Upload app.py

app.py

@@ -20,40 +20,37 @@ pipeline = transformers.pipeline(
     use_auth_token=hf_token  # Use the Hugging Face token here
 )
 
-# Function to calculate scores and rankings
-def calculate_ranking(data):
-    for institution in data:
-        institution["Total"] = (
-            institution["TLR"] + institution["GO"] + institution["OI"] + institution["PR"]
-        )
-    ranked_data = sorted(data, key=lambda x: x["Total"], reverse=True)
-    for rank, institution in enumerate(ranked_data, start=1):
-        institution["Rank"] = rank
-    return ranked_data
-
-# Predefined ranking data
+# Predefined data
 example_data = [
     {"Institution": "A", "TLR": 70, "GO": 85, "OI": 90, "PR": 75},
     {"Institution": "B", "TLR": 80, "GO": 88, "OI": 85, "PR": 90},
     {"Institution": "C", "TLR": 65, "GO": 80, "OI": 70, "PR": 60},
 ]
 
+# Format predefined data into a readable string
+predefined_context = "Here are the institution rankings based on scores:\n"
+for institution in sorted(example_data, key=lambda x: x["TLR"] + x["GO"] + x["OI"] + x["PR"], reverse=True):
+    total_score = institution["TLR"] + institution["GO"] + institution["OI"] + institution["PR"]
+    predefined_context += f"- {institution['Institution']} (Total Score: {total_score})\n"
+
+# System prompt to provide context to the model
+system_prompt = f"""You are an intelligent assistant. Here is some contextual information:
+{predefined_context}
+
+When a user asks about rankings, respond with this information. If the user asks general questions, respond appropriately.
+"""
+
 # Chatbot function
 def chatbot_response(user_message):
-    # Check for predefined data queries
-    if "rank" in user_message.lower():
-        ranked_data = calculate_ranking(example_data)
-        response = "Here are the ranks of the institutions:\n"
-        for institution in ranked_data:
-            response += f"Rank {institution['Rank']}: {institution['Institution']} (Total Score: {institution['Total']})\n"
-        return response
+    # Combine system prompt with the user's message
+    full_prompt = f"{system_prompt}\nUser: {user_message}\nAssistant:"
     
-    # Fallback to model-generated response for out-of-scope questions
+    # Generate a response using the model
     outputs = pipeline(
-        user_message,
-        max_new_tokens=100,  # Restrict length for unexpected questions
+        full_prompt,
+        max_new_tokens=150,  # Adjust token limit as needed
         do_sample=True,
-        temperature=0.7,  # Slightly random responses for more natural output
+        temperature=0.7,
         top_p=0.9,
     )
     return outputs[0]["generated_text"]
@@ -61,8 +58,8 @@ def chatbot_response(user_message):
 # Gradio interface
 def build_gradio_ui():
     with gr.Blocks() as demo:
-        gr.Markdown("## Chatbot with Predefined Data and AI Responses")
-        gr.Markdown("Ask about institution rankings or any other general query!")
+        gr.Markdown("## Intelligent Chatbot with Predefined Context and AI Responses")
+        gr.Markdown("Ask about institution rankings or any general query!")
         with gr.Row():
             user_input = gr.Textbox(label="Your Message", placeholder="Type your message here...")
             chatbot_output = gr.Textbox(label="Chatbot Response", interactive=False)