Update app.py

app.py

@@ -7,6 +7,7 @@ import edge_tts
 import asyncio
 import os
 import io
+import tempfile
 
 
 # Initialize session state for storing data
@@ -20,8 +21,8 @@ if 'story_zh' not in st.session_state:
     st.session_state.story_zh = None
 if 'audio_generated' not in st.session_state:
     st.session_state.audio_generated = False
-if 'audio_data' not in st.session_state:
-    st.session_state.audio_data = None
+if 'audio_path' not in st.session_state:
+    st.session_state.audio_path = None
 
 
 # function part
@@ -65,21 +66,23 @@ async def text2audio_cantonese(text):
         voice = "zh-HK-HiuMaanNeural"  # Female Cantonese voice
         # Alternative: "zh-HK-WanLungNeural" for male voice
         
+        # Create a temporary file
+        temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3")
+        temp_file.close()
+        
+        # Configure edge-tts to save to the file path
         communicate = edge_tts.Communicate(text, voice)
+        await communicate.save(temp_file.name)
         
-        # Save to BytesIO object
-        audio_bytes = io.BytesIO()
-        await communicate.save(audio_bytes)
-        audio_bytes.seek(0)  # Reset the pointer to the start
-
+        # Return the path to the audio file
         return {
-            'audio': audio_bytes,
+            'path': temp_file.name,
             'success': True
         }
     except Exception as e:
         st.error(f"音頻製作出左問題: {str(e)}")
         return {
-            'audio': None,
+            'path': None,
             'success': False
         }
 
@@ -350,12 +353,14 @@ if uploaded_file is not None:
             if not st.session_state.audio_generated:
                 # Need to run async function with asyncio
                 audio_result = asyncio.run(text2audio_cantonese(st.session_state.story_zh))
-                st.session_state.audio_data = audio_result['audio']
+                st.session_state.audio_path = audio_result['path']
                 st.session_state.audio_generated = audio_result['success']
                 
             # Play the audio
-            if st.session_state.audio_data:
-                st.audio(st.session_state.audio_data, format="audio/mp3")
+            if st.session_state.audio_path and os.path.exists(st.session_state.audio_path):
+                with open(st.session_state.audio_path, "rb") as audio_file:
+                    audio_bytes = audio_file.read()
+                st.audio(audio_bytes, format="audio/mp3")
             else:
                 st.error("哎呀！再試多次啦！")
 
@@ -364,12 +369,19 @@ if uploaded_file is not None:
         os.remove(temp_file_path)
 else:
     # Clear session state when no file is uploaded
+    # Also clean up any temporary audio files
+    if st.session_state.audio_path and os.path.exists(st.session_state.audio_path):
+        try:
+            os.remove(st.session_state.audio_path)
+        except:
+            pass
+            
     st.session_state.scenario = None
     st.session_state.scenario_zh = None
     st.session_state.story = None
     st.session_state.story_zh = None
     st.session_state.audio_generated = False
-    st.session_state.audio_data = None
+    st.session_state.audio_path = None
     
     # Welcome message in Cantonese
     st.markdown("""