Update audio2text/a2t.py

audio2text/a2t.py

@@ -9,14 +9,14 @@ class A2T:
     def __init__(self, mic):
         self.mic = mic
 
-    def __transcribe(self, inputs, task: str = None):
+    def __generate_text(self, inputs, task: str = None):
         if inputs is  None:
-            print("Inputs None")
+            raise Exception("Inputs is None")
 
         transcribed_text = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=True)["text"]
         return transcribed_text
 
-    def __preprocces(self, raw):
+    def __preprocces(self, raw: bytes):
         print(f"Raw type : {type(raw)}")
         chunk = io.BytesIO(raw)
         audio, sample_rate = librosa.load(chunk, sr=16000)
@@ -28,12 +28,12 @@ class A2T:
             if self.mic is not None:
                 raw = self.mic
                 audio = self.__preprocces(raw=raw)
-                print(f"audio type : {type(audio)} \n shape : {audio.shape}")
+                print(f"audio type : {type(audio)} \n shape : {audio.shape} \n audio max value : {np.max(audio)}")
             else:
                 raise Exception("please provide audio")
 
             if isinstance(audio , np.ndarray):
-                return self.__transcribe(inputs=inputs, task=TASK)
+                return self.__generate_text(inputs=audio, task=TASK)
             else:
                 raise Exception("Audio is not np array")