Update handler.py

handler.py

@@ -2,17 +2,16 @@ from typing import  Dict
 from pyannote.audio import Pipeline
 from transformers.pipelines.audio_utils import ffmpeg_read
 import torch 
+import base64
+import numpy as np
 
 SAMPLE_RATE = 16000
 
-
-
 class EndpointHandler():
     def __init__(self, path=""):
         # load the model
         self.pipeline = Pipeline.from_pretrained("philschmid/pyannote-speaker-diarization-endpoint")
 
-
     def __call__(self, data: Dict[str, bytes]) -> Dict[str, str]:
         """
         Args:
@@ -23,14 +22,16 @@ class EndpointHandler():
         """
         # process input
         inputs = data.pop("inputs", data)
-        parameters = data.pop("parameters", None) #  min_speakers=2, max_speakers=5
+        parameters = data.pop("parameters", None) # min_speakers=2, max_speakers=5
 
+        # decode the base64 audio data
+        audio_data = base64.b64decode(inputs)
+        audio_nparray = np.frombuffer(audio_data, dtype=np.int16)
         
         # prepare pynannote input
-        audio_nparray = ffmpeg_read(inputs, SAMPLE_RATE)
         audio_tensor= torch.from_numpy(audio_nparray).unsqueeze(0)
         pyannote_input = {"waveform": audio_tensor, "sample_rate": SAMPLE_RATE}
-        
+
         # apply pretrained pipeline
         # pass inputs with all kwargs in data
         if parameters is not None:
@@ -45,3 +46,4 @@ class EndpointHandler():
         ]
         
         return {"diarization": processed_diarization}
+