Fix Backlink Processing

classes/Seo_Off_Page.py

@@ -1,3 +1,5 @@
+from io import StringIO
+from urllib.parse import urlparse
 import streamlit as st
 import requests
 from dotenv import load_dotenv
@@ -102,8 +104,19 @@ class SeoOffPageAnalyst:
                         for file_info in st.session_state['uploaded_files'].values():
                             if file_info['type'] == 'pdf':
                                 combined_text += file_info['content'] + "\n"
-                            elif file_info['type'] == 'csv':
-                                combined_text += file_info['content'].to_csv(index=True) + "\n"
+                            elif file_info['type'] == 'csv':                                    
+                                # Load CSV
+                                df = pd.read_csv(StringIO(file_info['content'].to_csv(index=True)))
+                                
+                                # Count total rows
+                                num_rows = len(df)
+                                
+                                # Extract unique domains from 'Source url'
+                                df['Source Domain'] = df['Source url'].apply(lambda x: urlparse(x).netloc)
+                                unique_domains = df['Source Domain'].nunique()
+                                    
+                                combined_text += f"Total rows: {num_rows}\n"
+                                combined_text += f"Referring Domain: {unique_domains}"
                         
                         # OUTPUT FOR SEO ANALYST
                         payload_txt = {"question": combined_text}