Update ui/ui_core.py

ui/ui_core.py

@@ -3,6 +3,7 @@ import os
 import pandas as pd
 import pdfplumber
 import gradio as gr
+import re
 from typing import List
 
 # ✅ Fix: Add src to Python path
@@ -10,6 +11,9 @@ sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), "..",
 
 from txagent.txagent import TxAgent
 
+def sanitize_utf8(text: str) -> str:
+    return re.sub(r'[\ud800-\udfff]', '', text)
+
 def extract_all_text_from_csv_or_excel(file_path: str, progress=None, index=0, total=1) -> str:
     try:
         if not os.path.exists(file_path):
@@ -100,8 +104,9 @@ def create_ui(agent: TxAgent):
                             extracted_text += f"[Error processing file: {os.path.basename(path)}] — {str(file_error)}\n"
                             continue
 
+                    sanitized = sanitize_utf8(extracted_text.strip())
                     message = (
-                        f"{context}\n\n--- Uploaded File Content ---\n\n{extracted_text.strip()}\n\n--- End of File ---\n\nNow begin your reasoning:"
+                        f"{context}\n\n--- Uploaded File Content ---\n\n{sanitized}\n\n--- End of File ---\n\nNow begin your reasoning:"
                     )
 
                 generator = agent.run_gradio_chat(
@@ -131,7 +136,7 @@ def create_ui(agent: TxAgent):
                             if cleaned:
                                 yield cleaned
                         elif isinstance(update, str) and not update.strip().startswith("\ud83e\udde0"):
-                            yield update.encode("utf-8", "replace").decode("utf-8")
+                            yield sanitize_utf8(update.encode("utf-8", "replace").decode("utf-8"))
                     except Exception as update_error:
                         print(f"Error processing update: {update_error}")
                         continue
@@ -150,4 +155,4 @@ def create_ui(agent: TxAgent):
             ["Is there anything abnormal in the attached blood work report?"]
         ], inputs=message_input)
 
-    return demo
+    return demo