Update app.py

app.py

@@ -322,79 +322,79 @@ import requests
 #     else:
 #         return {"error": response.json()}
 
-# def process_audio(audio):
-#     audio_file = open(audio, "rb")    
-#     try:
-#         audio = AudioSegment.from_file(audio_file)
-#     except CouldntDecodeError:
-#         os.remove(audio)
-#         return "Unsupported audio format"
-
-#     # بررسی مدت زمان فایل صوتی
-#     duration_seconds = len(audio_file) / 1000.0  # مدت زمان به ثانیه
-#     if duration_seconds > 900:
-#         os.remove(audio)
-#         return "Audio file is too long"
-
-#     project_root = os.path.dirname(os.path.dirname(__file__))
-#     voice_id = str(uuid.uuid4())
-#     # تبدیل به فرمت WAV سازگار با Whisper
-#     whisper_path = f'{voice_id}.wav'
-#     whisper_voice_path = os.path.join(project_root, whisper_path)
-    
-#     audio.export(whisper_voice_path, format='wav')
-    
-#     # ارسال درخواست به API Hugging Face
-#     output = query(whisper_voice_path)
-    
-#     # حذف فایل موقت
-#     os.remove(whisper_voice_path)
-    
-#     if "error" in output:
-#         print("Error:", output["error"])
-#         return "Transcription failed"
-#     else:
-#         text_question = output.get("text", "")
-#         print("text_question =", text_question)
-#         return text_question
-
 def process_audio(audio):
-    # باز کردن فایل صوتی ضبط‌شده
-    audio_file = open(audio, "rb")
-
-    # Load and convert the audio file
+    audio_file = open(audio, "rb")    
     try:
         audio = AudioSegment.from_file(audio_file)
-
     except CouldntDecodeError:
-        os.remove(audio_file)
+        os.remove(audio)
         return "Unsupported audio format"
 
-    # Check duration
-    duration_seconds = len(audio) / 1000.0  # Duration in seconds
+    # بررسی مدت زمان فایل صوتی
+    duration_seconds = len(audio_file) / 1000.0  # مدت زمان به ثانیه
     if duration_seconds > 900:
-        os.remove(whisper_voice_path)
+        os.remove(audio)
         return "Audio file is too long"
 
     project_root = os.path.dirname(os.path.dirname(__file__))
     voice_id = str(uuid.uuid4())
-    # Convert to WAV format compatible with Whisper
+    # تبدیل به فرمت WAV سازگار با Whisper
     whisper_path = f'{voice_id}.wav'
     whisper_voice_path = os.path.join(project_root, whisper_path)
     
     audio.export(whisper_voice_path, format='wav')
     
-    client = OpenAI(api_key=OPENAI_API_KEY)
+    # ارسال درخواست به API Hugging Face
+    output = query(whisper_voice_path)
     
-    with open(whisper_voice_path, 'rb') as audio_file:
-        text_question = client.audio.transcriptions.create(model="whisper-1", 
-                                                        file=audio_file, 
-                                                        response_format="text",
-                                                        language="fa")
     # حذف فایل موقت
     os.remove(whisper_voice_path)
-    print("text_question=", text_question)
-    return text_question
+    
+    if "error" in output:
+        print("Error:", output["error"])
+        return "Transcription failed"
+    else:
+        text_question = output.get("text", "")
+        print("text_question =", text_question)
+        return text_question
+
+# def process_audio(audio):
+#     # باز کردن فایل صوتی ضبط‌شده
+#     audio_file = open(audio, "rb")
+
+#     # Load and convert the audio file
+#     try:
+#         audio = AudioSegment.from_file(audio_file)
+
+#     except CouldntDecodeError:
+#         os.remove(audio_file)
+#         return "Unsupported audio format"
+
+#     # Check duration
+#     duration_seconds = len(audio) / 1000.0  # Duration in seconds
+#     if duration_seconds > 900:
+#         os.remove(whisper_voice_path)
+#         return "Audio file is too long"
+
+#     project_root = os.path.dirname(os.path.dirname(__file__))
+#     voice_id = str(uuid.uuid4())
+#     # Convert to WAV format compatible with Whisper
+#     whisper_path = f'{voice_id}.wav'
+#     whisper_voice_path = os.path.join(project_root, whisper_path)
+    
+#     audio.export(whisper_voice_path, format='wav')
+    
+#     client = OpenAI(api_key=OPENAI_API_KEY)
+    
+#     with open(whisper_voice_path, 'rb') as audio_file:
+#         text_question = client.audio.transcriptions.create(model="whisper-1", 
+#                                                         file=audio_file, 
+#                                                         response_format="text",
+#                                                         language="fa")
+#     # حذف فایل موقت
+#     os.remove(whisper_voice_path)
+#     print("text_question=", text_question)
+#     return text_question