Update app.py

app.py

@@ -5,10 +5,11 @@ import faiss
 from sentence_transformers import SentenceTransformer
 from huggingface_hub import InferenceClient
 
-# --- Load data ---
+# --- Load and clean data ---
 df = pd.read_csv("tariff_codes.csv", encoding="latin1", low_memory=False)
-descriptions = df["Description"].astype(str).tolist()
-codes = df["Code"].astype(str).tolist()
+df.columns = df.columns.str.strip()
+descriptions = df["brief_description"].astype(str).tolist()
+codes = df["hts8"].astype(str).tolist()
 
 # --- Create embeddings ---
 embedding_model = SentenceTransformer("all-MiniLM-L6-v2")
@@ -30,7 +31,11 @@ def generate_answer(user_query):
     _, indices = index.search(query_embedding, k=5)
 
     context = "\n".join([f"{codes[i]}: {descriptions[i]}" for i in indices[0]])
-    prompt = f"""Here are some tariff code descriptions:\n{context}\n\nQuestion: {user_query}\nAnswer:"""
+    prompt = f"""Here are some tariff code descriptions:
+{context}
+
+Question: {user_query}
+Answer:"""
 
     response = client.text_generation(
         prompt,
@@ -40,7 +45,7 @@ def generate_answer(user_query):
     )
     return response.strip()
 
-# --- Gradio UI ---
+# --- Gradio Chat Interface ---
 gr.ChatInterface(
     fn=generate_answer,
     title="Tariff Code RAG Bot (FAISS + Inference API)"