fixed transcription

app.py

@@ -59,6 +59,7 @@ def reset_member():
     global total_audio
     print(total_audio)
     process.delete_files_in_directory(total_audio)
+    process.delete_files_in_directory('/tmp/data/transcription_audio')
     try:
         data = request.get_json()
         if not data or "names" not in data:
@@ -97,11 +98,10 @@ def transcription():
     global transcription_text
     global total_audio
     try: 
-        audio_directory = transcripter.merge_segments(total_audio)
+        audio_directory = transcripter.merge_segments(total_audio,'/tmp/data/transcription_audio')
         transcription_text = transcripter.create_transcription(audio_directory)
         with open(transcription_text,'r',encoding='utf-8') as file:
             file_content = file.read()
-            print(file_content)
         return jsonify({'transcription': file_content}),200
     except Exception as e:
         return jsonify({"error": str(e)}),500
@@ -110,6 +110,13 @@ def transcription():
 @app.route('/analyze',methods =['GET','POST'])
 def analyze():
     global transcription_text
+    if transcription_text == "":
+        try: 
+            audio_directory = transcripter.merge_segments(total_audio,'/tmp/data/transcription_audio')
+            transcription_text = transcripter.create_transcription(audio_directory)
+        except Exception as e:
+            return jsonify({'error making transcription': e})
+    
     analyzer = TextAnalyzer(transcription_text, harassment_keywords)
     api_key = os.environ.get("DEEPSEEK")
     if api_key is None:

static/feedback.js

@@ -15,8 +15,6 @@ async function getAnalysis() {
   const loader = document.getElementById("loader");
   loader.style.display = "block";
   try {
-    await getTranscription();
-
     const response = await fetch("/analyze");
     if (!response.ok) {
       throw new Error(`HTTP error! status: ${response.status}`);

transcription.py

@@ -19,7 +19,6 @@ class TranscriptionMaker():
         self.output_dir = output_dir
         os.makedirs(self.output_dir, exist_ok=True)
 
-
     #音声ファイルのディレクトリを受け取り、書き起こしファイルを作成する
     def create_transcription(self,audio_directory):
         results = []
@@ -45,6 +44,7 @@ class TranscriptionMaker():
                     "end": segment.end,
                     "text": segment.text
                 })
+
         #ファイルの書き込み。ファイル名は"transcription.txt"
         output_file=os.path.join(self.output_dir,"transcription.txt")
         try:
@@ -56,54 +56,30 @@ class TranscriptionMaker():
             raise
         return output_file
     
-    #ファイル名が連続しているならくっつける
-    def merge_segments(self,segments_dir,output_dir = "/tmp/data/merged_segment"):
+    #ディレクトリ内の音声ファイルをくっつける
+    def merge_segments(self, segments_dir, output_dir="/tmp/data/merged_segment"):
         if not os.path.exists(output_dir):
             os.makedirs(output_dir, exist_ok=True)
 
         files = sorted([f for f in os.listdir(segments_dir) if f.endswith('.wav')])
 
-        merged_files = []
-        current_group = []
-        previous_index = None
+        if len(files) <= 1:
+            print('No need to merge')
+            return output_dir
+
+        combined_audio = AudioSegment.empty()
 
         for file in files:
-            # ファイル名から番号を抽出（例: "0.wav" -> 0）
-            file_index = int(file.split('.')[0])
-        
-            # 番号が連続していない場合、新しいグループを作成
-            if previous_index is not None and file_index != previous_index + 1:
-                # 現在のグループを結合して保存
-                if current_group:
-                    merged_files.append(current_group)
-                current_group = []
-        
-            # 現在のファイルをグループに追加
-            current_group.append(file)
-            previous_index = file_index
+            file_path = os.path.join(segments_dir, file)
+            segment = AudioSegment.from_file(file_path)
+            combined_audio += segment
 
-        # 最後のグループを追加
-        if current_group:
-            merged_files.append(current_group)
-    
-        # グループごとに結合して保存
-        for i, group in enumerate(merged_files):
-            combined_audio = AudioSegment.empty()
-            for file in group:
-                file_path = os.path.join(segments_dir, file)
-                segment = AudioSegment.from_file(file_path)
-                combined_audio += segment
-            # 出力ファイル名を設定して保存
-            output_file = os.path.join(output_dir, self.generate_filename(3))
-            combined_audio.export(output_file, format='wav')
+        output_file = os.path.join(output_dir, self.generate_filename())
 
+        combined_audio.export(output_file, format="wav")
         return output_dir
-    
-    def generate_random_string(self,length):
-        letters = string.ascii_letters + string.digits
-        return ''.join(random.choice(letters) for i in range(length))
 
-    def generate_filename(self,random_length):
+    def generate_filename(self):
         current_time = datetime.now().strftime("%Y%m%d%H%M%S")
         filename = f"{current_time}.wav"
         return filename