json format subtitles support

fasterWhisperGUIConfig.json

@@ -16,7 +16,7 @@
         "preciese": 0,
         "thread_num": "4",
         "num_worker": "1",
-        "download_root": "G:/Program Files (x86)/FasterWhisperGUI/cache",
+        "download_root": "C:/Users/12059/.cache/huggingface/hub",
         "local_files_only": false
     },
     "vad_param": {
@@ -39,7 +39,7 @@
     },
     "Transcription_param": {
         "aggregate_contents": true,
-        "language": 1,
+        "language": 3,
         "task": false,
         "beam_size": "5",
         "best_of": "5",
@@ -67,11 +67,11 @@
         "tabMovable": false,
         "tabScrollable": false,
         "tabShadowEnabled": false,
-        "tabMaxWidth": 366,
+        "tabMaxWidth": 257,
         "closeDisplayMode": 0,
         "whisperXMinSpeaker": 2,
         "whisperXMaxSpeaker": 3,
-        "outputFormat": 4,
+        "outputFormat": 0,
         "outputEncoding": 1
     }
 }

faster_whisper_GUI/config.py

@@ -137,7 +137,7 @@
 
 STR_BOOL = {"False" : False, "True" : True}
 
-SUBTITLE_FORMAT = ["SRT", "TXT", "VTT", "LRC", "SMI"]
+SUBTITLE_FORMAT = ["SRT", "TXT", "VTT", "LRC", "SMI", "JSON"]
 
 CAPTURE_PARA = [
     {"rate": 44100

faster_whisper_GUI/mainWindows.py

@@ -58,7 +58,7 @@
 from .whisper_x import WhisperXWorker
 from .de_mucs import DemucsWorker
 # from .style_sheet import StyleSheet
-from .subtitleFileRead import readSRTFileToSegments
+from .subtitleFileRead import readSRTFileToSegments, readJSONFileToSegments
 from .config import ENCODING_DICT
 from .util import outputWithDateTime
 from .split_audio import SplitAudioFileWithSpeakersWorker
@@ -1133,36 +1133,52 @@ def openExcitedFiles(self):
         dataDir,_ = os.path.split(file)
         self.page_process.fileNameListView.avDataRootDir = dataDir
         # filesList = os.listdir(dataDir)
+
+        # json 格式字幕文件为首选
+        file_subtitle_fileName = ".".join(file.split(".")[:-1]+["json"])
+        ext_ = "json"
+        # 检测 json 格式字幕文件存在性
+        if not os.path.exists(file_subtitle_fileName):
+            # 没有 json 格式字幕文件的时候将会尝试获取 srt 格式字幕
+            file_subtitle_fileName = ".".join(file.split(".")[:-1]+["srt"])
+            ext_ = "srt"
 
-        file_subtitle_fileName = ".".join(file.split(".")[:-1]+["srt"])
-        fileName_subtitle_without_Ext = '.'.join(os.path.split(file_subtitle_fileName)[-1].split('.')[:-1])
+        # fileName_subtitle_without_Ext = '.'.join(os.path.split(file_subtitle_fileName)[-1].split('.')[:-1])
 
+        # 当字幕文件目录所指向的文件存在时
         if os.path.exists(file_subtitle_fileName):
             print(f"find existed srt file: {file_subtitle_fileName}")
+            # 获取文件的后缀名
+            # ext_ = file_subtitle_fileName.split(".")[-1]
+
         else:
-            file_subtitle_fileName,_ = QFileDialog.getOpenFileName(
+            file_subtitle_fileName,ext_ = QFileDialog.getOpenFileName(
                                                                     self, 
                                                                     self.tr("选择字幕文件"), 
                                                                     file_subtitle_fileName, 
                                                                     # self.page_process.fileNameListView.avDataRootDir, 
-                                                                    "SRT file(*.srt)",
-
+                                                                    "JSON file(*.json);;SRT file(*.srt)",
                                                                 )
+            # print(ext_)
+
             if file_subtitle_fileName and os.path.isfile(file_subtitle_fileName):
                 print(f"get srt file: {file_subtitle_fileName}")
             else:
-                messageBoxDia_ = MessageBox(self.tr("无效文件"),self.tr("必须要有有效的字幕文件"),self)
+                messageBoxDia_ = MessageBox(self.tr("没有字幕文件"),self.tr("必须要有有效的字幕文件"),self)
                 messageBoxDia_.show()
                 return
 
         code_ = self.page_output.combox_output_code.currentText()
 
         try:
-            segments = readSRTFileToSegments(file_subtitle_fileName, file_code=ENCODING_DICT[code_])
-        except UnicodeDecodeError as e:
-            print("read subtitle file failed")
-            print(str(e))
-            self.raiseErrorInfoBar(self.tr("读取字幕文件失败"), self.tr("读取失败 文件可能使用了不同的编码\n检查 fasterwhispergui.log 文件可能会获取更多信息"))
+            if ext_ in ["JSON file(*.json)" ,"json"]:
+                segments = readJSONFileToSegments(file_subtitle_fileName, file_code=ENCODING_DICT[code_])
+            else:
+                segments = readSRTFileToSegments(file_subtitle_fileName, file_code=ENCODING_DICT[code_])
+        except Exception as e:
+            print("read subtitle file failed:")
+            print(f"    {str(e)}")
+            self.raiseErrorInfoBar(self.tr("读取失败"), self.tr("读取字幕文件失败 \n检查日志文件可能会获取更多信息"))
             return
         # 输出字幕文件内容
         # for segment in segments:
@@ -1509,9 +1525,11 @@ def deleteResultTableEvent(self, routeKey:str):
         self.outputWithDateTime("deleteTable")
 
         print(f"len_DataModel:{len(self.tableModel_list)}")
-        # print(routeKey)
+        for tb in self.tableModel_list.items():
+            print(f"    {tb[0]}")
+        print(f"data to delete: {routeKey}")
         file_key ="_".join(routeKey.split("_")[1:]) 
-        print(file_key)
+        print(f"key: {file_key}")
         self.tableModel_list.pop(file_key)
         print(f"len_DataModel_after_pop:{len(self.tableModel_list)}")
 

faster_whisper_GUI/subtitleFileRead.py

@@ -1,7 +1,25 @@
 # coding:utf-8
 
+import os
 from typing import List
 from .seg_ment import segment_Transcribe
+import json
+from faster_whisper import Word
+
+def readJSONFileToSegments(file:str, file_code = "utf8") -> List[segment_Transcribe]:
+
+    with open(os.path.abspath(file),"r", encoding= file_code) as fp:
+        subtitles_str = json.load(fp=fp)["body"]
+    segments = [ segment_Transcribe(
+                                start=subtitle["from"], 
+                                end=subtitle["to"], 
+                                text=subtitle["content"],
+                                words=[ Word(word["start"],word["end"],word["word"],word["probability"]) for word in subtitle["words"]],
+                                speaker=subtitle["speaker"] or None
+                            ) for subtitle in subtitles_str
+                ]
+    return segments
+
 
 def readSRTFileToSegments(file:str, file_code = "utf8") -> List[segment_Transcribe]:
 

faster_whisper_GUI/tableViewInterface.py

@@ -248,13 +248,15 @@ def onCurrentIndexChanged(self, index):
     def removeTab(self, index):
 
         item = self.tabBar.tabItem(index)
+        print(f"removeTab: {item.routeKey()}")
+
         self.signal_delete_table.emit(item.routeKey())
 
         widget = self.findChild(TableView, item.routeKey())
-
+        # print(f"table:{widget}")
         self.stackedWidget.removeWidget(widget)
         self.tabBar.removeTab(index)
-        widget.deleteLater()
+        # widget.deleteLater()
 
 
 class CustomTableView(TableView):

faster_whisper_GUI/transcribe.py

@@ -9,6 +9,7 @@
 import torch
 import numpy as np 
 import av
+import json
 
 from faster_whisper import (
                             WhisperModel
@@ -410,8 +411,38 @@ def writeSubtitles(outputFileName:str,
         wirteLRC(outputFileName, segments,language=language, file_code=file_code)
     elif format == "SMI":
         writeSMI(outputFileName, segments, language=language, avFile=fileName, file_code=file_code)
+    elif format == "JSON":
+        writeJson(outputFileName, segments, language=language, avFile=fileName, file_code=file_code)
+
     print(f"write over | {outputFileName}")
 
+def writeJson(fileName:str,segments:List[segment_Transcribe],language:str,avFile="",file_code="utf8"):
+    result = {}
+    result["language"] = language
+    result["file"]=avFile
+
+    result["body"] = []
+
+    for segment in segments:
+        result["body"].append(
+                                {
+                                    "from":segment.start,
+                                    "to":segment.end,
+                                    "location":2,
+                                    "content":segment.text,
+                                    "speaker":segment.speaker or "",
+                                    "words":[{"start":word.start,"end":word.end,"word":word.word,"probability":word.probability }for word in segment.words]
+                                }
+                            )
+    with open(os.path.abspath(fileName),'w',encoding=file_code) as fp:
+
+        json.dump(
+                    result,
+                    fp,
+                    ensure_ascii=False,
+                    indent=4
+                )
+
 
 def writeSMI(fileName:str, segments:List[segment_Transcribe], language:str, avFile = "",file_code="utf8"):