Improve text formatting in text_to_speech function to handle punctuation correctly

app.py

@@ -76,7 +76,11 @@ async def text_to_speech(text, voice, rate, pitch, generate_subtitles=False):
                 phrase_start = start_time
                 
             current_phrase.append(boundary)
-            current_text += word + " "
+            
+            if word in ['.', ',', '!', '?', ':', ';'] or word.startswith(('.', ',', '!', '?', ':', ';')):
+                current_text = current_text.rstrip() + word + " "
+            else:
+                current_text += word + " "
             
             # Determine if we should end this phrase and start a new one
             should_break = False