Update app.py

app.py

@@ -87,7 +87,17 @@ def speaker_diarization():
     for turn, _, speaker in diarization.itertracks(yield_label=True):
         df = df.append({'start': turn.start, 'stop': turn.end, 'speaker': speaker}, ignore_index=True)
 
-    df.to_csv('result.csv', index=False)
+    # if the speaker is the same for 2 rows or more then merge them
+    new_df = pd.DataFrame(columns=['start', 'stop', 'speaker'])
+    for i in range(len(df)):
+        if i == 0:
+            new_df = new_df.append({'start': df['start'][i], 'stop': df['stop'][i], 'speaker': df['speaker'][i]}, ignore_index=True)
+        else:
+            if df['speaker'][i] == df['speaker'][i-1]:
+                new_df['stop'][len(new_df)-1] = df['stop'][i]
+            else:
+                new_df = new_df.append({'start': df['start'][i], 'stop': df['stop'][i], 'speaker': df['speaker'][i]}, ignore_index=True)
+    new_df.to_csv('result.csv', index=False)
 
 def save_audio_chunks(data_path, new_df):
     # load the audio file