add new funtion #82 and fixed bug #71

__init__.py

@@ -1,3 +1,11 @@
 import whisperx
 import faster_whisper_GUI
 
+# fixed bugs after complited with Nuitka
+from av.audio.codeccontext import AudioCodecContext
+from av.video.codeccontext import VideoCodecContext
+from av.data import stream
+from av.packet import Packet
+from av.subtitles import stream
+from av.subtitles import subtitle
+from av.subtitles import codeccontext

fasterWhisperGUIConfig.json

@@ -16,7 +16,7 @@
         "preciese": 0,
         "thread_num": "4",
         "num_worker": "1",
-        "download_root": "",
+        "download_root": "G:/Program Files (x86)/FasterWhisperGUI/cache",
         "local_files_only": false
     },
     "vad_param": {
@@ -38,10 +38,11 @@
         "themeColor": "#009faa"
     },
     "Transcription_param": {
-        "language": 0,
+        "aggregate_contents": true,
+        "language": 1,
         "task": false,
         "beam_size": "5",
-        "best_of": "6",
+        "best_of": "5",
         "patience": "1.0",
         "length_penalty": "1.0",
         "temperature": "0.0,0.2,0.4,0.6,0.8,1.0",
@@ -66,11 +67,11 @@
         "tabMovable": false,
         "tabScrollable": false,
         "tabShadowEnabled": false,
-        "tabMaxWidth": 220,
+        "tabMaxWidth": 367,
         "closeDisplayMode": 0,
-        "whisperXMinSpeaker": 2,
-        "whisperXMaxSpeaker": 2,
-        "outputFormat": 0,
+        "whisperXMinSpeaker": 3,
+        "whisperXMaxSpeaker": 3,
+        "outputFormat": 2,
         "outputEncoding": 1
     }
 }

faster_whisper_GUI/mainWindows.py

@@ -265,7 +265,7 @@ def getParam_model(self) -> dict:
         compute_type: str = self.page_model.preciese_combox.currentText()
         cpu_threads: int = int(self.page_model.LineEdit_cpu_threads.text().replace(" ", ""))
         num_workers: int = int(self.page_model.LineEdit_num_workers.text().replace(" ", ""))
-        download_root: str = self.page_model.LineEdit_download_root.text().replace(" ", "")
+        download_root: str = self.page_model.LineEdit_download_root.text().strip()
         local_files_only: bool = self.page_model.switchButton_local_files_only.isChecked()
         use_v3_model: bool = self.page_model.switchButton_use_v3.isChecked()
 
@@ -359,9 +359,7 @@ def getParamWhisperX(self) -> dict:
         if dict_WhisperXParams["min_speaker"] == 0 and dict_WhisperXParams["max_speaker"] == 0:
             dict_WhisperXParams["min_speaker"] = None
             dict_WhisperXParams["max_speaker"] = None
-        else:
-            dict_WhisperXParams["min_speaker"] = None
-            dict_WhisperXParams["max_speaker"] = None
+
 
         return dict_WhisperXParams
 
@@ -845,9 +843,10 @@ def outputSubtitleFile(self):
         output_dir = self.page_output.outputGroupWidget.LineEdit_output_dir.text()
         code_ = self.page_output.combox_output_code.currentText()
 
+        aggregate_contents_according_to_the_speaker = self.page_transcribes.switchButton_aggregate_contents_according_to_the_speaker.isChecked()
         result_to_write = self.current_result # self.result_faster_whisper if (self.result_whisperx_aligment is None and self.result_whisperx_speaker_diarize is None) else (self.result_whisperx_aligment if self.result_whisperx_speaker_diarize is None else self.result_whisperx_speaker_diarize)
 
-        self.outputWorker = OutputWorker(result_to_write, output_dir, format, code_,self)
+        self.outputWorker = OutputWorker(result_to_write, output_dir, format, code_, aggregate_contents_according_to_the_speaker , self)
         self.outputWorker.signal_write_over.connect(self.outputOver)    
         self.outputWorker.start()
         self.setStateTool(self.tr("保存文件"), self.tr("输出字幕文件"), False)
@@ -947,6 +946,9 @@ def whisperXDiarizeSpeakers(self):
 
         if self.whisperXWorker is None:
 
+            print(f"min_speaker: {whisperParams['min_speaker']}")
+            print(f"max_speaker: {whisperParams['max_speaker']}")
+
             self.whisperXWorker = WhisperXWorker(result_needed
                                                 , alignment=False
                                                 , speaker_diarize=True

faster_whisper_GUI/modelLoad.py

@@ -63,13 +63,20 @@ def run(self) -> None:
     def stop(self):
         self.isRunning = False
 
-    def loadModel(self,model_size_or_path:str=None):
+    def loadModel(self, model_size_or_path:str=None):
 
         model = None
         try:
             if model_size_or_path is None:
                 model_size_or_path = self.model_size_or_path
 
+            # 尝试替换空格，以处理带有空格的路径
+            # model_size_or_path = model_size_or_path.replace("\\", "/")
+            # model_size_or_path = model_size_or_path.replace(" ", "\ ")
+
+            # self.download_root = self.download_root.replace("\\", "/")
+            # self.download_root = self.download_root.replace(" ", "\ ")
+
             model = WhisperModel(
                                     model_size_or_path, 
                                     device=self.device, 

faster_whisper_GUI/tranccribePageNavigationInterface.py

@@ -166,6 +166,15 @@ def setupUI(self):
 
         widget_list.append(self.word_level_timestampels_param_widget)
 
+        # --------------------------------------------------------------------------------------------
+        self.switchButton_aggregate_contents_according_to_the_speaker = SwitchButton()
+        self.switchButton_aggregate_contents_according_to_the_speaker.setChecked(False)
+        self.aggregate_contents_param_widget = ParamWidget(self.__tr("根据说话人聚合内容"),
+                                                                self.__tr("按顺序讲相同说话人的内容聚合到一起，仅支持 txt 格式输出"),
+                                                                self.switchButton_aggregate_contents_according_to_the_speaker
+                                                            )
+        widget_list.append(self.aggregate_contents_param_widget)
+
         # =======================================================================================================
         self.titleLabel_auditory_hallucination = TitleLabel(self.__tr("幻听参数"))
         widget_list.append(self.titleLabel_auditory_hallucination)
@@ -472,6 +481,8 @@ def loadParamsFromFile(self):
 
     def setParam(self, Transcribe_params:dict) -> None:
 
+        self.switchButton_aggregate_contents_according_to_the_speaker.setChecked(Transcribe_params["aggregate_contents"])
+
         self.combox_language.setCurrentIndex(Transcribe_params["language"])
         # Transcribe_params["language"] = language_index
 
@@ -546,6 +557,8 @@ def getParam(self) -> dict:
         Transcribe_params = {}
 
         # 从数据模型获取文件列表
+
+        Transcribe_params["aggregate_contents"] = self.switchButton_aggregate_contents_according_to_the_speaker.isChecked()
 
         language = self.combox_language.currentIndex()
         Transcribe_params["language"] = language

faster_whisper_GUI/transcribe.py

@@ -127,6 +127,7 @@ def __init__(self,
                     output_dir:str, 
                     format:str, 
                     output_code = "UTF-8",
+                    aggregate_contents = False,
                     parent=None
                 ) -> None:
 
@@ -136,6 +137,7 @@ def __init__(self,
         self.format = format
         self.output_dir = output_dir
         self.output_code = output_code
+        self.aggregate_contents = aggregate_contents
 
     def stop(self):
         self.is_running = False
@@ -178,6 +180,7 @@ def run(self):
                             , language=info.language
                             , fileName=path
                             , file_code = self.output_code
+                            , aggregate_contents = self.aggregate_contents
                         )
 
         print("\n【Over】")
@@ -208,7 +211,7 @@ def __init__(self
         self.segments_path_info = []
 
 
-    def transcribe_file(self, file) -> (TranscriptionInfo, List):
+    def transcribe_file(self, file) -> (TranscriptionInfo, List): # type: ignore
         # try:
         #     is_av_file = self.try_decode_avFile(file)
         #     if not is_av_file:
@@ -388,15 +391,16 @@ def writeSubtitles(outputFileName:str,
                     format:str, 
                     language:str="",
                     fileName = "",
-                    file_code = "UTF-8"
+                    file_code = "UTF-8",
+                    aggregate_contents = False
                 ):
 
     file_code = ENCODING_DICT[file_code]
 
     if format == "SRT":
         writeSRT(outputFileName, segments, file_code = file_code)
     elif format == "TXT":
-        writeTXT(outputFileName, segments, file_code=file_code)
+        writeTXT(outputFileName, segments, file_code=file_code, aggregate_contents = aggregate_contents)
     elif format == "VTT":
         writeVTT(outputFileName, segments,language=language, file_code=file_code)
     elif format == "LRC":
@@ -608,21 +612,31 @@ def writeVTT(fileName:str, segments:List[segment_Transcribe],language:str,file_c
 
     vtt.save(fileName, file_code)
 
-def writeTXT(fileName:str, segments,file_code="utf8"):
+def writeTXT(fileName:str, segments, file_code="utf8", aggregate_contents=False):
     with codecs.open(fileName, "w", encoding=file_code) as f:
+        speaker_temp = ""
+
         for segment in segments:
+
             text:str = segment.text
             try:
                 speaker = segment.speaker + ": "
             except:
                 speaker = ""
-
-            text = speaker + text
+
+            if speaker_temp != speaker and aggregate_contents:
+                f.write(f"\n{speaker.encode('utf8').decode('utf8')} \n")
+                speaker_temp = speaker
+            elif not aggregate_contents:
+                text = speaker + text
 
             # 重编码为 utf-8 
             text:str = text.encode("utf8").decode("utf8")
+            f.write(f"{text} \n")
 
-            f.write(f"{text} \n\n")
+            # text = speaker + text
+
+
 
 def writeSRT(fileName:str, segments, file_code="UTF-8"):
     index = 1