Update app.py

app.py

@@ -1,102 +0,0 @@
-import json
-import os
-import gradio as gr
-from huggingface_hub import InferenceClient
-
-# Function to call the Llama 3.1 8B model through Hugging Face API
-def call_llama_model(user_query):
-    # Initialize the inference client with access token
-    client = InferenceClient(
-        "meta-llama/Meta-Llama-3.1-8B-Instruct",
-        token=os.environ.get("HF_TOKEN")
-    )
-    
-    # Define the addition function schema
-    function_schema = {
-        "name": "add_numbers",
-        "description": "Add two numbers together",
-        "parameters": {
-            "type": "object",
-            "properties": {
-                "num1": {
-                    "type": "number",
-                    "description": "First number to add"
-                },
-                "num2": {
-                    "type": "number",
-                    "description": "Second number to add"
-                }
-            },
-            "required": ["num1", "num2"]
-        }
-    }
-    
-    # Create the system prompt with function definition
-    system_prompt = f"""You have access to the following function:
-{json.dumps(function_schema, indent=2)}
-
-When given a query about adding numbers, you must:
-1. Extract the two numbers from the query
-2. Call the add_numbers function with these numbers
-3. Respond ONLY with a valid JSON object in this exact format:
-{{"function_call": {{"name": "add_numbers", "parameters": {{"num1": [first number], "num2": [second number]}}}}}}
-
-DO NOT include any explanation, just the JSON.
-"""
-    
-    # Call the model with the appropriate format for Llama 3.1
-    response = client.text_generation(
-        prompt=f"<|system|>\n{system_prompt}\n<|user|>\n{user_query}\n<|assistant|>",
-        max_new_tokens=256,
-        temperature=0.1,
-        return_full_text=False
-    )
-    
-    return response
-
-# Function to parse the model response and calculate the result
-def process_addition(query):
-    try:
-        # Get model response
-        model_response = call_llama_model(query)
-        
-        # Create a debug output with the raw response
-        debug_info = f"Raw model response:\n{model_response}\n\n"
-        
-        # Try to parse the JSON response
-        try:
-            # Find the JSON part in the response (it might have additional text)
-            json_start = model_response.find('{')
-            json_end = model_response.rfind('}') + 1
-            
-            if json_start >= 0 and json_end > json_start:
-                json_str = model_response[json_start:json_end]
-                response_data = json.loads(json_str)
-                debug_info += f"Parsed JSON:\n{json.dumps(response_data, indent=2)}\n\n"
-            else:
-                return f"Error: No valid JSON found in response.\n\n{debug_info}"
-            
-            # Check if it has a function call
-            if "function_call" in response_data:
-                function_name = response_data["function_call"]["name"]
-                params = response_data["function_call"]["parameters"]
-                
-                if function_name == "add_numbers":
-                    # Extract the numbers
-                    num1 = float(params["num1"])
-                    num2 = float(params["num2"])
-                    result = num1 + num2
-                    
-                    # Return a formatted response
-                    return f"""
-### Input Processed Successfully
-
-**Numbers extracted:** {num1} and {num2}
-
-**Function called:** `{function_name}`
-
-**Result:** {result}
-
-**JSON Function Call:**
-```json
-{json.dumps(response_data, indent=2)}