Update app.py

app.py

@@ -4,6 +4,8 @@ import edge_tts
 import time
 import os
 import uuid
+import re
+import html
 import firebase_admin
 from firebase_admin import credentials, firestore
 from openai import OpenAI
@@ -64,7 +66,6 @@ header {visibility: hidden;}
     margin-top: 0.1em;
     position: relative;
 }
-
 .clear-chat-btn-top {
     position: absolute;
     top: 10px;
@@ -77,21 +78,17 @@ header {visibility: hidden;}
     z-index: 1000;
     transition: color 0.2s ease;
 }
-
 .clear-chat-btn-top:hover {
     color: #fff;
 }
-
 .stChatMessage { max-width: 85%; border-radius: 12px; padding: 8px; margin-bottom: 10px; }
 .stChatMessage[data-testid="stChatMessage-user"] { background: #f0f0f0; color: #000000; }
 .stChatMessage[data-testid="stChatMessage-assistant"] { background: #e3f2fd; color: #000000; }
-
 .chat-history-wrapper {
     margin-top: 0.5em;
-    padding-bottom: 9em; /* enough space so input does not cover messages */
+    padding-bottom: 9em;
     min-height: 60vh;
 }
-
 .input-bottom-bar {
     position: fixed;
     bottom: 3.5em;
@@ -180,6 +177,22 @@ def display_chat_history():
     st.markdown('<div class="chat-history-wrapper">' + "".join(chat_msgs) + '</div>', unsafe_allow_html=True)
     st.markdown('<div id="chat-top-anchor"></div>', unsafe_allow_html=True)
 
+# --- TTS sanitize
+def sanitize_for_tts(text):
+    text = html.unescape(text)
+    text = re.sub(r'[^\x00-\x7F]+', ' ', text)
+    text = re.sub(r'\[([^\]]+)\]\([^\)]+\)', r'\1', text)
+    text = re.sub(r'(\*\*|__)(.*?)\1', r'\2', text)
+    text = re.sub(r'(\*|_)(.*?)\1', r'\2', text)
+    text = re.sub(r'^#{1,6}\s+', '', text, flags=re.MULTILINE)
+    text = re.sub(r'^\s*[-*+]\s+', ' • ', text, flags=re.MULTILINE)
+    text = re.sub(r'^\s*\d+\.\s+', ' • ', text, flags=re.MULTILINE)
+    text = re.sub(r'[!?]{2,}', '.', text)
+    text = re.sub(r'\.{3,}', '.', text)
+    text = re.sub(r'\n{2,}', '. ', text)
+    text = re.sub(r'\s+', ' ', text).strip()
+    return text
+
 # --- Edge TTS synth
 async def edge_tts_synthesize(text, voice, user_id):
     out_path = f"output_{user_id}.mp3"
@@ -241,7 +254,8 @@ if user_input:
     mute_voice = st.session_state.get("mute_voice", False)
     audio_path = None
     if not mute_voice and assistant_message.strip():
-        audio_path = synthesize_voice(assistant_message, st.session_state["selected_voice"], user_id)
+        clean_text = sanitize_for_tts(assistant_message)
+        audio_path = synthesize_voice(clean_text, st.session_state["selected_voice"], user_id)
         st.session_state["last_audio_path"] = audio_path
 
     time.sleep(0.2)