- fix some bugs.

- certain adjustments to the project code.

.gitignore

@@ -2,4 +2,5 @@
 *.npy
 *.nyc
 *.csv
-*.fts
+*.fts
\ No newline at end of file
+*.h5
\ No newline at end of file

README.md

@@ -82,7 +82,7 @@
 
 - 我猜测可能时机器太久没有关机了(平时我都是休眠),导致系统出现了一些错误
 
-  - 机器发生错误是很有可能的,就比如学校图书馆的刷脸系统,验证通过平时屏幕显示绿色,然而最近通过显示的也是红色
+  - 机器发生错误是很有可能的,就比如学校图书馆的刷脸系统,`验证成功`平时的话屏幕显示绿色提醒,然而最近`验证成功`显示的也是红色
   - 而在早期的windows7上,有时候从休眠中回复直接会失败
 
 - 然后我重启机器,发现任务栏多出了个搜索框,系统更新了没有重启可能也造成了一些影响
@@ -831,6 +831,43 @@ print("Accuracy:", clf.score(X_test, y_test))
 
   在使用MLPRegressor时，需要根据具体的数据集和任务需求，选择合适的参数来构建模型。同时，还可以通过交叉验证等技术来评估模型的性能和调整参数，以获得更好的预测结果。需要注意的是，MLPRegressor算法对于数据集的特征缩放和标准化非常敏感，因此在使用MLPRegressor时需要对数据进行预处理。
 
+#### BaggingClassifier
+
+A Bagging classifier.
+
+A Bagging classifier is an ensemble meta-estimator that fits base classifiers each on random subsets of the original dataset and then aggregate their individual predictions (either by voting or by averaging) to form a final prediction. Such a meta-estimator can typically be used as a way to reduce the variance of a black-box estimator (e.g., a decision tree), by introducing randomization into its construction procedure and then making an ensemble out of it.
+
+This algorithm encompasses several works from the literature. When random subsets of the dataset are drawn as random subsets of the samples, then this algorithm is known as Pasting [[1\]](https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.BaggingClassifier.html#rb1846455d0e5-1). If samples are drawn with replacement, then the method is known as Bagging [[2\]](https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.BaggingClassifier.html#rb1846455d0e5-2). When random subsets of the dataset are drawn as random subsets of the features, then the method is known as Random Subspaces [[3\]](https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.BaggingClassifier.html#rb1846455d0e5-3). Finally, when base estimators are built on subsets of both samples and features, then the method is known as Random Patches [[4\]](https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.BaggingClassifier.html#rb1846455d0e5-4).
+
+Bagging分类器是一种元估计器集成方法，它在原始数据集的随机子集上分别拟合基分类器，然后聚合它们的个体预测结果（通过投票或平均）以形成最终预测。这种元估计器通常可以用作一种降低黑盒估计器（例如决策树）方差的方法，通过将随机化引入其构建过程，然后将其组合起来。
+
+此算法包括文献中的几项工作。当随机抽取数据集的子集作为样本的随机子集时，此算法称为Pasting [1]。如果使用放回抽样，则该方法称为Bagging [2]。当随机抽取数据集的子集作为特征的随机子集时，该方法称为随机子空间方法[3]。最后，当基估计器基于样本和特征的子集构建时，该方法称为Random Patches [4]。
+
+In ensemble algorithms, bagging methods form a class of algorithms which build several instances of a black-box estimator on random subsets of the original training set and then aggregate their individual predictions to form a final prediction. These methods are used as a way to reduce the variance of a base estimator (e.g., a decision tree), by introducing randomization into its construction procedure and then making an ensemble out of it. In many cases, bagging methods constitute a very simple way to improve with respect to a single model, without making it necessary to adapt the underlying base algorithm. As they provide a way to reduce overfitting, bagging methods work best with strong and complex models (e.g., fully developed decision trees), in contrast with boosting methods which usually work best with weak models (e.g., shallow decision trees).
+
+Bagging methods come in many flavours but mostly differ from each other by the way they draw random subsets of the training set:
+
+> - When random subsets of the dataset are drawn as random subsets of the samples, then this algorithm is known as Pasting [[B1999\]](https://scikit-learn.org/stable/modules/ensemble.html#b1999).
+> - When samples are drawn with replacement, then the method is known as Bagging [[B1996\]](https://scikit-learn.org/stable/modules/ensemble.html#b1996).
+> - When random subsets of the dataset are drawn as random subsets of the features, then the method is known as Random Subspaces [[H1998\]](https://scikit-learn.org/stable/modules/ensemble.html#h1998).
+> - Finally, when base estimators are built on subsets of both samples and features, then the method is known as Random Patches [[LG2012\]](https://scikit-learn.org/stable/modules/ensemble.html#lg2012).
+
+In scikit-learn, bagging methods are offered as a unified [`BaggingClassifier`](https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.BaggingClassifier.html#sklearn.ensemble.BaggingClassifier) meta-estimator (resp. [`BaggingRegressor`](https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.BaggingRegressor.html#sklearn.ensemble.BaggingRegressor)), taking as input a user-specified estimator along with parameters specifying the strategy to draw random subsets. In particular, `max_samples` and `max_features` control the size of the subsets (in terms of samples and features), while `bootstrap` and `bootstrap_features` control whether samples and features are drawn with or without replacement. When using a subset of the available samples the generalization accuracy can be estimated with the out-of-bag samples by setting `oob_score=True`. As an example, the snippet below illustrates how to instantiate a bagging ensemble of `KNeighborsClassifier` estimators, each built on random subsets of 50% of the samples and 50% of the features.
+
+在集成算法中，bagging方法是一类算法，它在原始训练集的随机子集上构建多个黑盒估计器实例，然后聚合它们的个体预测结果以形成最终预测。这些方法通过将随机化引入其构建过程，并将其组合起来，以降低基本估计器（例如决策树）的方差。在许多情况下，bagging方法是一种非常简单的改进单个模型的方法，而无需调整底层基础算法。由于它们提供了一种减少过拟合的方法，因此bagging方法适用于强大而复杂的模型（例如完全发展的决策树），而提升方法通常适用于弱模型（例如浅层决策树）。
+
+Bagging方法有许多变种，但它们主要通过绘制训练集的随机子集的方式而不同：
+
+当随机抽取数据集的子集作为样本的随机子集时，此算法称为Pasting [B1999]。
+
+如果使用放回抽样，则该方法称为Bagging [B1996]。
+
+当随机抽取数据集的子集作为特征的随机子集时，该方法称为随机子空间方法[ H1998]。
+
+最后，当基估计器基于样本和特征的子集构建时，该方法称为Random Patches [LG2012]。
+
+在scikit-learn中，Bagging方法提供了一个统一的BaggingClassifier元估计器（resp. BaggingRegressor），以用户指定的估计器为输入，并指定绘制随机子集的策略的参数。特别地，max_samples和max_features控制子集的大小（以样本和特征为单位），而bootstrap和bootstrap_features控制是否使用有放回或无放回的方式来抽取样本和特征。当使用可用样本的子集时，可以通过设置oob_score=True来估计外部样本的泛化精度。例如，下面的代码段说明如何实例化一个KNeighborsClassifier估计器的bagging集合，每个估计器都是在50％的样本和50％的特征的随机子集上构建的。
+
 #### BaggingRegressor
 
 BaggingRegressor是一种基于袋装法（Bagging）的回归模型，常用于机器学习中。
@@ -995,11 +1032,13 @@ SVR（Support Vector Regression）是一种基于支持向量机（SVM）的回
   - 使用大纲阅读一个大文件是好办法
   - 对于没有封装在函数或者类中的代码,可以设置`fold level`折叠至level 2来快速把握代码结构
 
-## 部分实验结果
+# 部分实验结果
 
-### 跨库识别
+## 跨库识别
 
-#### angry&sad
+#### 
+
+### angry&sad@emodb-ravdess
 
 - `train_emodb_AS.csv+test_ravdess_AS.csv`
 
@@ -1053,6 +1092,62 @@ SVR（Support Vector Regression）是一种基于支持向量机（SVM）的回
     test_score=0.770935960591133
     ```
 
+### ravdess-emodb
+
+```bash
+(d:\condaPythonEnvs\tf2.10) PS D:\repos\CCSER\SER> py "d:\repos\CCSER\SER\recognizer\basic.py"
+@{model}
+partition='train'
+D:\repos\CCSER\SER\meta_files\train_ravdess_AS.csv @🎈{meta_file}
+[I] Loading audio file paths and its corresponding labels...
+meta_file存在D:\repos\CCSER\SER\meta_files\train_ravdess_AS.csv文件!
+检查特征文件D:\repos\CCSER\SER\features\ravdess_mfcc_AS_1216.npy是否存在...
+self.e_config=['angry', 'sad']
+self.f_config=['mfcc']
+npy文件不存在,尝试创建...
+{} @{self.feature_transforms}🎈
+True @{save_obj}
+Extracting features for partition:: 100%|███████████████████████████████████████████████████████████| 1216/1216 [00:19<00:00, 63.43it/s] 
+fts参数key合法
+🎈🎈🎈特征提取
+(1216, 40) @{feature.shape}
+[Info] Adding  train samples
+partition='test'
+D:\repos\CCSER\SER\meta_files\test_emodb_AS.csv @🎈{meta_file}
+[I] Loading audio file paths and its corresponding labels...
+meta_file存在D:\repos\CCSER\SER\meta_files\test_emodb_AS.csv文件!
+检查特征文件D:\repos\CCSER\SER\features\emodb_mfcc_AS_38.npy是否存在...
+self.e_config=['angry', 'sad']
+self.f_config=['mfcc']
+特征矩阵文件(.npy)已经存在,直接导入:loading...
+(38, 40) @{feature.shape}
+[Info] Adding  test samples
+[I] Data loaded
+
+self.ae=<audio.extractor.AudioExtractor object at 0x0000028D91CE85E0>
+self.ae.pca=None🎈
+Evaluating <SVC>:   0%|                                                                                           | 0/5 [00:00<?, ?it/s]@{model}
+
+[I] SVC with 0.7105263157894737 test accuracy
+Evaluating <RandomForestClassifier>:  20%|████████████▊                                                   | 1/5 [00:05<00:23,  5.80s/it]@{model}
+Evaluating <BaggingClassifier>: 100%|█████████████████████████████████████████████████████████████████████| 5/5 [00:14<00:00,  2.99s/it] 
+[🎈] Best model : RandomForestClassifier with 97.368% test accuracy
+train_score=0.9868421052631579verbose=0               precision    recall  f1-score   support
+
+       angry       1.00      0.92      0.96        24
+         sad       0.88      1.00      0.93        14
+
+    accuracy                           0.95        38
+   macro avg       0.94      0.96      0.94        38
+weighted avg       0.95      0.95      0.95        38 RandomForestClassifier
+test_score=0.9473684210526315
+(1216, 40) (1216,) 🎈
+n_splits=5
+cv_score=0.9508196721311475
+(d:\condaPythonEnvs\tf2.10) PS D:\repos\CCSER\SER>
+```
+
+
 
 #### angry&happy&sad
 

SG/ccser_gui.py

@@ -118,7 +118,7 @@ def get_algos_elements_list(ava_algorithms=ava_algorithms):
                 algo.title(),
                 "algorithm",
                 key=f"{algo}",
-                default=(i == 1),
+                default=(i == 0),
                 enable_events=True,
             )
         )
@@ -360,7 +360,7 @@ def get_file_choose_layout():
             # ],
             # [
             sg.B(
-                "recognize it",
+                "Recognize it",
                 key="recognize it",
                 tooltip=lang["recognize_the_audio_emotion"],
             ),
@@ -482,8 +482,7 @@ def get_logging_viewer_layout():
 
 
 def get_analyzer_layout():
-    analyzer_layout = (
-        [
+    analyzer_log_printer_layout = [
             [bt.h2("Anything printed will display here!")],
             [
                 sg.Multiline(
@@ -499,9 +498,17 @@ def get_analyzer_layout():
                 )
             ],
         ]
-        + dv.layout
-        + q.query_layout
-    )
+    
+    # analyzer_layout = (
+    #     analyzer_log_printer
+    #     + dv.layout
+    #     + q.query_layout
+    # )
+    analyzer_layout = [
+        *analyzer_log_printer_layout,
+        *dv.layout,
+        *q.query_layout,
+    ]
 
     return analyzer_layout
 
@@ -663,12 +670,12 @@ def get_algo_layout():
 def get_e_config_layout():
     emotion_config_checboxes_layout = [
         [
-            sg.Checkbox("angry", key="angry", default=True, enable_events=True),
-            sg.Checkbox("happy", key="happy", enable_events=True),
+            sg.Checkbox("angry", key="angry", default=False, enable_events=True),
+            sg.Checkbox("happy", key="happy", default=True, enable_events=True),
             sg.Checkbox("neutral", key="neutral", default=True, enable_events=True),
             sg.Checkbox("ps", key="ps", enable_events=True),
             sg.Checkbox("sad", key="sad", default=True, enable_events=True),
-            sg.Checkbox("others", key="others", default=True, enable_events=True),
+            sg.Checkbox("others", key="others", default=False, enable_events=True),
         ]
     ]
 
@@ -692,34 +699,6 @@ def get_e_config_layout():
     return e_config_layout
 
 
-tooltip_pca_components = """
-PCA components
-Number of components to keep. if n_components is not set all components are kept:
-
-n_components == min(n_samples, n_features)
-If n_components == 'mle' and svd_solver == 'full', Minka’s MLE is used to guess the dimension. Use of n_components == 'mle' will interpret svd_solver == 'auto' as svd_solver == 'full'.
-
-If 0 < n_components < 1 and svd_solver == 'full', select the number of components such that the amount of variance that needs to be explained is greater than the percentage specified by n_components.
-
-If svd_solver == 'arpack', the number of components must be strictly less than the minimum of n_features and n_samples.
-
-Hence, the None case results in:
-
-n_components == min(n_samples, n_features) - 1
-"""
-tooltip_pca_svd_solver = """
-If auto :
-The solver is selected by a default policy based on X.shape and n_components: if the input data is larger than 500x500 and the number of components to extract is lower than 80% of the smallest dimension of the data, then the more efficient ‘randomized’ method is enabled. Otherwise the exact full SVD is computed and optionally truncated afterwards.
-
-If full :
-run exact full SVD calling the standard LAPACK solver via scipy.linalg.svd and select the components by postprocessing
-
-If arpack :
-run SVD truncated to n_components calling ARPACK solver via scipy.sparse.linalg.svds. It requires strictly 0 < n_components < min(X.shape)
-
-If randomized :
-run randomized SVD by the method of Halko et al.
-"""
 
 
 def get_f_config_layout():
@@ -760,16 +739,17 @@ def get_f_transform_layout():
                 ),
             ],
             [
+                sg.T('n_components:',tooltip="input the number of components to keep."),
                 sg.Input(
                     key=pca_components_key,
-                    default_text="35",
-                    tooltip=tooltip_pca_components,
+                    default_text="None",
+                    tooltip=bt.pca_components_tooltip,
                     enable_events=True,
                 ),
                 sg.Combo(
                     values=ava_svd_solver,
                     default_value="auto",
-                    tooltip=tooltip_pca_svd_solver,
+                    tooltip=bt.pca_svd_solver_tooltip,
                     enable_events=True,
                     key=pca_svd_solver_key,
                 ),
@@ -791,7 +771,7 @@ def get_f_transform_layout():
         ],
     )
     f_transform_layout = [
-        [bt.h2(lang["feature_transfomr_config"])],
+        [bt.h2(lang["feature_transform_config"])],
         [f_transform_frame_layout],
     ]
     return f_transform_layout

SG/constants/beauty.py

@@ -33,6 +33,52 @@ logo = """
 """
 
 
+pca_components_tooltip = """
+PCA components
+Number of components to keep. if n_components is not set all 
+components are kept:
+
+n_components == min(n_samples, n_features)
+If n_components == 'mle' and svd_solver == 'full', Minka’s MLE 
+is used to guess the dimension. Use of n_components == 'mle' will 
+interpret svd_solver == 'auto' as svd_solver == 'full'.
+
+If 0 < n_components < 1 and svd_solver == 'full', select 
+the number of components such that the amount of variance that 
+needs to be explained is greater than the percentage specified 
+by n_components.
+
+If svd_solver == 'arpack', the number of components must
+ be strictly less than the minimum of n_features and n_samples.
+
+Hence, the None case results in:
+
+n_components == min(n_samples, n_features) - 1
+"""
+pca_svd_solver_tooltip = """
+If auto :
+The solver is selected by a default policy based on X.shape and 
+n_components:
+ if the input data is larger than 500x500 and the number of components
+   to extract is lower than 80% of the smallest dimension of the data, 
+   then the more efficient ‘randomized’ method is enabled. Otherwise 
+   the exact full SVD is computed and optionally truncated afterwards.
+
+If full :
+run exact full SVD calling the standard LAPACK solver via scipy.
+linalg.svd and select the components by postprocessing
+
+If arpack :
+run SVD truncated to n_components calling ARPACK solver via 
+scipy.sparse.linalg.svds. It requires strictly
+ 0 < n_components < min(X.shape)
+
+If randomized :
+run randomized SVD by the method of Halko et al.
+"""
+
+
+
 db_introduction = """
 ## SpeechDatabases
 

SG/data_visualization.py

@@ -14,7 +14,7 @@ emotion_counts = df["emotion"].value_counts()
 
 # 创建窗口布局
 layout = [
-    [sg.Text("情感成分分析图表")],
+    [sg.Text("Emotional Component Analysis")],
     [sg.Canvas(key="-CANVAS-")],
     [sg.Button("generate pie graph")],
 ]

SG/demo_programs/Demo_Matplotlib.py

-##!/usr/bin/env python
-from matplotlib.ticker import NullFormatter  # useful for `logit` scale
+##
+# from matplotlib.ticker import NullFormatter  # useful for `logit` scale
+
 import matplotlib.pyplot as plt
 import numpy as np
 from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg
 import PySimpleGUI as sg
 import matplotlib
-matplotlib.use('TkAgg')
+
+matplotlib.use("TkAgg")
 
 """
 Demonstrates one way of embedding Matplotlib figures into a PySimpleGUI window.
@@ -23,82 +25,41 @@ Basic steps are:
 
 
 # ------------------------------- PASTE YOUR MATPLOTLIB CODE HERE -------------------------------
-#
-# # Goal is to have your plot contained in the variable  "fig"
-#
-# # Fixing random state for reproducibility
-# np.random.seed(19680801)
-#
-# # make up some data in the interval ]0, 1[
-# y = np.random.normal(loc=0.5, scale=0.4, size=1000)
-# y = y[(y > 0) & (y < 1)]
-# y.sort()
-# x = np.arange(len(y))
-#
-# # plot with various axes scales
-# plt.figure(1)
-#
-# # linear
-# plt.subplot(221)
-# plt.plot(x, y)
-# plt.yscale('linear')
-# plt.title('linear')
-# plt.grid(True)
-#
-# # log
-# plt.subplot(222)
-# plt.plot(x, y)
-# plt.yscale('log')
-# plt.title('log')
-# plt.grid(True)
-#
-# # symmetric log
-# plt.subplot(223)
-# plt.plot(x, y - y.mean())
-# plt.yscale('symlog', linthreshy=0.01)
-# plt.title('symlog')
-# plt.grid(True)
-#
-# # logit
-# plt.subplot(224)
-# plt.plot(x, y)
-# plt.yscale('logit')
-# plt.title('logit')
-# plt.grid(True)
-# plt.gca().yaxis.set_minor_formatter(NullFormatter())
-# plt.subplots_adjust(top=0.92, bottom=0.08, left=0.10, right=0.95, hspace=0.25,
-#                     wspace=0.35)
-# fig = plt.gcf()
-#
 
 
 fig = matplotlib.figure.Figure(figsize=(5, 4), dpi=100)
-t = np.arange(0, 3, .01)
+t = np.arange(0, 3, 0.01)
 fig.add_subplot(111).plot(t, 2 * np.sin(2 * np.pi * t))
 
 # ------------------------------- END OF YOUR MATPLOTLIB CODE -------------------------------
 
 # ------------------------------- Beginning of Matplotlib helper code -----------------------
 
+
 def draw_figure(canvas, figure):
     figure_canvas_agg = FigureCanvasTkAgg(figure, canvas)
     figure_canvas_agg.draw()
-    figure_canvas_agg.get_tk_widget().pack(side='top', fill='both', expand=1)
+    figure_canvas_agg.get_tk_widget().pack(side="top", fill="both", expand=1)
     return figure_canvas_agg
 
+
 # ------------------------------- Beginning of GUI CODE -------------------------------
 
 # define the window layout
-layout = [[sg.Text('Plot test')],
-          [sg.Canvas(key='-CANVAS-')],
-          [sg.Button('Ok')]]
+layout = [[sg.Text("Plot test")], [sg.Canvas(key="-CANVAS-")], [sg.Button("Ok")]]
 
 # create the form and show it without the plot
-window = sg.Window('Demo Application - Embedding Matplotlib In PySimpleGUI', layout, finalize=True, element_justification='center', font='Helvetica 18')
+window = sg.Window(
+    "Demo Application - Embedding Matplotlib In PySimpleGUI",
+    layout,
+    finalize=True,
+    element_justification="center",
+    font="Helvetica 18",
+)
 
 # add the plot to the window
-fig_canvas_agg = draw_figure(window['-CANVAS-'].TKCanvas, fig)
-
+fig_canvas_agg = draw_figure(window["-CANVAS-"].TKCanvas, fig)
+# display the window
 event, values = window.read()
 
 window.close()

SG/fviewer.py

@@ -106,7 +106,10 @@ def get_audios_regex(
         else:
             path = path.absolute()
         # 对路径进行正则过滤
+        #todo 对括号的识别有问题(得益于模块化,可以直接在这个模块内启动图形界面进行调试)
         if filter_regex:
+            if verbose:
+                print('filter_regex:> ', filter_regex)
             s = re.search(filter_regex, str(path), re.IGNORECASE)
             if s:
                 filtered_audios.append(path)
@@ -226,7 +229,7 @@ audio_viewer_layout = [
     ],
     [
         sg.B(filter_audios_key, tooltip="click to manual refresh the files listbox"),
-        sg.Button(ufg.close),
+        # sg.Button(ufg.close),
     ],
     [sg.Text(f"{len_default_folder_file_list} files", key="num_files_text")],
     [

SG/table_show.py

@@ -32,7 +32,7 @@ class TableShow():
                     # background_color="lightblue",
                     auto_size_columns=True,
                     justification="center",
-                    num_rows=min(25, len(self.data_rows)),
+                    num_rows=min(15, len(self.data_rows)),#每次最多显示15条数据
                     expand_x=True,
                     expand_y=True,
                 )
@@ -43,7 +43,7 @@ class TableShow():
 
         ]
     def run(self):
-        window = sg.Window("result table", self.layout,resizable=True,size=(500,400))
+        window = sg.Window("result table", self.layout,resizable=True,size=(500,500))
         # 事件循环
         while True:
             event, values = window.read()

SG/translations/en.json

@@ -2,7 +2,7 @@
     "welcome_message": "Welcome to My App!",
     "choose_emotion_config": "Please select an emotional combination for testing: recommended combinations are AS, HNS, AHNS, AHNPS. \nNote that there is a difference between 'surprise' and 'pleasantSurprise' in the SAVEE dataset, \nso the AHNPS combination is not recommended for use on SAVEE.",
     "choose_feature_config": "Please choose one or more features",
-    "feature_transfomr_config":"feature transfomer config",
+    "feature_transform_config":"feature transformer config",
     "choose_algorithm": "Choose an algorithm for testing",
     "choose_audio": "Please select an audio sample file to recognize its emotion.",
     "recognize_the_audio_emotion": "Recognize the emotion of the selected audio file.",
@@ -11,5 +11,6 @@
     "welcome_title": "𝒲ℯ𝓁𝒸ℴ𝓂ℯ 𝓉ℴ ℯ𝓍𝓅ℯ𝓇𝒾ℯ𝓃𝒸ℯ 𝒞𝒞𝒮ℰℛ 𝒞𝓁𝒾ℯ𝓃𝓉!",
     "result_training":"result of model training:",
     "train_result_title":"Train Result",
-    "result_frame":"Emotion Of Select File(Predict Result)"
+    "result_frame":"Emotion Of Select File(Predict Result)",
+    "recognize_it":"Recognize it"
 }
\ No newline at end of file

SG/user.py

@@ -134,7 +134,8 @@ class UserAuthenticatorGUI:
                     key="confirm_password", password_char="*", enable_events=True
                 ),
             ],
-            [sg.Submit(button_text="Register"), sg.Cancel(button_text="Cancel")],
+            # sg.Submit默认绑定了enter快捷键,为了避免误操作,需要谨慎使用
+            [sg.Submit(button_text="Register",bind_return_key=False), sg.Cancel(button_text="Cancel")],
         ]
 
         # 定义登录Tab页布局
@@ -152,7 +153,7 @@ class UserAuthenticatorGUI:
                     key=passowrd_login_key, password_char="*", enable_events=True
                 ),
             ],
-            [sg.Submit(button_text="Login"), sg.Cancel(button_text="Cancel")],
+            [sg.Submit(button_text="Login",bind_return_key=False), sg.Cancel(button_text="Cancel")],
         ]
 
         return register_layout, login_layout

audio/extractor.py

+##
+
 import os
-from collections import defaultdict
 import sys
-import ipdb
-from joblib import load
+from collections import defaultdict
+from pathlib import Path
 
+import ipdb
 import numpy as np
 import pandas as pd
 import tqdm
+from joblib import load
 
-from audio.create_meta import create_csv_by_metaname
-from config.EF import AHNPS, HNS, AHNPS_dict, HNS_dict, e_config_def, f_config_def
 import config.MetaPath as mp
-from config.MetaPath import (
-    create_tag_name,
-    features_dir,
-    get_first_letters,
-    train_emodb_csv,
-    validate_partition,
-    ava_dbs,
-    ava_fts_params
-)
 from audio.core import extract_feature_of_audio
+from audio.create_meta import create_csv_by_metaname
+from config.EF import (AHNPS, HNS, AHNPS_dict, HNS_dict, e_config_def,
+                       f_config_def)
+from config.MetaPath import (ava_dbs, ava_fts_params, create_tag_name, emodb,
+                             features_dir, get_first_letters, meta_dir,
+                             project_dir, train_emodb_csv, validate_partition)
 
 # from pathlib import Path
 Series = pd.Series
@@ -100,6 +98,11 @@ class AudioExtractor:
         self.test_features = []
         # 使用字典打包
         self.pca = None
+    def pathlike_to_list(self, meta_paths):
+        if isinstance(meta_paths, str) or isinstance(meta_paths,Path):
+            # print(f"cast the '{meta_paths}' to [str]")
+            meta_paths = [meta_paths]
+        return meta_paths
 
     def get_partition_features(self, partition) -> np.ndarray:
         """将包含若干个二维ndarray的列表vstack成1个二维ndarray
@@ -134,14 +137,19 @@ class AudioExtractor:
 
     def load_metadata(self, meta_files):
         """
-        从meta_files(文件)中读取语料库各条语音的信息;
+        从给定meta_files(文件)路径中读取语料库各条语音的信息;
+        如果需要读取的meta_files不存在,那么尝试解析meta_files(如果meta_files参数是一个符合可解析规范的字符串)
+        这种情况下会调用create_meta模块中的create_csv_by_metaname函数进行meta文件构造
 
-        Read metadata from a  file & Extract and loads features of audio files
+        Read metadata from a  file & Extract meta if according meta_files and loads features of audio files
 
         Parameters
         ----------
         meta_files : list[str]|str
             需要读取的meta文件
+        Return
+        -
+        从meta中读取的信息:包括各语音文件的路径和情感标签
 
         """
         # empty dataframe
@@ -157,12 +165,13 @@ class AudioExtractor:
         # print("meta_files:", meta_files)
 
         # print("type(meta_files)", type(meta_files))
-        if isinstance(meta_files, str):
-            meta_files = [meta_files]
+
+        meta_files = self.pathlike_to_list(meta_files)
         for meta_file in meta_files:
             if not os.path.exists(meta_file):
                 # create_csv_by_meta_name
                 print(f"{meta_file} does not exist,creating...😂")
+
                 create_csv_by_metaname(meta_file, shuffle=self.shuffle)
             else:
                 print(f"meta_file存在{meta_file}文件!")
@@ -326,7 +335,7 @@ class AudioExtractor:
 
         # 尝试计算语料库的名字和情感配置名字
         db = self.fields_parse(meta_path)
-
+        # 特征保存的目录检查(不存在则创建之)
         if not os.path.isdir(self.features_dir):
             os.mkdir(self.features_dir)
 
@@ -378,6 +387,10 @@ class AudioExtractor:
         return features, audio_paths, emotions
 
     def get_features_file_path(self, partition, db, n_samples,ext=""):
+        fts=self.feature_transforms
+
+        if fts is None:
+            self.feature_transforms = {}
         features_file_name = create_tag_name(
             db=db,
             partition=partition,  # 建议保存特征文件时,这个字段置空即可
@@ -474,8 +487,6 @@ class AudioExtractor:
         # 考虑特征预处理
         from sklearn.preprocessing import StandardScaler
 
-        
-
         # X为特征矩阵,y为标签
 
         fts = self.feature_transforms
@@ -508,16 +519,20 @@ class AudioExtractor:
 
             n_components = pca_params_dict.get("n_components")
 
-            if n_components == "None":
-                # pca_params_dict["n_components"] = None
-                n_components=None
-            elif n_components=='mle':
+      
+            if n_components=='mle':
                 pass
-            else:
+            elif isinstance(n_components, int):
+                pass
+            elif n_components and n_components.isdigit():
                 # if n_components.isdigit():
                 # int()函数自带类型错误检测,有非法输入会自动抛出错误,所以这里直接使用,而不去手动检测输入的合法性
                 # pca_params_dict['n_components'] = int(n_components)
                 n_components=int(n_components)
+            # elif n_components == "None":
+            else:
+                # pca_params_dict["n_components"] = None
+                n_components=None
 
             # 将检验&处理后的n_components写入到pca字典中
             pca_params_dict['n_components']=n_components
@@ -573,21 +588,23 @@ class AudioExtractor:
         return features
 
     def extract_update(self, partition="", meta_paths="", verbose=1):
-        """特征提取和self属性更新维护
+        """
+        根据meta_paths进行特征提取任务
+        提取完特征后对相关self属性更新维护
+
         多次调用将执行增量提取,根据partition的取值,将每次的提取结果增量更新self的相应属性集
+
         Parameters
         ----------
         partition : str
             "train"|"test"
-        meta_files : list[str]|str
-            _description_
+        meta_paths : list[str]|str
+            需要提取特征的meta文件路径
         """
         if not meta_paths:
             raise ValueError("meta_files cannot be empty")
             # return meta_files
-        if isinstance(meta_paths, str):
-            # print(f"cast the '{meta_paths}' to [str]")
-            meta_paths = [meta_paths]
+        meta_paths = self.pathlike_to_list(meta_paths)
 
         # 执行特征提取
         for meta_file in meta_paths:
@@ -614,6 +631,7 @@ class AudioExtractor:
                 features=features,
             )
 
+
     def load_data_preprocessing(self, meta_files=None, partition="", shuffle=False):
         """将特征提取和属性设置以及打乱和平衡操作打包处理
         AE对象在如数据集后可选的数据处理操作(balance&shuffle)
@@ -883,21 +901,22 @@ def load_data_from_meta(
     balance=False,
     feature_transforms=None,
 ) -> dict:
-    """导入语音数据,并返回numpy打包train/test dataset相关属性的ndarray类型
-    如果只想提取train/test dataset中的一方,那么另一方就传None(或者不传对应参数)
+    """
+    根据meta文件,提取/导入语音数据(numpy特征),并返回numpy打包train/test dataset相关属性的ndarray类型
+    如果只想提取train/test dataset中的一方,那么另一方就传None(或者不传对应参数)即,前两个参数中允许其中一个为None
 
     Parameters
     ----------
     train_desc_files : list
-        train_meta_files
+        需要提取特征的语音文件列表信息,作为训练集
     test_desc_files : list
-        test_meta_files
-    f_config : dict, optional
-        需要提取的特征, by default None
-    e_config : list, optional
-        需要使用的情感类别字符串构成的列表, by default ['sad', 'neutral', 'happy']
+        需要提取特征的语音文件列表信息,作为测试集
+    f_config : list[str], optional
+        需要提取的特征组合, by default None
+    e_config : list[str], optional
+        需要使用的情感组合,类别字符串构成的列表, by default ['sad', 'neutral', 'happy']
     classification_task : bool, optional
-        是否采用分类器(否则使用回归模型), by default True
+        是否采用分类模型(否则使用回归模型), by default True
     shuffle : bool, optional
         是否打乱顺序, by default True
     balance : bool, optional
@@ -948,15 +967,27 @@ def load_data_from_meta(
     }
 
 
-if __name__ == "__main__":
-    ftd = dict(std_scaler=False, pca_params=dict(n_components=39))
-    ae = AudioExtractor(
-        e_config=e_config_def,
-        f_config=f_config_def, shuffle=True, feature_transforms_dict=ftd
+
+def load_data_from_meta_demo():
+    meta_dict=dict(
+        train_meta_files=meta_dir/'train_emodb_HNS.csv',
+        test_meta_files=meta_dir/'test_emodb_HNS.csv',
     )
-    print(ae)
-    ae._extract_feature_in_meta(meta_path=train_emodb_csv)
+    res=load_data_from_meta(**meta_dict,f_config=f_config_def)
 
+    return res
+
+if __name__ == "__main__":
+
+    load_data_from_meta_demo()
+
+    # ftd = dict(std_scaler=False, pca_params=dict(n_components=3))
+    # ae = AudioExtractor(
+    #     e_config=e_config_def,
+    #     f_config=f_config_def, shuffle=True, feature_transforms_dict=ftd
+    # )
+    # print(ae)
+    # ae._extract_feature_in_meta(meta_path=train_emodb_csv)
     # data = load_data_from_meta(
     #     # train_meta_files=train_emodb_csv,
     #     test_meta_files=test_emodb_csv,
@@ -964,3 +995,4 @@ if __name__ == "__main__":
     #     # balance=False,
     #     balance=True,
     # )
+    

audio/graph.py

@@ -59,24 +59,32 @@ def showMelFreqGraph(audio_path=audio_file):
     y, sr = librosa.load(audio_path)
 
     # 计算Mel频谱
-    S = librosa.feature.melspectrogram(y, sr=sr, n_mels=128)
+    S = librosa.feature.melspectrogram(y=y, sr=sr, n_mels=128)
+    #另一种等效写法:(使用S参数)
+    # D=np.abs(librosa.stft(y))**2
+    # S=librosa.feature.melspectrogram(S=D, sr=sr, n_mels=128) 
 
     # 将Mel频谱转换为分贝表示
     S_db = librosa.power_to_db(S, ref=np.max)
 
-    # 绘制Mel频谱图
+    # 绘制Mel频谱(系数)图
+    fig,ax=plt.subplots()
     # plt.figure(figsize=(10, 4))
-    librosa.display.specshow(S_db, x_axis='time', y_axis='mel', sr=sr, fmax=8000)
-    plt.colorbar(format='%+2.0f dB')
+    img=librosa.display.specshow(S_db, x_axis='time', y_axis='mel', sr=sr, fmax=8000,ax=ax)
+
+    fig.colorbar(img,ax=ax,format='%+2.0f dB')
+    ax.set(title='Mel-frequencey spectrogram')
 
     # plt.title('Mel频谱图')
     # plt.xlabel('时间')
     # plt.ylabel('频率（Mel刻度）')
-    # plt.show()
-
+    plt.show()
+def showMFCCGraph(audio_path=audio_file):
+    pass
 
 if __name__=="__main__":
+    pass
     # showWaveForm()
-    showFreqGraph()
-    # showMelFreqForm()
-    
+    # showFreqGraph()
+    showMelFreqGraph()
+##
\ No newline at end of file

charts/mindmap.md

+- 启动系统
+- 选择语料库组合
+- 配置情感组合
+- 配置情感特征
+- 配置配置预处理参数
+- 
\ No newline at end of file

charts/tmp.txt

+本文是本人本科阶段第一次完成一篇完整的论文写作.在本文完成的过程中，除了个人努力以外还得到了许多人的支持和帮助。
+首先，我要感谢指导老师蒋海华老师,他在整个毕设的过程中给予了我宝贵的指导和建议，使我能够更好地理解和掌握语音情感识别的相关知识和技术。蒋老师的专业知识、严谨态度和无私奉献精神，对我的毕设研究起到了至关重要的作用。
+其次，我要感谢陪伴我共同学习的同学们和教授过我专业课的老师们，在写作本文期间,他们直接或间接的向我提供了帮助。感谢学校为我提供了良好的学习和研究的环境与氛围，使我能够完成此次毕业设计和论文写作。
+在此，我还要感谢为本研究提供了宝贵语音数据的人员以及开发工具框架的工程师，他们为本研究做出了重要贡献。最后，我要感谢我的家人和朋友们，在我完成毕设的过程中给予了我无尽的关心和支持，他们的关爱和鼓励，是我不断前行和超越自我的动力源泉。

meta_files/test_emodb_HNS.csv

 ,path,emotion
-0,D:\repos\CCSER\SER\data\emodb\wav\12b02Fb.wav,happy
-19,D:\repos\CCSER\SER\data\emodb\wav\08a05Fe.wav,happy
-27,D:\repos\CCSER\SER\data\emodb\wav\16a07Fb.wav,happy
-29,D:\repos\CCSER\SER\data\emodb\wav\11a05Fb.wav,happy
-12,D:\repos\CCSER\SER\data\emodb\wav\10b10Fc.wav,happy
-20,D:\repos\CCSER\SER\data\emodb\wav\10b01Fa.wav,happy
-8,D:\repos\CCSER\SER\data\emodb\wav\14b02Fb.wav,happy
-41,D:\repos\CCSER\SER\data\emodb\wav\11b09Fd.wav,happy
-36,D:\repos\CCSER\SER\data\emodb\wav\14b01Fc.wav,happy
-39,D:\repos\CCSER\SER\data\emodb\wav\09b03Fd.wav,happy
-40,D:\repos\CCSER\SER\data\emodb\wav\16b03Fd.wav,happy
-1,D:\repos\CCSER\SER\data\emodb\wav\08b09Fd.wav,happy
-7,D:\repos\CCSER\SER\data\emodb\wav\09a01Fa.wav,happy
-35,D:\repos\CCSER\SER\data\emodb\wav\14b02Na.wav,neutral
-37,D:\repos\CCSER\SER\data\emodb\wav\14a02Nc.wav,neutral
-32,D:\repos\CCSER\SER\data\emodb\wav\03b10Nc.wav,neutral
-38,D:\repos\CCSER\SER\data\emodb\wav\15b10Nb.wav,neutral
-31,D:\repos\CCSER\SER\data\emodb\wav\15b10Nc.wav,neutral
-30,D:\repos\CCSER\SER\data\emodb\wav\11b10Nc.wav,neutral
-26,D:\repos\CCSER\SER\data\emodb\wav\14a05Na.wav,neutral
-25,D:\repos\CCSER\SER\data\emodb\wav\10a01Nb.wav,neutral
-33,D:\repos\CCSER\SER\data\emodb\wav\13a07Na.wav,neutral
-21,D:\repos\CCSER\SER\data\emodb\wav\03a02Nc.wav,neutral
-42,D:\repos\CCSER\SER\data\emodb\wav\09a04Nb.wav,neutral
-2,D:\repos\CCSER\SER\data\emodb\wav\13b09Na.wav,neutral
-18,D:\repos\CCSER\SER\data\emodb\wav\13b01Nc.wav,neutral
-17,D:\repos\CCSER\SER\data\emodb\wav\03b03Nb.wav,neutral
-16,D:\repos\CCSER\SER\data\emodb\wav\09b10Nd.wav,neutral
-3,D:\repos\CCSER\SER\data\emodb\wav\13b02Nb.wav,neutral
-4,D:\repos\CCSER\SER\data\emodb\wav\10a04Nb.wav,neutral
-22,D:\repos\CCSER\SER\data\emodb\wav\09a07Na.wav,neutral
-10,D:\repos\CCSER\SER\data\emodb\wav\14a07Na.wav,neutral
-9,D:\repos\CCSER\SER\data\emodb\wav\15a05Na.wav,neutral
+27,D:\repos\CCSER\SER\data\emodb\wav\14b02Fb.wav,happy
+2,D:\repos\CCSER\SER\data\emodb\wav\08b01Fd.wav,happy
+26,D:\repos\CCSER\SER\data\emodb\wav\11b09Fd.wav,happy
+25,D:\repos\CCSER\SER\data\emodb\wav\08b09Fd.wav,happy
+38,D:\repos\CCSER\SER\data\emodb\wav\10b10Fc.wav,happy
+23,D:\repos\CCSER\SER\data\emodb\wav\09b03Fa.wav,happy
+8,D:\repos\CCSER\SER\data\emodb\wav\15a05Fb.wav,happy
+36,D:\repos\CCSER\SER\data\emodb\wav\13a04Fc.wav,happy
+30,D:\repos\CCSER\SER\data\emodb\wav\11a04Fd.wav,happy
+42,D:\repos\CCSER\SER\data\emodb\wav\08b01Na.wav,neutral
+24,D:\repos\CCSER\SER\data\emodb\wav\03b01Nb.wav,neutral
+41,D:\repos\CCSER\SER\data\emodb\wav\09b02Na.wav,neutral
+32,D:\repos\CCSER\SER\data\emodb\wav\09b09Nd.wav,neutral
+19,D:\repos\CCSER\SER\data\emodb\wav\10b02Na.wav,neutral
+33,D:\repos\CCSER\SER\data\emodb\wav\16a07Nb.wav,neutral
+34,D:\repos\CCSER\SER\data\emodb\wav\03b09Nc.wav,neutral
+14,D:\repos\CCSER\SER\data\emodb\wav\08b10Nc.wav,neutral
+15,D:\repos\CCSER\SER\data\emodb\wav\13b02Nb.wav,neutral
+35,D:\repos\CCSER\SER\data\emodb\wav\09b01Na.wav,neutral
+12,D:\repos\CCSER\SER\data\emodb\wav\11a04Nd.wav,neutral
+10,D:\repos\CCSER\SER\data\emodb\wav\13b01Nc.wav,neutral
+37,D:\repos\CCSER\SER\data\emodb\wav\13a05Nb.wav,neutral
+7,D:\repos\CCSER\SER\data\emodb\wav\09a07Na.wav,neutral
+6,D:\repos\CCSER\SER\data\emodb\wav\09a04Nb.wav,neutral
+39,D:\repos\CCSER\SER\data\emodb\wav\15b09Nb.wav,neutral
+40,D:\repos\CCSER\SER\data\emodb\wav\15b10Nb.wav,neutral
+1,D:\repos\CCSER\SER\data\emodb\wav\16b03Nb.wav,neutral
+16,D:\repos\CCSER\SER\data\emodb\wav\08a02Na.wav,neutral
+29,D:\repos\CCSER\SER\data\emodb\wav\13a01Nb.wav,neutral
+31,D:\repos\CCSER\SER\data\emodb\wav\11b10Td.wav,sad
+0,D:\repos\CCSER\SER\data\emodb\wav\12b01Ta.wav,sad
+22,D:\repos\CCSER\SER\data\emodb\wav\13a02Ta.wav,sad
+20,D:\repos\CCSER\SER\data\emodb\wav\10b03Tb.wav,sad
+18,D:\repos\CCSER\SER\data\emodb\wav\13a05Tc.wav,sad
+17,D:\repos\CCSER\SER\data\emodb\wav\09a05Tb.wav,sad
+13,D:\repos\CCSER\SER\data\emodb\wav\16a02Tc.wav,sad
+11,D:\repos\CCSER\SER\data\emodb\wav\08b02Tc.wav,sad
+9,D:\repos\CCSER\SER\data\emodb\wav\12b03Ta.wav,sad
 5,D:\repos\CCSER\SER\data\emodb\wav\10a05Tb.wav,sad
-6,D:\repos\CCSER\SER\data\emodb\wav\16a07Td.wav,sad
-23,D:\repos\CCSER\SER\data\emodb\wav\14b03Ta.wav,sad
-13,D:\repos\CCSER\SER\data\emodb\wav\11a07Ta.wav,sad
-14,D:\repos\CCSER\SER\data\emodb\wav\15b09Ta.wav,sad
-15,D:\repos\CCSER\SER\data\emodb\wav\16b01Tb.wav,sad
+4,D:\repos\CCSER\SER\data\emodb\wav\08a07Ta.wav,sad
+3,D:\repos\CCSER\SER\data\emodb\wav\14a07Tc.wav,sad
 28,D:\repos\CCSER\SER\data\emodb\wav\08b03Tc.wav,sad
-24,D:\repos\CCSER\SER\data\emodb\wav\14a04Tb.wav,sad
-34,D:\repos\CCSER\SER\data\emodb\wav\14b02Tc.wav,sad
-11,D:\repos\CCSER\SER\data\emodb\wav\16b10Td.wav,sad
+21,D:\repos\CCSER\SER\data\emodb\wav\08a07Tb.wav,sad

meta_files/train_emodb_HNS.csv

 ,path,emotion
-0,D:\repos\CCSER\SER\data\emodb\wav\03a01Fa.wav,happy
-98,D:\repos\CCSER\SER\data\emodb\wav\14a05Fa.wav,happy
-97,D:\repos\CCSER\SER\data\emodb\wav\11a02Fb.wav,happy
-96,D:\repos\CCSER\SER\data\emodb\wav\03b01Fa.wav,happy
-94,D:\repos\CCSER\SER\data\emodb\wav\14a07Fd.wav,happy
-93,D:\repos\CCSER\SER\data\emodb\wav\09a04Fd.wav,happy
-91,D:\repos\CCSER\SER\data\emodb\wav\11b02Fd.wav,happy
-89,D:\repos\CCSER\SER\data\emodb\wav\15a01Fb.wav,happy
-88,D:\repos\CCSER\SER\data\emodb\wav\11a05Fc.wav,happy
-99,D:\repos\CCSER\SER\data\emodb\wav\13b09Fb.wav,happy
-86,D:\repos\CCSER\SER\data\emodb\wav\08a02Fe.wav,happy
-77,D:\repos\CCSER\SER\data\emodb\wav\09b03Fa.wav,happy
-72,D:\repos\CCSER\SER\data\emodb\wav\16b02Fd.wav,happy
-71,D:\repos\CCSER\SER\data\emodb\wav\08b03Fe.wav,happy
-69,D:\repos\CCSER\SER\data\emodb\wav\15a07Fb.wav,happy
-68,D:\repos\CCSER\SER\data\emodb\wav\15a05Fb.wav,happy
-64,D:\repos\CCSER\SER\data\emodb\wav\12a01Fb.wav,happy
-61,D:\repos\CCSER\SER\data\emodb\wav\03a07Fb.wav,happy
-56,D:\repos\CCSER\SER\data\emodb\wav\16a07Fa.wav,happy
-167,D:\repos\CCSER\SER\data\emodb\wav\13a01Fd.wav,happy
-53,D:\repos\CCSER\SER\data\emodb\wav\13b01Fc.wav,happy
-100,D:\repos\CCSER\SER\data\emodb\wav\08b10Fd.wav,happy
-104,D:\repos\CCSER\SER\data\emodb\wav\13b02Fb.wav,happy
-157,D:\repos\CCSER\SER\data\emodb\wav\14a02Fd.wav,happy
-148,D:\repos\CCSER\SER\data\emodb\wav\11a04Fd.wav,happy
-143,D:\repos\CCSER\SER\data\emodb\wav\14b09Fc.wav,happy
-142,D:\repos\CCSER\SER\data\emodb\wav\14a05Fb.wav,happy
-139,D:\repos\CCSER\SER\data\emodb\wav\16b09Fb.wav,happy
-135,D:\repos\CCSER\SER\data\emodb\wav\13a04Fc.wav,happy
-133,D:\repos\CCSER\SER\data\emodb\wav\03a04Fd.wav,happy
-131,D:\repos\CCSER\SER\data\emodb\wav\16b10Fb.wav,happy
-103,D:\repos\CCSER\SER\data\emodb\wav\08b02Ff.wav,happy
-129,D:\repos\CCSER\SER\data\emodb\wav\08b01Fd.wav,happy
-122,D:\repos\CCSER\SER\data\emodb\wav\15a04Fd.wav,happy
-120,D:\repos\CCSER\SER\data\emodb\wav\16a04Fa.wav,happy
-119,D:\repos\CCSER\SER\data\emodb\wav\16a05Fc.wav,happy
-117,D:\repos\CCSER\SER\data\emodb\wav\13b10Fa.wav,happy
-114,D:\repos\CCSER\SER\data\emodb\wav\16b03Fa.wav,happy
-112,D:\repos\CCSER\SER\data\emodb\wav\08b01Fe.wav,happy
-110,D:\repos\CCSER\SER\data\emodb\wav\03a02Fc.wav,happy
-106,D:\repos\CCSER\SER\data\emodb\wav\08a04Ff.wav,happy
-127,D:\repos\CCSER\SER\data\emodb\wav\13a02Fa.wav,happy
-49,D:\repos\CCSER\SER\data\emodb\wav\16b01Fa.wav,happy
-84,D:\repos\CCSER\SER\data\emodb\wav\13b09Fc.wav,happy
-168,D:\repos\CCSER\SER\data\emodb\wav\03a05Fc.wav,happy
-7,D:\repos\CCSER\SER\data\emodb\wav\03a07Fa.wav,happy
-22,D:\repos\CCSER\SER\data\emodb\wav\08a07Fd.wav,happy
-21,D:\repos\CCSER\SER\data\emodb\wav\08a01Fd.wav,happy
-11,D:\repos\CCSER\SER\data\emodb\wav\13a07Fd.wav,happy
-1,D:\repos\CCSER\SER\data\emodb\wav\11b01Fc.wav,happy
-13,D:\repos\CCSER\SER\data\emodb\wav\16a01Fc.wav,happy
-38,D:\repos\CCSER\SER\data\emodb\wav\13b03Fd.wav,happy
-14,D:\repos\CCSER\SER\data\emodb\wav\15a07Fa.wav,happy
-43,D:\repos\CCSER\SER\data\emodb\wav\10a02Fa.wav,happy
-5,D:\repos\CCSER\SER\data\emodb\wav\11b03Fc.wav,happy
-2,D:\repos\CCSER\SER\data\emodb\wav\14b01Fa.wav,happy
-8,D:\repos\CCSER\SER\data\emodb\wav\10a04Fd.wav,happy
-15,D:\repos\CCSER\SER\data\emodb\wav\15b09Fa.wav,happy
-126,D:\repos\CCSER\SER\data\emodb\wav\15a07Nc.wav,neutral
-116,D:\repos\CCSER\SER\data\emodb\wav\10b02Na.wav,neutral
-111,D:\repos\CCSER\SER\data\emodb\wav\09b01Na.wav,neutral
-113,D:\repos\CCSER\SER\data\emodb\wav\11a04Nd.wav,neutral
-108,D:\repos\CCSER\SER\data\emodb\wav\14b01Na.wav,neutral
-105,D:\repos\CCSER\SER\data\emodb\wav\15a04Nc.wav,neutral
-19,D:\repos\CCSER\SER\data\emodb\wav\13a01Nb.wav,neutral
-20,D:\repos\CCSER\SER\data\emodb\wav\11a02Nc.wav,neutral
-101,D:\repos\CCSER\SER\data\emodb\wav\13a05Nb.wav,neutral
-23,D:\repos\CCSER\SER\data\emodb\wav\08b02Nb.wav,neutral
-24,D:\repos\CCSER\SER\data\emodb\wav\15b01Na.wav,neutral
-115,D:\repos\CCSER\SER\data\emodb\wav\13a02Nc.wav,neutral
-128,D:\repos\CCSER\SER\data\emodb\wav\08a02Na.wav,neutral
-132,D:\repos\CCSER\SER\data\emodb\wav\09a05Nb.wav,neutral
-25,D:\repos\CCSER\SER\data\emodb\wav\08a05Nb.wav,neutral
-166,D:\repos\CCSER\SER\data\emodb\wav\03a07Nc.wav,neutral
-163,D:\repos\CCSER\SER\data\emodb\wav\16b03Nb.wav,neutral
-161,D:\repos\CCSER\SER\data\emodb\wav\15b09Nb.wav,neutral
-160,D:\repos\CCSER\SER\data\emodb\wav\13b10Nc.wav,neutral
-159,D:\repos\CCSER\SER\data\emodb\wav\08b01Na.wav,neutral
-156,D:\repos\CCSER\SER\data\emodb\wav\14b10Nb.wav,neutral
-153,D:\repos\CCSER\SER\data\emodb\wav\03a04Nc.wav,neutral
-152,D:\repos\CCSER\SER\data\emodb\wav\15a01Nb.wav,neutral
-9,D:\repos\CCSER\SER\data\emodb\wav\11b03Nb.wav,neutral
-151,D:\repos\CCSER\SER\data\emodb\wav\03a05Nd.wav,neutral
-147,D:\repos\CCSER\SER\data\emodb\wav\08a01Na.wav,neutral
-144,D:\repos\CCSER\SER\data\emodb\wav\11b02Na.wav,neutral
-3,D:\repos\CCSER\SER\data\emodb\wav\09b02Na.wav,neutral
-141,D:\repos\CCSER\SER\data\emodb\wav\09b09Nd.wav,neutral
-140,D:\repos\CCSER\SER\data\emodb\wav\03b10Na.wav,neutral
-137,D:\repos\CCSER\SER\data\emodb\wav\15a02Na.wav,neutral
-136,D:\repos\CCSER\SER\data\emodb\wav\03b01Nb.wav,neutral
-6,D:\repos\CCSER\SER\data\emodb\wav\16a01Nc.wav,neutral
-149,D:\repos\CCSER\SER\data\emodb\wav\12b02Na.wav,neutral
-26,D:\repos\CCSER\SER\data\emodb\wav\15b02Nd.wav,neutral
-18,D:\repos\CCSER\SER\data\emodb\wav\08a07Na.wav,neutral
-92,D:\repos\CCSER\SER\data\emodb\wav\08b09Nb.wav,neutral
-33,D:\repos\CCSER\SER\data\emodb\wav\09a01Nb.wav,neutral
-75,D:\repos\CCSER\SER\data\emodb\wav\11b01Nc.wav,neutral
-73,D:\repos\CCSER\SER\data\emodb\wav\08a04Nc.wav,neutral
-34,D:\repos\CCSER\SER\data\emodb\wav\12a01Nb.wav,neutral
-35,D:\repos\CCSER\SER\data\emodb\wav\16a02Nb.wav,neutral
-70,D:\repos\CCSER\SER\data\emodb\wav\14a01Na.wav,neutral
-37,D:\repos\CCSER\SER\data\emodb\wav\11a05Na.wav,neutral
-59,D:\repos\CCSER\SER\data\emodb\wav\11b09Na.wav,neutral
-57,D:\repos\CCSER\SER\data\emodb\wav\10a02Na.wav,neutral
-54,D:\repos\CCSER\SER\data\emodb\wav\16a04Nc.wav,neutral
-50,D:\repos\CCSER\SER\data\emodb\wav\12a02Nb.wav,neutral
-42,D:\repos\CCSER\SER\data\emodb\wav\16a07Nb.wav,neutral
-48,D:\repos\CCSER\SER\data\emodb\wav\03b09Nc.wav,neutral
-46,D:\repos\CCSER\SER\data\emodb\wav\09b03Nb.wav,neutral
-45,D:\repos\CCSER\SER\data\emodb\wav\03a01Nc.wav,neutral
-78,D:\repos\CCSER\SER\data\emodb\wav\08b03Nb.wav,neutral
-79,D:\repos\CCSER\SER\data\emodb\wav\08b10Nc.wav,neutral
-36,D:\repos\CCSER\SER\data\emodb\wav\13b03Na.wav,neutral
-80,D:\repos\CCSER\SER\data\emodb\wav\15b03Nb.wav,neutral
-85,D:\repos\CCSER\SER\data\emodb\wav\11a01Nd.wav,neutral
-90,D:\repos\CCSER\SER\data\emodb\wav\03b02Na.wav,neutral
-82,D:\repos\CCSER\SER\data\emodb\wav\12a05Nd.wav,neutral
-28,D:\repos\CCSER\SER\data\emodb\wav\08b10Tc.wav,sad
-58,D:\repos\CCSER\SER\data\emodb\wav\13a07Tc.wav,sad
-145,D:\repos\CCSER\SER\data\emodb\wav\08a04Tb.wav,sad
-146,D:\repos\CCSER\SER\data\emodb\wav\13a02Ta.wav,sad
-31,D:\repos\CCSER\SER\data\emodb\wav\14a04Tc.wav,sad
-40,D:\repos\CCSER\SER\data\emodb\wav\11b02Td.wav,sad
-55,D:\repos\CCSER\SER\data\emodb\wav\11b03Td.wav,sad
-95,D:\repos\CCSER\SER\data\emodb\wav\16b10Tb.wav,sad
-150,D:\repos\CCSER\SER\data\emodb\wav\08a07Tb.wav,sad
-87,D:\repos\CCSER\SER\data\emodb\wav\08a07Ta.wav,sad
-41,D:\repos\CCSER\SER\data\emodb\wav\16a04Tc.wav,sad
-52,D:\repos\CCSER\SER\data\emodb\wav\11a05Td.wav,sad
-154,D:\repos\CCSER\SER\data\emodb\wav\09a05Tb.wav,sad
-102,D:\repos\CCSER\SER\data\emodb\wav\14b10Tc.wav,sad
-155,D:\repos\CCSER\SER\data\emodb\wav\03a05Tc.wav,sad
-51,D:\repos\CCSER\SER\data\emodb\wav\16a02Tc.wav,sad
-30,D:\repos\CCSER\SER\data\emodb\wav\08b02Tc.wav,sad
-158,D:\repos\CCSER\SER\data\emodb\wav\15b03Tc.wav,sad
-29,D:\repos\CCSER\SER\data\emodb\wav\08a02Tb.wav,sad
-165,D:\repos\CCSER\SER\data\emodb\wav\14a07Tc.wav,sad
-47,D:\repos\CCSER\SER\data\emodb\wav\13a05Tc.wav,sad
-162,D:\repos\CCSER\SER\data\emodb\wav\13b03Td.wav,sad
-164,D:\repos\CCSER\SER\data\emodb\wav\16a05Tb.wav,sad
-4,D:\repos\CCSER\SER\data\emodb\wav\03a02Ta.wav,sad
-39,D:\repos\CCSER\SER\data\emodb\wav\16b03Ta.wav,sad
-16,D:\repos\CCSER\SER\data\emodb\wav\10b03Tb.wav,sad
-17,D:\repos\CCSER\SER\data\emodb\wav\08b09Tb.wav,sad
-109,D:\repos\CCSER\SER\data\emodb\wav\14a05Ta.wav,sad
-76,D:\repos\CCSER\SER\data\emodb\wav\09a07Ta.wav,sad
-81,D:\repos\CCSER\SER\data\emodb\wav\13a04Ta.wav,sad
-118,D:\repos\CCSER\SER\data\emodb\wav\03b01Td.wav,sad
-74,D:\repos\CCSER\SER\data\emodb\wav\14a02Tb.wav,sad
-12,D:\repos\CCSER\SER\data\emodb\wav\11b10Td.wav,sad
-121,D:\repos\CCSER\SER\data\emodb\wav\12b09Td.wav,sad
-107,D:\repos\CCSER\SER\data\emodb\wav\16a01Tb.wav,sad
-123,D:\repos\CCSER\SER\data\emodb\wav\11b09Td.wav,sad
-124,D:\repos\CCSER\SER\data\emodb\wav\12a05Ta.wav,sad
-60,D:\repos\CCSER\SER\data\emodb\wav\03a04Ta.wav,sad
-125,D:\repos\CCSER\SER\data\emodb\wav\15b02Tc.wav,sad
-83,D:\repos\CCSER\SER\data\emodb\wav\03b09Tc.wav,sad
-32,D:\repos\CCSER\SER\data\emodb\wav\15a02Ta.wav,sad
-130,D:\repos\CCSER\SER\data\emodb\wav\09b02Tb.wav,sad
-27,D:\repos\CCSER\SER\data\emodb\wav\09b03Ta.wav,sad
-67,D:\repos\CCSER\SER\data\emodb\wav\03b03Tc.wav,sad
-134,D:\repos\CCSER\SER\data\emodb\wav\14a05Tc.wav,sad
-66,D:\repos\CCSER\SER\data\emodb\wav\11a02Tc.wav,sad
-65,D:\repos\CCSER\SER\data\emodb\wav\08a05Ta.wav,sad
-63,D:\repos\CCSER\SER\data\emodb\wav\03b02Tb.wav,sad
-138,D:\repos\CCSER\SER\data\emodb\wav\12b03Ta.wav,sad
-62,D:\repos\CCSER\SER\data\emodb\wav\10a07Ta.wav,sad
-10,D:\repos\CCSER\SER\data\emodb\wav\14b09Td.wav,sad
-44,D:\repos\CCSER\SER\data\emodb\wav\12b01Ta.wav,sad
+92,D:\repos\CCSER\SER\data\emodb\wav\03a01Fa.wav,happy
+136,D:\repos\CCSER\SER\data\emodb\wav\14b09Fc.wav,happy
+35,D:\repos\CCSER\SER\data\emodb\wav\13a07Fd.wav,happy
+74,D:\repos\CCSER\SER\data\emodb\wav\09b03Fd.wav,happy
+37,D:\repos\CCSER\SER\data\emodb\wav\14a05Fa.wav,happy
+38,D:\repos\CCSER\SER\data\emodb\wav\08b10Fd.wav,happy
+125,D:\repos\CCSER\SER\data\emodb\wav\14a02Fd.wav,happy
+40,D:\repos\CCSER\SER\data\emodb\wav\15a07Fb.wav,happy
+41,D:\repos\CCSER\SER\data\emodb\wav\13a02Fa.wav,happy
+167,D:\repos\CCSER\SER\data\emodb\wav\08b01Fe.wav,happy
+123,D:\repos\CCSER\SER\data\emodb\wav\10a04Fd.wav,happy
+95,D:\repos\CCSER\SER\data\emodb\wav\11a02Fb.wav,happy
+121,D:\repos\CCSER\SER\data\emodb\wav\13b02Fb.wav,happy
+46,D:\repos\CCSER\SER\data\emodb\wav\16a07Fa.wav,happy
+119,D:\repos\CCSER\SER\data\emodb\wav\11a05Fc.wav,happy
+48,D:\repos\CCSER\SER\data\emodb\wav\11b01Fc.wav,happy
+117,D:\repos\CCSER\SER\data\emodb\wav\09a01Fa.wav,happy
+116,D:\repos\CCSER\SER\data\emodb\wav\08b03Fe.wav,happy
+51,D:\repos\CCSER\SER\data\emodb\wav\08a04Ff.wav,happy
+52,D:\repos\CCSER\SER\data\emodb\wav\15a04Fd.wav,happy
+53,D:\repos\CCSER\SER\data\emodb\wav\16a07Fb.wav,happy
+55,D:\repos\CCSER\SER\data\emodb\wav\16a04Fa.wav,happy
+111,D:\repos\CCSER\SER\data\emodb\wav\08a01Fd.wav,happy
+97,D:\repos\CCSER\SER\data\emodb\wav\08a07Fd.wav,happy
+108,D:\repos\CCSER\SER\data\emodb\wav\16b02Fd.wav,happy
+59,D:\repos\CCSER\SER\data\emodb\wav\14b01Fc.wav,happy
+70,D:\repos\CCSER\SER\data\emodb\wav\13b01Fc.wav,happy
+98,D:\repos\CCSER\SER\data\emodb\wav\03a04Fd.wav,happy
+100,D:\repos\CCSER\SER\data\emodb\wav\03a07Fb.wav,happy
+101,D:\repos\CCSER\SER\data\emodb\wav\13b09Fb.wav,happy
+137,D:\repos\CCSER\SER\data\emodb\wav\09a04Fd.wav,happy
+64,D:\repos\CCSER\SER\data\emodb\wav\13b09Fc.wav,happy
+103,D:\repos\CCSER\SER\data\emodb\wav\16b01Fa.wav,happy
+156,D:\repos\CCSER\SER\data\emodb\wav\10b01Fa.wav,happy
+2,D:\repos\CCSER\SER\data\emodb\wav\14a05Fb.wav,happy
+81,D:\repos\CCSER\SER\data\emodb\wav\16b03Fd.wav,happy
+7,D:\repos\CCSER\SER\data\emodb\wav\15a01Fb.wav,happy
+8,D:\repos\CCSER\SER\data\emodb\wav\08a05Fe.wav,happy
+9,D:\repos\CCSER\SER\data\emodb\wav\12b02Fb.wav,happy
+1,D:\repos\CCSER\SER\data\emodb\wav\03b01Fa.wav,happy
+163,D:\repos\CCSER\SER\data\emodb\wav\11a05Fb.wav,happy
+162,D:\repos\CCSER\SER\data\emodb\wav\15a07Fa.wav,happy
+80,D:\repos\CCSER\SER\data\emodb\wav\12a01Fb.wav,happy
+13,D:\repos\CCSER\SER\data\emodb\wav\16b09Fb.wav,happy
+76,D:\repos\CCSER\SER\data\emodb\wav\08b02Ff.wav,happy
+155,D:\repos\CCSER\SER\data\emodb\wav\15b09Fa.wav,happy
+4,D:\repos\CCSER\SER\data\emodb\wav\14b01Fa.wav,happy
+153,D:\repos\CCSER\SER\data\emodb\wav\03a05Fc.wav,happy
+150,D:\repos\CCSER\SER\data\emodb\wav\11b03Fc.wav,happy
+19,D:\repos\CCSER\SER\data\emodb\wav\08a02Fe.wav,happy
+20,D:\repos\CCSER\SER\data\emodb\wav\14a07Fd.wav,happy
+149,D:\repos\CCSER\SER\data\emodb\wav\16b10Fb.wav,happy
+148,D:\repos\CCSER\SER\data\emodb\wav\13b03Fd.wav,happy
+145,D:\repos\CCSER\SER\data\emodb\wav\03a02Fc.wav,happy
+24,D:\repos\CCSER\SER\data\emodb\wav\10a02Fa.wav,happy
+144,D:\repos\CCSER\SER\data\emodb\wav\16b03Fa.wav,happy
+78,D:\repos\CCSER\SER\data\emodb\wav\16a05Fc.wav,happy
+27,D:\repos\CCSER\SER\data\emodb\wav\03a07Fa.wav,happy
+28,D:\repos\CCSER\SER\data\emodb\wav\13b10Fa.wav,happy
+142,D:\repos\CCSER\SER\data\emodb\wav\11b02Fd.wav,happy
+90,D:\repos\CCSER\SER\data\emodb\wav\13a01Fd.wav,happy
+88,D:\repos\CCSER\SER\data\emodb\wav\16a01Fc.wav,happy
+91,D:\repos\CCSER\SER\data\emodb\wav\14b02Na.wav,neutral
+96,D:\repos\CCSER\SER\data\emodb\wav\03b10Na.wav,neutral
+89,D:\repos\CCSER\SER\data\emodb\wav\15b01Na.wav,neutral
+0,D:\repos\CCSER\SER\data\emodb\wav\14b10Nb.wav,neutral
+124,D:\repos\CCSER\SER\data\emodb\wav\08a05Nb.wav,neutral
+104,D:\repos\CCSER\SER\data\emodb\wav\03b03Nb.wav,neutral
+166,D:\repos\CCSER\SER\data\emodb\wav\10a04Nb.wav,neutral
+165,D:\repos\CCSER\SER\data\emodb\wav\16a02Nb.wav,neutral
+164,D:\repos\CCSER\SER\data\emodb\wav\15b10Nc.wav,neutral
+159,D:\repos\CCSER\SER\data\emodb\wav\12a05Nd.wav,neutral
+151,D:\repos\CCSER\SER\data\emodb\wav\15b03Nb.wav,neutral
+143,D:\repos\CCSER\SER\data\emodb\wav\09a05Nb.wav,neutral
+140,D:\repos\CCSER\SER\data\emodb\wav\03a07Nc.wav,neutral
+102,D:\repos\CCSER\SER\data\emodb\wav\15a01Nb.wav,neutral
+139,D:\repos\CCSER\SER\data\emodb\wav\13b10Nc.wav,neutral
+129,D:\repos\CCSER\SER\data\emodb\wav\12a01Nb.wav,neutral
+122,D:\repos\CCSER\SER\data\emodb\wav\10a02Na.wav,neutral
+113,D:\repos\CCSER\SER\data\emodb\wav\08b09Nb.wav,neutral
+109,D:\repos\CCSER\SER\data\emodb\wav\11b03Nb.wav,neutral
+107,D:\repos\CCSER\SER\data\emodb\wav\08a04Nc.wav,neutral
+106,D:\repos\CCSER\SER\data\emodb\wav\09a01Nb.wav,neutral
+105,D:\repos\CCSER\SER\data\emodb\wav\03a04Nc.wav,neutral
+138,D:\repos\CCSER\SER\data\emodb\wav\11a01Nd.wav,neutral
+87,D:\repos\CCSER\SER\data\emodb\wav\12b02Na.wav,neutral
+84,D:\repos\CCSER\SER\data\emodb\wav\09b10Nd.wav,neutral
+168,D:\repos\CCSER\SER\data\emodb\wav\11b02Na.wav,neutral
+57,D:\repos\CCSER\SER\data\emodb\wav\11a02Nc.wav,neutral
+56,D:\repos\CCSER\SER\data\emodb\wav\08b02Nb.wav,neutral
+33,D:\repos\CCSER\SER\data\emodb\wav\03a05Nd.wav,neutral
+16,D:\repos\CCSER\SER\data\emodb\wav\10a01Nb.wav,neutral
+49,D:\repos\CCSER\SER\data\emodb\wav\14a07Na.wav,neutral
+17,D:\repos\CCSER\SER\data\emodb\wav\13b03Na.wav,neutral
+45,D:\repos\CCSER\SER\data\emodb\wav\15a02Na.wav,neutral
+58,D:\repos\CCSER\SER\data\emodb\wav\08a07Na.wav,neutral
+44,D:\repos\CCSER\SER\data\emodb\wav\03b10Nc.wav,neutral
+42,D:\repos\CCSER\SER\data\emodb\wav\13a02Nc.wav,neutral
+86,D:\repos\CCSER\SER\data\emodb\wav\13a07Na.wav,neutral
+36,D:\repos\CCSER\SER\data\emodb\wav\14a05Na.wav,neutral
+23,D:\repos\CCSER\SER\data\emodb\wav\16a01Nc.wav,neutral
+26,D:\repos\CCSER\SER\data\emodb\wav\03a01Nc.wav,neutral
+30,D:\repos\CCSER\SER\data\emodb\wav\15b02Nd.wav,neutral
+34,D:\repos\CCSER\SER\data\emodb\wav\14b01Na.wav,neutral
+43,D:\repos\CCSER\SER\data\emodb\wav\09b03Nb.wav,neutral
+60,D:\repos\CCSER\SER\data\emodb\wav\11b10Nc.wav,neutral
+32,D:\repos\CCSER\SER\data\emodb\wav\11a05Na.wav,neutral
+12,D:\repos\CCSER\SER\data\emodb\wav\14a02Nc.wav,neutral
+3,D:\repos\CCSER\SER\data\emodb\wav\03b02Na.wav,neutral
+82,D:\repos\CCSER\SER\data\emodb\wav\03a02Nc.wav,neutral
+79,D:\repos\CCSER\SER\data\emodb\wav\08a01Na.wav,neutral
+77,D:\repos\CCSER\SER\data\emodb\wav\16a04Nc.wav,neutral
+75,D:\repos\CCSER\SER\data\emodb\wav\14a01Na.wav,neutral
+73,D:\repos\CCSER\SER\data\emodb\wav\15a04Nc.wav,neutral
+71,D:\repos\CCSER\SER\data\emodb\wav\13b09Na.wav,neutral
+69,D:\repos\CCSER\SER\data\emodb\wav\11b01Nc.wav,neutral
+10,D:\repos\CCSER\SER\data\emodb\wav\08b03Nb.wav,neutral
+72,D:\repos\CCSER\SER\data\emodb\wav\11b09Na.wav,neutral
+68,D:\repos\CCSER\SER\data\emodb\wav\12a02Nb.wav,neutral
+11,D:\repos\CCSER\SER\data\emodb\wav\15a05Na.wav,neutral
+67,D:\repos\CCSER\SER\data\emodb\wav\15a07Nc.wav,neutral
+160,D:\repos\CCSER\SER\data\emodb\wav\11a07Ta.wav,sad
+161,D:\repos\CCSER\SER\data\emodb\wav\09a07Ta.wav,sad
+141,D:\repos\CCSER\SER\data\emodb\wav\16b01Tb.wav,sad
+29,D:\repos\CCSER\SER\data\emodb\wav\08a04Tb.wav,sad
+157,D:\repos\CCSER\SER\data\emodb\wav\11b03Td.wav,sad
+25,D:\repos\CCSER\SER\data\emodb\wav\16b03Ta.wav,sad
+152,D:\repos\CCSER\SER\data\emodb\wav\14a04Tc.wav,sad
+146,D:\repos\CCSER\SER\data\emodb\wav\14a05Tc.wav,sad
+147,D:\repos\CCSER\SER\data\emodb\wav\03a05Tc.wav,sad
+22,D:\repos\CCSER\SER\data\emodb\wav\03b01Td.wav,sad
+6,D:\repos\CCSER\SER\data\emodb\wav\14a05Ta.wav,sad
+15,D:\repos\CCSER\SER\data\emodb\wav\15b02Tc.wav,sad
+158,D:\repos\CCSER\SER\data\emodb\wav\16a01Tb.wav,sad
+154,D:\repos\CCSER\SER\data\emodb\wav\09b02Tb.wav,sad
+21,D:\repos\CCSER\SER\data\emodb\wav\15a02Ta.wav,sad
+18,D:\repos\CCSER\SER\data\emodb\wav\11b02Td.wav,sad
+5,D:\repos\CCSER\SER\data\emodb\wav\08a02Tb.wav,sad
+14,D:\repos\CCSER\SER\data\emodb\wav\14b09Td.wav,sad
+85,D:\repos\CCSER\SER\data\emodb\wav\16a04Tc.wav,sad
+135,D:\repos\CCSER\SER\data\emodb\wav\08b09Tb.wav,sad
+83,D:\repos\CCSER\SER\data\emodb\wav\08b10Tc.wav,sad
+93,D:\repos\CCSER\SER\data\emodb\wav\14b03Ta.wav,sad
+94,D:\repos\CCSER\SER\data\emodb\wav\14a04Tb.wav,sad
+99,D:\repos\CCSER\SER\data\emodb\wav\09b03Ta.wav,sad
+66,D:\repos\CCSER\SER\data\emodb\wav\03a04Ta.wav,sad
+65,D:\repos\CCSER\SER\data\emodb\wav\16a07Td.wav,sad
+63,D:\repos\CCSER\SER\data\emodb\wav\08a05Ta.wav,sad
+62,D:\repos\CCSER\SER\data\emodb\wav\15b09Ta.wav,sad
+61,D:\repos\CCSER\SER\data\emodb\wav\03b09Tc.wav,sad
+110,D:\repos\CCSER\SER\data\emodb\wav\12b09Td.wav,sad
+112,D:\repos\CCSER\SER\data\emodb\wav\14b02Tc.wav,sad
+54,D:\repos\CCSER\SER\data\emodb\wav\11a02Tc.wav,sad
+31,D:\repos\CCSER\SER\data\emodb\wav\03b02Tb.wav,sad
+114,D:\repos\CCSER\SER\data\emodb\wav\13a04Ta.wav,sad
+50,D:\repos\CCSER\SER\data\emodb\wav\14b10Tc.wav,sad
+118,D:\repos\CCSER\SER\data\emodb\wav\11a05Td.wav,sad
+47,D:\repos\CCSER\SER\data\emodb\wav\11b09Td.wav,sad
+120,D:\repos\CCSER\SER\data\emodb\wav\03a02Ta.wav,sad
+39,D:\repos\CCSER\SER\data\emodb\wav\10a07Ta.wav,sad
+126,D:\repos\CCSER\SER\data\emodb\wav\16a05Tb.wav,sad
+128,D:\repos\CCSER\SER\data\emodb\wav\14a02Tb.wav,sad
+130,D:\repos\CCSER\SER\data\emodb\wav\12a05Ta.wav,sad
+131,D:\repos\CCSER\SER\data\emodb\wav\16b10Td.wav,sad
+132,D:\repos\CCSER\SER\data\emodb\wav\13a07Tc.wav,sad
+133,D:\repos\CCSER\SER\data\emodb\wav\15b03Tc.wav,sad
+134,D:\repos\CCSER\SER\data\emodb\wav\03b03Tc.wav,sad
+115,D:\repos\CCSER\SER\data\emodb\wav\13b03Td.wav,sad
+127,D:\repos\CCSER\SER\data\emodb\wav\16b10Tb.wav,sad

recognizer/basic.py

@@ -32,7 +32,7 @@ from sklearn.svm import SVC
 from tqdm import tqdm
 
 from audio.extractor import AudioExtractor, load_data_from_meta
-from config.EF import e_config_def, f_config_def, validate_emotions
+from config.EF import AHNPS, e_config_def, f_config_def, validate_emotions
 from config.MetaPath import (
     emodb,
     meta_paths_of_db,
@@ -713,24 +713,24 @@ def main():
     passive_emo_others = passive_emo + ["others"]
     typical_emo = ["happy", "neutral", "sad"]
     AHSO = ["angry", "neutral", "sad", "others"]
-    e_config = passive_emo
+    e_config = AHNPS
     f_config = ["mfcc"]
 
     # my_model = RandomForestClassifier(max_depth=3, max_features=0.2)
     my_model = SVC(C=0.001, gamma=0.001, kernel="poly", probability=True)
     # my_model=KNeighborsClassifier(n_neighbors=3, p=1, weights='distance')
-    # my_model = None
+    my_model = None
 
     # rec = EmotionRecognizer(model=my_model,e_config=AHNPS,f_config=f_config_def,test_dbs=[ravdess],train_dbs=[ravdess], verbose=1)
     # rec = EmotionRecognizer(model=my_model,e_config=AHNPS,f_config=f_config_def,test_dbs=emodb,train_dbs=emodb, verbose=1)
 
-    single_db = emodb
-    meta_dict = {"train_dbs": single_db, "test_dbs": single_db}
+    # single_db = emodb
+    # meta_dict = {"train_dbs": single_db, "test_dbs": single_db}
 
-    # meta_dict=dict(
-    #     train_dbs=emodb,
-    #     test_dbs=ravdess
-    # )
+    meta_dict=dict(
+        train_dbs=emodb,
+        test_dbs=emodb
+    )
 
     er = EmotionRecognizer(
         model=my_model,
@@ -739,8 +739,8 @@ def main():
         f_config=f_config,
         verbose=1,
 
-        std_scaler=False, 
-        pca_params=dict(n_components=39)
+        # std_scaler=False, 
+        # pca_params=dict(n_components=39)
 
         # std_scaler=False,
         # pca={"n_components":"mle"}
@@ -755,7 +755,7 @@ def main():
 
     cv_score = er.model_cv_score()
     print(f"{cv_score=}")
-
+    print(er.confusion_matrix())
     return er
 
 

recognizer/check_gpu.ipynb

+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "True"
+      ]
+     },
+     "execution_count": 1,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "import tensorflow as tf\n",
+    "tf.test.is_built_with_cuda()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Num GPUs Available:  1\n"
+     ]
+    }
+   ],
+   "source": [
+    "import tensorflow as tf\n",
+    "print(\"Num GPUs Available: \", len(tf.config.list_physical_devices('GPU')))"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "tf2.10",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.16"
+  },
+  "orig_nbformat": 4
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

recognizer/cross_validation_demo.ipynb

+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "X.shape, y.shape:  (150, 4) (150,)\n",
+      "X_train.shape, y_train.shape:  (90, 4) (90,)\n",
+      "X_test.shape, y_test.shape:  (60, 4) (60,)\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "0.9666666666666667"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "from sklearn.model_selection import train_test_split\n",
+    "from sklearn import datasets\n",
+    "from sklearn import svm\n",
+    "\n",
+    "X, y = datasets.load_iris(return_X_y=True)\n",
+    "X.shape, y.shape\n",
+    "print('X.shape, y.shape: ', X.shape, y.shape)\n",
+    "X_train, X_test, y_train, y_test = train_test_split(\n",
+    "    X, y, test_size=0.4, random_state=0)\n",
+    "\n",
+    "X_train.shape, y_train.shape\n",
+    "print('X_train.shape, y_train.shape: ', X_train.shape, y_train.shape)\n",
+    "\n",
+    "X_test.shape, y_test.shape\n",
+    "print('X_test.shape, y_test.shape: ', X_test.shape, y_test.shape)\n",
+    "\n",
+    "\n",
+    "clf = svm.SVC(kernel='linear', C=1).fit(X_train, y_train)\n",
+    "clf.score(X_test, y_test)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "tf2.10",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.16"
+  },
+  "orig_nbformat": 4
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

recognizer/deep.py

 import os
+
 # disable keras loggings
 import sys
 
-from config.EF import e_config_def
+from config.EF import e_config_def, f_config_def
 
 stderr = sys.stderr
 sys.stderr = open(os.devnull, "w")
 import random
-
+from config.MetaPath import savee, emodb, ravdess
 import numpy as np
 import pandas as pd
-from sklearn.metrics import (accuracy_score, confusion_matrix,
-                             mean_absolute_error)
+from sklearn.metrics import accuracy_score, confusion_matrix, mean_absolute_error
+
 # import tensorflow as tf
 from tensorflow.keras.callbacks import ModelCheckpoint, TensorBoard
-from tensorflow.keras.layers import (LSTM, Dense, Dropout)
+from tensorflow.keras.layers import LSTM, Dense, Dropout
 from tensorflow.keras.models import Sequential
 from tensorflow.keras.utils import to_categorical
 
+from audio.core import extract_feature_of_audio, get_dropout_str
 from config.EF import validate_emotions
-from recognizer.basic import EmotionRecognizer
+
 # from ER import EmotionRecognizer
 from config.MetaPath import get_first_letters
-from audio.core import extract_feature_of_audio, get_dropout_str
+from recognizer.basic import EmotionRecognizer
 
 
 class DeepEmotionRecognizer(EmotionRecognizer):
@@ -128,7 +130,6 @@ class DeepEmotionRecognizer(EmotionRecognizer):
         )
         # number of classes ( emotions )
         self.output_dim = len(self.e_config)
-  
 
         # optimization attributes
         self.optimizer = optimizer if optimizer else "adam"
@@ -276,7 +277,7 @@ class DeepEmotionRecognizer(EmotionRecognizer):
             )
 
         # reshape labels
-        if(self.y_train is  None or self.y_test is  None): 
+        if self.y_train is None or self.y_test is None:
             raise ValueError("y_train and y_test must be array_like ")
         y_train_shape = self.y_train.shape
         y_test_shape = self.y_test.shape
@@ -349,9 +350,9 @@ class DeepEmotionRecognizer(EmotionRecognizer):
 
     def predict_proba(self, audio_path):
         if self.classification_task:
-            feature = extract_feature_of_audio(audio_path, **self._f_config_dict).reshape(
-                (1, 1, self.input_length)
-            )
+            feature = extract_feature_of_audio(
+                audio_path, **self._f_config_dict
+            ).reshape((1, 1, self.input_length))
             proba = self.model.predict(feature)[0][0]
             result = {}
             for prob, emotion in zip(proba, self.e_config):
@@ -418,7 +419,7 @@ class DeepEmotionRecognizer(EmotionRecognizer):
         if partition == "test":
             if self.classification_task:
                 # np.squeeze去除数组中所有维度大小为 1 的维度，从而将数组的维度降低。如果数组没有大小为 1 的维度，则不会有任何变化。
-                #这里的y_test可能采用oneHotEncoder,因此可以如下计算
+                # 这里的y_test可能采用oneHotEncoder,因此可以如下计算
                 y_test = np.array(
                     [
                         np.argmax(y, axis=None, out=None) + 1
@@ -449,8 +450,12 @@ class DeepEmotionRecognizer(EmotionRecognizer):
         test_samples = []
         total = []
         for emotion in self.e_config:
-            n_train = self.count_samples_in_partition(self.emotions2int[emotion] + 1, "train")
-            n_test = self.count_samples_in_partition(self.emotions2int[emotion] + 1, "test")
+            n_train = self.count_samples_in_partition(
+                self.emotions2int[emotion] + 1, "train"
+            )
+            n_test = self.count_samples_in_partition(
+                self.emotions2int[emotion] + 1, "test"
+            )
             train_samples.append(n_train)
             test_samples.append(n_test)
             total.append(n_train + n_test)
@@ -463,7 +468,7 @@ class DeepEmotionRecognizer(EmotionRecognizer):
             data={"train": train_samples, "test": test_samples, "total": total},
             index=self.e_config + ["total"],
         )
-    
+
     def get_training_and_testing_samples_per_emotion(self):
         """
         Returns a dataframe with the number of training and testing samples
@@ -473,8 +478,12 @@ class DeepEmotionRecognizer(EmotionRecognizer):
         test_samples_per_emotion = []
         total_samples_per_emotion = []
         for emotion in self.e_config:
-            n_train = self.count_samples_in_partition(self.emotions2int[emotion] + 1, "train")
-            n_test = self.count_samples_in_partition(self.emotions2int[emotion] + 1, "test")
+            n_train = self.count_samples_in_partition(
+                self.emotions2int[emotion] + 1, "train"
+            )
+            n_test = self.count_samples_in_partition(
+                self.emotions2int[emotion] + 1, "test"
+            )
             train_samples_per_emotion.append(n_train)
             test_samples_per_emotion.append(n_test)
             total_samples_per_emotion.append(n_train + n_test)
@@ -492,7 +501,6 @@ class DeepEmotionRecognizer(EmotionRecognizer):
             index=self.e_config + ["Total"],
         )
 
-
     def get_random_emotion_index(self, emotion, partition="train"):
         """
         Returns random `emotion` data sample index on `partition`
@@ -517,21 +525,21 @@ class DeepEmotionRecognizer(EmotionRecognizer):
         return index
 
 
-
 ##
 if __name__ == "__main__":
     from config.MetaPath import ravdess
-    meta_dict = {
-        "train_dbs":ravdess,
-        "test_dbs":ravdess
-    }
+
+    meta_dict = {"train_dbs": savee, "test_dbs": emodb}
     print(meta_dict)
-    
-    der = DeepEmotionRecognizer(**meta_dict, emotions=e_config_def, verbose=0)
+
+    der = DeepEmotionRecognizer(
+        **meta_dict, e_config=e_config_def, f_config=f_config_def, epochs=400, verbose=0
+    )
     # #train
     der.train()
-    print("train_score",der.train_score())
-    print("test_score",der.test_score())
+    print("train_score", der.train_score())
+    print("test_score", der.test_score())
+
     # print(der.train_meta_files,der.test_meta_files)
     # der.train(override=False)
     # print("Test accuracy score:", der.test_score() * 100, "%")

release_note.md

@@ -21,7 +21,16 @@
 
 ### versions with time
 
+- **<u>input version under this line:</u>**
+  - new notes:"Sort update records in descending order of time".That would be convenient to update new version notes in the future.
+- 
+- 2023-05-09@19:10:05
+  - fix some bugs.
+  - certain adjustments to the project code.
 
+### old notes
+
+- old notes:"Sort update records in ascending order of time".
 
 - 2023-04-25@21:35:50
   - new:
@@ -41,4 +50,5 @@
 - 2023-04-30@00:02:03
   - "update PCA transformer to the SER system"
   - "there are some problems with recognize the single audio with pca preprocessing,the idea is to save the pca transformer when it was first fited,and then try to load corresponding file(joblib or pickle) when you need to recognize several audios that need extract the required dimension feature"
-- 
+- 
\ No newline at end of file
+

test_playground/tmp.py

-d = {'A': [1, 2, 3], 'B': [4, 5, 6], 'C':{"a":[7, 8, 9],"b":7}}
-import json
-
-def dict_to_filetag(d):
-    # Convert dictionary to JSON string
-    json_str = json.dumps(d)
-    remove_chars=['"',' ']
-    for c in remove_chars:
-        json_str=json_str.replace(c,'')
-    # Replace invalid characters with hyphen
-    rep_dict={
-        ":":"=",
-        # '"':'',
-        # "'":""
-    }
-    for char in json_str:
-        if rep_dict.get(char):
-            json_str = json_str.replace(char, rep_dict[char])
-    # Truncate string if too long
-    # max_len = 260
-    # if len(json_str) > max_len:
-    #     json_str = json_str[:max_len]
-
-    return json_str
-
-
-res=dict_to_filetag(d)
-print(res)
+import tkinter
+
+from matplotlib.backends.backend_tkagg import (
+    FigureCanvasTkAgg, NavigationToolbar2Tk)
+# Implement the default Matplotlib key bindings.
+from matplotlib.backend_bases import key_press_handler
+from matplotlib.figure import Figure
+
+import numpy as np
+
+
+root = tkinter.Tk()
+root.wm_title("Embedding in Tk")
+
+def get_fig():
+    """
+    Returns a matplotlib Figure object, a numpy array of time values, and a matplotlib Line2D object
+    representing the plot of 2*sin(2*pi*t) over the range [0,3] 
+    with a label of "f(t)" on the y-axis and "time [s]"
+    on the x-axis. 
+    The Figure has a size of 5x4 inches and a DPI of 100. No parameters are required.
+    """
+    fig = Figure(figsize=(5, 4), dpi=100)
+    ax = fig.add_subplot()
+    ax.set_xlabel("time [s]")
+    ax.set_ylabel("f(t)")
+
+    t = np.arange(0, 3, .01)
+    line, = ax.plot(t, 2 * np.sin(2 * np.pi * t))
+    return fig,t,line
+
+fig, t, line = get_fig()
+
+canvas = FigureCanvasTkAgg(fig, master=root)  # A tk.DrawingArea.
+canvas.draw()
+
+# pack_toolbar=False will make it easier to use a layout manager later on.
+toolbar = NavigationToolbar2Tk(canvas, root, pack_toolbar=False)
+toolbar.update()
+
+canvas.mpl_connect(
+    "key_press_event", lambda event: print(f"you pressed {event.key}"))
+canvas.mpl_connect("key_press_event", key_press_handler)
+
+button_quit = tkinter.Button(master=root, text="Quit", command=root.destroy)
+
+
+def update_frequency(new_val):
+    """
+    Update plot with new frequency value.
+
+    :param new_val: A float representing the new frequency value.
+    """
+
+    # retrieve frequency
+    f = float(new_val)
+
+    # update data
+    y = 2 * np.sin(2 * np.pi * f * t)
+    line.set_data(t, y)
+
+    # required to update canvas and attached toolbar!
+    canvas.draw()
+
+
+slider_update = tkinter.Scale(root, from_=1, to=5, orient=tkinter.HORIZONTAL,
+                              command=update_frequency, label="Frequency [Hz]")
+
+# Packing order is important. Widgets are processed sequentially and if there
+# is no space left, because the window is too small, they are not displayed.
+# The canvas is rather flexible in its size, so we pack it last which makes
+# sure the UI controls are displayed as long as possible.
+button_quit.pack(side=tkinter.BOTTOM)
+slider_update.pack(side=tkinter.BOTTOM)
+toolbar.pack(side=tkinter.BOTTOM, fill=tkinter.X)
+canvas.get_tk_widget().pack(side=tkinter.TOP, fill=tkinter.BOTH, expand=True)
+
+tkinter.mainloop()
\ No newline at end of file