Update app.py

app.py

@@ -202,96 +202,38 @@ def query_crm_data_with_context(prompt, top_k=3):
         return ["Error in querying recommendations."]
 
 
-def load_objection_responses(file_path):
-    df = pd.read_csv(file_path)
-    objection_response_dict = dict(zip(df["Objection"], df["Response"]))
-    return objection_response_dict
+sentence_model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2')
+faiss_index = faiss.IndexFlatL2(384)
+
+def load_objection_responses(csv_file_path):
+    try:
+        df = pd.read_csv(csv_file_path)
+        objection_response_pairs = dict(zip(df['Objection'], df['Response']))
+        return objection_response_pairs
+    except Exception as e:
+        print(f"Error loading objections CSV: {e}")
+        return {}
 
-# Load objections from CSV file
 objection_response_pairs = load_objection_responses(r"C:\Users\bhagy\OneDrive\Desktop\INFOSYS PROJECT\objections_responses.csv")
 objections = list(objection_response_pairs.keys())
-
-# Load SentenceTransformer model
-sentence_model = SentenceTransformer("all-MiniLM-L6-v2")  # Change model if needed
-
-# Convert objections into embeddings
 objection_embeddings = sentence_model.encode(objections)
+faiss_index.add(np.array(objection_embeddings, dtype="float32"))
 
-# Ensure embeddings are 2D
-objection_embeddings = np.array(objection_embeddings, dtype="float32")
-if len(objection_embeddings.shape) == 1:
-    objection_embeddings = objection_embeddings.reshape(-1, 1)
-
-# Debug: Print shape of embeddings
-print("Objection Embeddings Shape:", objection_embeddings.shape)
-
-# Initialize FAISS index
-d = objection_embeddings.shape[1]  # Get embedding dimension
-faiss_index = faiss.IndexFlatL2(d)  # L2 similarity index
-
-# Add embeddings to FAISS index
-faiss_index.add(objection_embeddings)
-
-# Function to find closest objection
 def find_closest_objection(query):
     query_embedding = sentence_model.encode([query])
-
-    # Ensure query_embedding is 2D
-    query_embedding = np.array(query_embedding, dtype="float32")
-    if len(query_embedding.shape) == 1:
-        query_embedding = query_embedding.reshape(-1, 1)
-
-    distances, indices = faiss_index.search(query_embedding, 1)
-
+    distances, indices = faiss_index.search(np.array(query_embedding, dtype="float32"), 1)
     closest_index = indices[0][0]
     closest_objection = objections[closest_index]
     response = objection_response_pairs[closest_objection]
-
     if distances[0][0] > 0.6:
         return "No objection found", "No Response"
-
     return closest_objection, response
 
-# Function to handle objections and recommend responses
 def handle_objection_and_recommendation(prompt):
     closest_objection, objection_response = find_closest_objection(prompt)
-    
-    # Placeholder function: Replace with actual CRM data retrieval
     recommendations = query_crm_data_with_context(prompt)
 
     return closest_objection, objection_response, recommendations
-# sentence_model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2')
-# faiss_index = faiss.IndexFlatL2(384)
-
-# def load_objection_responses(csv_file_path):
-#     try:
-#         df = pd.read_csv(csv_file_path)
-#         objection_response_pairs = dict(zip(df['Objection'], df['Response']))
-#         return objection_response_pairs
-#     except Exception as e:
-#         print(f"Error loading objections CSV: {e}")
-#         return {}
-
-# objection_response_pairs = load_objection_responses(r"C:\Users\bhagy\OneDrive\Desktop\INFOSYS PROJECT\objections_responses.csv")
-# objections = list(objection_response_pairs.keys())
-# objection_embeddings = sentence_model.encode(objections)
-# faiss_index.add(np.array(objection_embeddings, dtype="float32"))
-
-# def find_closest_objection(query):
-#     query_embedding = sentence_model.encode([query])
-#     distances, indices = faiss_index.search(np.array(query_embedding, dtype="float32"), 1)
-#     closest_index = indices[0][0]
-#     closest_objection = objections[closest_index]
-#     response = objection_response_pairs[closest_objection]
-#     if distances[0][0] > 0.6:
-#         return "No objection found", "No Response"
-#     return closest_objection, response
-
-# def handle_objection_and_recommendation(prompt):
-#     closest_objection, objection_response = find_closest_objection(prompt)
-#     recommendations = query_crm_data_with_context(prompt)
-
-#     return closest_objection, objection_response, recommendations
 
 
 if "is_listening" not in st.session_state: