness

neverlib/.history/Docs/audio_aug/del_20250827162739.py

@@ -1,9 +0,0 @@
-'''
-Author: 凌逆战 | Never
-Date: 2025-08-27 16:25:31
-Description: 
-'''
-
-from audio_aug import volume_norm
-# volume_aug_linmax, volume_aug_dbrms, volume_aug_lufs, measure_loudness
-# from audio_aug import volume_aug_linmax, volume_aug_dbrms, volume_aug_lufs, measure_loudness

neverlib/.history/Docs/audio_aug/test_snr_20250806011331.py

@@ -1,75 +0,0 @@
-'''
-Author: 凌逆战 | Never
-Date: 2025-03-24 10:00:14
-Description: 
-'''
-# -*- coding:utf-8 -*-
-# Author:凌逆战 | Never
-# Date: 2024/1/29
-"""
-snr增强, 对安静的测试集加指定snr的白噪
-"""
-import numpy as np
-import random
-import soundfile as sf
-from neverlib.metrics.snr import get_snr
-from neverlib.audio_aug import snr_aug_changeNoise, snr_aug_changeNoise_v2
-from neverlib.utils import EPS
-
-
-def add_noise(clean, snr_range, snr_aug_rate):
-    """
-    Args:
-        clean: (*, C)
-        snr_range: snr范围 [min, max]
-        snr_aug_rate: snr增强率
-    Returns:
-    """
-    if random.random() < snr_aug_rate:
-        snr = random.uniform(snr_range[0], snr_range[1])
-        noise = np.random.randn(*clean.shape)  # 生成和clean等长的白噪
-        noisy = snr_aug_changeNoise(clean, noise, snr)
-    else:
-        noisy = clean
-        snr = 100
-    return noisy, snr
-
-
-def snr_aug1(clean, snr):
-    """ 白噪
-    Args:
-        clean: (*, C)
-        snr: snr值
-    Returns:
-    """
-    noise = np.random.randn(*clean.shape)  # 生成和clean等长的白噪
-
-    noisy, _ = snr_aug_changeNoise(clean, noise, snr)
-
-    return noisy
-
-
-def snr_aug2(clean, snr):
-    """
-    Args:
-        clean: (*, C)
-        snr_range: snr范围 [min, max]
-        snr_aug_rate: snr增强率
-    Returns:
-    """
-    noise = np.random.randn(*clean.shape)  # 生成和clean等长的白噪
-    noisy, _ = snr_aug_changeNoise_v2(clean, noise, snr)
-    return noisy
-
-
-if __name__ == "__main__":
-    clean_path = "../wav_data/TIMIT.wav"
-    clean, fs = sf.read(clean_path, always_2d=True, dtype="float32")
-
-    noisy1 = snr_aug1(clean, 10)
-    noisy2 = snr_aug2(clean, 10)
-    print(clean.shape)
-    print(noisy1.shape, noisy2.shape)
-
-    print(get_snr(clean, noisy1 - clean))  # 10.000000000480982
-    print(get_snr(clean, noisy2 - clean))  # 9.999870642663442