Update rag.py

rag.py

@@ -27,10 +27,10 @@ GREETINGS = [
     "hey there", "greetings"
 ]
 
-# Normalize user input for internal processing
+# Normalize user input for internal processing (with 'which' to 'what' replacement)
 def normalize_input(text):
     text = text.lower().strip()
-    text = text.replace("which", "what")
+    text = text.replace("which", "what")  # Add your requested replacement
     return text
 
 # Load local dataset
@@ -82,12 +82,17 @@ def query_groq_llm(prompt, model_name="llama3-70b-8192"):
         print(f"Error querying Groq API: {e}")
         return ""
 
-# Main logic function to be called from Gradio or elsewhere
+# Main logic function (with hidden 'which' to 'what' replacement)
 def get_best_answer(user_input):
     if not user_input.strip():
         return "Please enter a valid question."
 
-    user_input_lower = user_input.lower().strip()
+    # Preserve original input for display
+    original_input = user_input
+    
+    # Normalize input for processing (with hidden replacement)
+    processed_input = normalize_input(user_input)
+    user_input_lower = processed_input  # Use normalized version for processing
 
     if len(user_input_lower.split()) < 3 and not any(greet in user_input_lower for greet in GREETINGS):
         return "Please ask your question properly with at least 3 words."
@@ -95,7 +100,7 @@ def get_best_answer(user_input):
     if any(greet in user_input_lower for greet in GREETINGS):
         greeting_response = query_groq_llm(
             f"You are an official assistant for University of Education Lahore. "
-            f"Respond to this greeting in a friendly and professional manner: {user_input}"
+            f"Respond to this greeting in a friendly and professional manner: {original_input}"
         )
         return greeting_response if greeting_response else "Hello! How can I assist you today?"
 
@@ -106,15 +111,14 @@ def get_best_answer(user_input):
             "🔗 https://ue.edu.pk/allfeestructure.php"
         )
 
-    # Normalize input for similarity
-    normalized_input = normalize_input(user_input_lower)
-    user_embedding = similarity_model.encode(normalized_input, convert_to_tensor=True)
+    # Use normalized input for similarity matching
+    user_embedding = similarity_model.encode(user_input_lower, convert_to_tensor=True)
     similarities = util.pytorch_cos_sim(user_embedding, dataset_embeddings)[0]
     best_match_idx = similarities.argmax().item()
     best_score = similarities[best_match_idx].item()
 
     if best_score < 0.65:
-        manage_unmatched_queries(user_input)
+        manage_unmatched_queries(original_input)  # Store original query
 
     if best_score >= 0.65:
         original_answer = dataset_answers[best_match_idx]
@@ -123,7 +127,7 @@ Rephrase the following official answer clearly and professionally.
 Use structured formatting (like headings, bullet points, or numbered lists) where appropriate.
 DO NOT add any new or extra information. ONLY rephrase and improve the clarity and formatting of the original answer.
 ### Question:
-{user_input}
+{original_input}  # Show original to user
 ### Original Answer:
 {original_answer}
 ### Rephrased Answer:
@@ -133,7 +137,7 @@ DO NOT add any new or extra information. ONLY rephrase and improve the clarity a
 Include relevant details about university policies.
 If unsure, direct to official channels.
 ### Question:
-{user_input}
+{original_input}  # Show original to user
 ### Official Answer:
 """
 
@@ -150,4 +154,4 @@ If unsure, direct to official channels.
             "📞 +92-42-99262231-33\n"
             "✉️ info@ue.edu.pk\n"
             "🌐 https://ue.edu.pk"
-        )
+        )