fixed bugs #61

CheshireCC · Jan 9, 2024 · 22b1b9a · 22b1b9a
1 parent 3ddb113
commit 22b1b9a
Show file tree

Hide file tree

Showing 22 changed files with 439 additions and 227 deletions.
diff --git a/FasterWhisperGUI.py b/FasterWhisperGUI.py
@@ -25,15 +25,6 @@
 faster_whisper_logger_handler.setFormatter(formatter1)
 logger_faster_whisper.addHandler(faster_whisper_logger_handler)
 
-# pyside6 日志
-# logger_pyside = logging.getLogger("faster_whisper_GUI")
-# logger_pyside.setLevel(logging.DEBUG)
-# pyside_logger_handler = logging.FileHandler(r"./pyside.log", mode="w")
-# pyside_logger_handler.setLevel(logging.DEBUG)
-# formatter2 = logging.Formatter("%(asctime)s - %(name)s - %(levelname)s - %(message)s",datefmt='%Y-%m-%d_%H:%M:%S')
-# pyside_logger_handler.setFormatter(formatter2)
-# logger_pyside.addHandler(pyside_logger_handler)
-
 # # 将默认的递归深度修改为3000
 # sys.setrecursionlimit(7000)  
 
@@ -43,20 +34,17 @@
 
 from qfluentwidgets import ProgressBar
 
-
 class MySplashScreen(QSplashScreen):
     # 鼠标点击事件
     def mousePressEvent(self, event):
         pass
 
 from resource import rc_Image
 
-
 # 启动一个Qt程序，并使用传入的系统参数
 app = QApplication(sys.argv)
 app.setObjectName("FasterWhisperGUIAPP")
 
-
 #设置启动界面
 splash = MySplashScreen()
 
@@ -84,8 +72,6 @@ def mousePressEvent(self, event):
 # 显示启动界面
 splash.show()
 
-
-
 # app.processEvents()  # 处理主进程事件
 
 import os

diff --git a/config/config.json b/config/config.json
@@ -0,0 +1,6 @@
+{
+    "QFluentWidgets": {
+        "ThemeColor": "#ffdb4437",
+        "ThemeMode": "Light"
+    }
+}
diff --git a/fasterWhisperGUIConfig.json b/fasterWhisperGUIConfig.json
@@ -8,15 +8,15 @@
     "model_param": {
         "localModel": true,
         "onlineModel": false,
-        "model_path": "F:/WhisperModels/faster-whisper/large-v3-float32",
+        "model_path": "",
         "modelName": 0,
         "use_v3_model": true,
         "device": 1,
         "deviceIndex": "0",
         "preciese": 5,
         "thread_num": "4",
         "num_worker": "1",
-        "download_root": "C:/Users/12059/.cache/huggingface/hub",
+        "download_root": "",
         "local_files_only": false
     },
     "vad_param": {
@@ -34,16 +34,17 @@
         "language": 2,
         "huggingface_user_token": "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",
         "autoGoToOutputPage": 2,
-        "autoClearTempFiles": true
+        "autoClearTempFiles": true,
+        "themeColor": "#009faa"
     },
     "Transcription_param": {
-        "language": 0,
+        "language": 1,
         "task": false,
         "beam_size": "5",
         "best_of": "1",
         "patience": "1.0",
         "length_penalty": "1.0",
-        "temperature": "0.5",
+        "temperature": "0.3",
         "compression_ratio_threshold": "2.4",
         "log_prob_threshold": "-1.0",
         "no_speech_threshold": "0.6",
@@ -65,7 +66,7 @@
         "tabMovable": false,
         "tabScrollable": true,
         "tabShadowEnabled": false,
-        "tabMaxWidth": 300,
+        "tabMaxWidth": 403,
         "closeDisplayMode": 0,
         "whisperXMinSpeaker": 2,
         "whisperXMaxSpeaker": 2,

diff --git a/faster_whisper_GUI/UI_MainWindows.py b/faster_whisper_GUI/UI_MainWindows.py
@@ -27,6 +27,7 @@
                             NavigationAvatarWidget,
                             NavigationInterface
                             , setTheme
+                            , setThemeColor
                             , Theme
                             , FluentIcon
                             , NavigationItemPosition
@@ -86,9 +87,20 @@ def tr(self, text):
     #     app.installTranslator(self.translator)
 
 
-    def readConfigJson(self):
+    def readConfigJson(self, config_file_path: str = ""):
+        self.default_theme = "light"
+        self.model_param = {}
+        self.setting = {}
+        self.demucs = {}
+        self.Transcription_param = {}
+        self.output_whisperX_param = {}
+        self.vad_param = {}
+
+        if not config_file_path:
+            return
+
         self.use_auth_token_speaker_diarition= ""
-        with open(r"./fasterWhisperGUIConfig.json","r", encoding="utf8") as fp:
+        with open(os.path.abspath(config_file_path),"r", encoding="utf8") as fp:
             json_data = json.load(fp)
 
             try:
@@ -130,6 +142,7 @@ def readConfigJson(self):
     def setConfig(self):
 
         setTheme(Theme.DARK if self.default_theme == "dark" else Theme.LIGHT)
+        # setThemeColor("#aaff009f")
         if self.model_param != {}:
             self.page_model.setParam(self.model_param)
             if not self.model_param["download_root"]:
@@ -165,7 +178,6 @@ def __init__(self, translator:QTranslator=None):
         else:
             self.translator = translator
 
-        # TODO: 无边框窗口方案需要等待 pyside6>6.5.0 支持
         # self.setWindowFlags(Qt.FramelessWindowHint)
         # self.setAttribute(Qt.WA_TranslucentBackground)  
 
@@ -186,17 +198,24 @@ def __init__(self, translator:QTranslator=None):
         self.initWin()
 
         # 读配置文件
-        self.readConfigJson()
+        self.readConfigJson(r"./fasterWhisperGUIConfig.json")
         # 设置配置
         self.setConfig()
 
+        try:
+            self.setWidgetsStatusFromConfig()
+        except Exception as e:
+            print(str(e))
+
+    def setWidgetsStatusFromConfig(self):
         # 根据读取的配置设置完控件状态之后，根据控件状态设置相关属性
         self.page_output.tableTab.onDisplayModeChanged(self.page_output.tableTab.closeDisplayModeComboBox.currentIndex())
         self.page_output.tableTab.tabBar.setMovable(self.page_output.tableTab.movableCheckBox.isChecked())
         self.page_output.tableTab.tabBar.setScrollable(self.page_output.tableTab.scrollableCheckBox.isChecked())
         self.page_output.tableTab.tabBar.setTabShadowEnabled(self.page_output.tableTab.shadowEnabledCheckBox.isChecked())
         self.page_output.tableTab.tabBar.setTabMaximumWidth(self.page_output.tableTab.tabMaxWidthSpinBox.value())
 
+
     def initWin(self):
 
         self.setObjectName("FramlessMainWin")

diff --git a/faster_whisper_GUI/config.py b/faster_whisper_GUI/config.py
@@ -154,3 +154,31 @@
                     "GB2312":"gb18030", 
                     "ANSI":"ansi"
                 }
+
+THEME_COLORS = [
+    "#009faa",
+    "#ff009f",
+    "#84BE84",
+    "#aaff00",
+    "#FF9500",
+    "#00CD00",
+    "#DB4437",
+    "#23CD5E",
+    "#E61D34",
+    "#00FF00",
+    "#FF00FF",
+    "#1ABC9C",
+    "#FF3300",
+    "#FFFF00",
+    "#FFC019",
+    "#FF6600",
+    "#00FFFF",
+    "#FF7A1D",
+    "#E71A1B",
+    "#FF8800",
+    "#3388FF",
+    "#F4B400",
+    "#0069B7",
+    "#FFCC00",
+    "#0078D4",
+]
diff --git a/faster_whisper_GUI/de_mucs.py b/faster_whisper_GUI/de_mucs.py
@@ -55,7 +55,7 @@ def run(self) -> None:
         self.file_process_status.emit({"file":"", "status":False, "task": "load model"})
         if self.model is None:
             try:
-                self.loadModel(self.model_path,device=device)
+                self.loadModel(self.model_path, device=device)
             except Exception as e:
                 print(f"load model error: \n{str(e)}")
                 self.signal_vr_over.emit(False)
@@ -78,7 +78,9 @@ def run(self) -> None:
 
             try:
                 samples = self.reSampleAudio(audio, 44100, device=device)
-                samples = torch.tensor(samples,dtype=torch.float32).to(device)
+                samples = np.asarray(samples)
+                print("samples shape: ", samples.shape)
+                # samples = torch.tensor(samples,dtype=torch.float32).to(device)
             except Exception as e:
                 print(f"resample audio error:\n{str(e)}")
                 self.signal_vr_over.emit(False)
@@ -96,15 +98,16 @@ def run(self) -> None:
             self.file_process_status.emit({"file":audio, "status":False, "task": "separate sources"})
 
             try:
-                sources = self.separate_sources(self.model, 
+                sources = self.separate_sources(
+                                                self.model, 
                                                 samples[None], 
                                                 self.segment, 
                                                 self.overlap, 
                                                 device, 
                                                 self.sampleRate
                                             )
             except Exception as e:
-                print(f"separate audio sources error:\n{str(e)}")
+                print(f"\nseparate audio sources error:\n    {str(e)}")
                 self.signal_vr_over.emit(False)
                 self.stop()
 
@@ -192,7 +195,8 @@ def separate_sources(self,
                 be on `device`, while the entire tracks will be stored on `mix.device`.
         """
         if device is None:
-            device = mix.device
+            device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+            model.to(device)
         else:
             device = device
 
@@ -211,6 +215,7 @@ def separate_sources(self,
                 return None
 
             chunk = mix[:, :, start:end]
+            chunk = torch.tensor(chunk, dtype=torch.float32).to(device)
             with torch.no_grad():
                 out = model.forward(chunk)