Fix type issue

modules/diarize/diarize_pipeline.py

@@ -7,6 +7,7 @@
 from typing import Optional, Union
 import torch
 
+from modules.whisper.data_classes import *
 from modules.utils.paths import DIARIZATION_MODELS_DIR
 from modules.diarize.audio_loader import load_audio, SAMPLE_RATE
 
@@ -44,7 +45,8 @@ def __call__(self, audio: Union[str, np.ndarray], min_speakers=None, max_speaker
 def assign_word_speakers(diarize_df, transcript_result, fill_nearest=False):
     transcript_segments = transcript_result["segments"]
     for seg in transcript_segments:
-        seg = seg.dict()
+        if isinstance(seg, Segment):
+            seg = seg.model_dump()
         # assign speaker to segment (if any)
         diarize_df['intersection'] = np.minimum(diarize_df['end'], seg['end']) - np.maximum(diarize_df['start'],
                                                                                             seg['start'])
@@ -64,7 +66,7 @@ def assign_word_speakers(diarize_df, transcript_result, fill_nearest=False):
             seg["speaker"] = speaker
 
         # assign speaker to words
-        if 'words' in seg:
+        if 'words' in seg and seg['words'] is not None:
             for word in seg['words']:
                 if 'start' in word:
                     diarize_df['intersection'] = np.minimum(diarize_df['end'], word['end']) - np.maximum(
@@ -89,7 +91,7 @@ def assign_word_speakers(diarize_df, transcript_result, fill_nearest=False):
     return transcript_result
 
 
-class Segment:
+class DiarizationSegment:
     def __init__(self, start, end, speaker=None):
         self.start = start
         self.end = end