Update app.py

app.py

@@ -157,7 +157,51 @@ def save_feedback_to_excel(name, email, feedback):
         st.error(f"Error saving feedback to Excel: {e}")
 
 
-
+def clean_text(text):
+    # Replace escaped newlines with actual newlines
+    text = text.replace('\\n', '\n')
+    
+    # Remove any span and italic tags
+    text = re.sub(r'<span[^>]*>', '', text)
+    text = re.sub(r'</span>', '', text)
+    text = re.sub(r'<i[^>]*>', '', text)
+    text = re.sub(r'</i>', '', text)
+    
+    # Preserve and correctly format markdown links
+    text = re.sub(r'\[([^\]]+)\]\((https?://[^\)]+)\)', r'\1: \2', text)
+    
+    # Split the text into paragraphs
+    paragraphs = text.split('\n\n')
+    
+    cleaned_paragraphs = []
+    for paragraph in paragraphs:
+        lines = paragraph.split('\n')
+        cleaned_lines = []
+        for line in lines:
+            # Preserve bold formatting for headings
+            if line.strip().startswith('**') and line.strip().endswith('**'):
+                cleaned_line = line.strip()
+            else:
+                # Remove asterisks, special characters, and fix merged text
+                cleaned_line = re.sub(r'\*|\−|\∗', '', line)
+                cleaned_line = re.sub(r'([a-z])([A-Z])', r'\1 \2', cleaned_line)
+            
+            # Handle bullet points
+            if cleaned_line.strip().startswith('-'):
+                cleaned_line = '\n' + cleaned_line.strip()
+            
+            # Remove extra spaces
+            cleaned_line = re.sub(r'\s+', ' ', cleaned_line).strip()
+            cleaned_lines.append(cleaned_line)
+        
+        # Join the lines within each paragraph
+        cleaned_paragraph = '\n'.join(cleaned_lines)
+        cleaned_paragraphs.append(cleaned_paragraph)
+    
+    # Join the paragraphs back together
+    cleaned_text = '\n\n'.join(para for para in cleaned_paragraphs if para)
+    
+    return cleaned_text
 
 def side():        
     with st.sidebar.form(key='feedback_form'):
@@ -536,16 +580,15 @@ if prompt :
             })
             full_response = output["output"]
 
-            text = re.sub(r'\s+', ' ', full_response)  # Replace multiple spaces or newlines with a single space
-            text = re.sub(r'([a-z])(\d)', r'\1 \2', full_response)
-            text = re.sub(r'</span>', '', full_response)      # Remove closing </span> tags
+            cleaned_text = clean_text(full_response)
+
             # Display the response
-            if "**Trust Tip**" not in full_response and "**Suggestion**" not in full_response:
+            if "**Trust Tip**" not in full_response and "**Suggestion**" not in cleaned_text:
                 trust_tip = random.choice(trust_tips)
                 suggestion = random.choice(suggestions)
-                text += f"\n\n**Trust Tip**: {trust_tip}  \n\n **Suggestion**: {suggestion}"
+                cleaned_text += f"\n\n**Trust Tip**: {trust_tip}  \n\n **Suggestion**: {suggestion}"
             
-            st.markdown(text, unsafe_allow_html=True)
+            st.write(cleaned_text, unsafe_allow_html=True)