Update app.py

app.py

@@ -88,32 +88,27 @@ class AudioProcessor:
     def get_best_match(self, target_time, diarization_output):
         target_start_time = target_time['start_time']
         target_end_time = target_time['end_time']
-    
-        # 确保 diarization_output 是 Annotation 类型
-        if not isinstance(diarization_output, Annotation):
-            print(f"Error: Expected an Annotation object, but got {type(diarization_output)}")
-            return None  # 或者返回一个空的默认值
-    
+
         # 用于存储每个说话人时间段的重叠比例
         speaker_segments = []
         for segment, label in diarization_output.itertracks(yield_label=True):
             try:
                 start_seconds = segment.start
                 end_seconds = segment.end
-    
+
                 # 计算目标音频时间段和说话人时间段的重叠时间
                 overlap = self.calculate_overlap(target_start_time, target_end_time, start_seconds, end_seconds)
                 overlap_ratio = overlap / (end_seconds - start_seconds)
-    
+
                 # 记录说话人标签和重叠比例
                 speaker_segments.append((label, overlap_ratio, start_seconds, end_seconds))
-    
+
             except Exception as e:
                 print(f"处理行时出错: '{segment}'. 错误: {e}")
 
-    # 按照重叠比例排序，返回重叠比例最大的一段
-    best_match = max(speaker_segments, key=lambda x: x[1], default=None)
-    return best_match
+        # 按照重叠比例排序，返回重叠比例最大的一段
+        best_match = max(speaker_segments, key=lambda x: x[1], default=None)
+        return best_match
 
     # 获取该说话人除了目标语音时间段外的所有时间段
     def get_speaker_time_segments(self, diarization_output, target_time, speaker_label):
@@ -148,7 +143,7 @@ class AudioProcessor:
         # 执行说话人分离
         diarization_result = self.diarize_audio("final_output.wav")
 
-        if isinstance(diarization_result, str) and diarization_result.startswith("错误"):
+        if isinstance(diarization_result, Annotation) and diarization_result.startswith("错误"):
             return diarization_result, None  # 出错时返回错误信息
         else:
             # 获取最佳匹配的说话人标签和时间段