Update app.py

app.py

@@ -1,39 +1,61 @@
-""" Enhanced Multi-LLM Agent Evaluation Runner with Vector Database Integration"""
+""" Enhanced Multi-LLM Agent Evaluation Runner with Agno Integration"""
 import os
 import gradio as gr
 import requests
 import pandas as pd
 from langchain_core.messages import HumanMessage
-from veryfinal import build_graph, HybridLangGraphMultiLLMSystem
+
+# Import the enhanced classes from veryfinal.py in the same directory
+try:
+    from veryfinal import (
+        build_graph, 
+        UnifiedAgnoEnhancedSystem,
+        AgnoEnhancedAgentSystem,
+        AgnoEnhancedModelManager
+    )
+    VERYFINAL_AVAILABLE = True
+except ImportError as e:
+    print(f"Error importing from veryfinal.py: {e}")
+    VERYFINAL_AVAILABLE = False
 
 # --- Constants ---
 DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
 # --- Enhanced Agent Definition ---
 class EnhancedMultiLLMAgent:
-    """A multi-provider LangGraph agent with vector database integration."""
+    """A multi-provider Agno agent with NVIDIA + open-source model integration."""
     def __init__(self):
-        print("Enhanced Multi-LLM Agent with Vector Database initialized.")
+        print("Enhanced Multi-LLM Agent with Agno Integration initialized.")
+        
+        if not VERYFINAL_AVAILABLE:
+            print("Error: veryfinal.py not properly imported")
+            self.system = None
+            self.graph = None
+            return
+        
         try:
-            self.system = HybridLangGraphMultiLLMSystem(provider="groq")
+            # Use the unified Agno enhanced system
+            self.system = UnifiedAgnoEnhancedSystem()
             self.graph = self.system.graph
             
-            # Load metadata if available
-            if os.path.exists("metadata.jsonl"):
-                print("Loading question metadata...")
-                count = self.system.load_metadata_from_jsonl("metadata.jsonl")
-                print(f"Loaded {count} questions into vector database")
+            # Display system information
+            if self.system.agno_system:
+                info = self.system.get_system_info()
+                print(f"System initialized with {info.get('total_models', 0)} models")
+                if info.get('nvidia_available'):
+                    print("✅ NVIDIA NIM models available")
+                print(f"Active agents: {info.get('active_agents', [])}")
             
-            print("Enhanced Multi-LLM Graph built successfully.")
+            print("Enhanced Agno Multi-LLM System built successfully.")
         except Exception as e:
-            print(f"Error building graph: {e}")
+            print(f"Error building enhanced system: {e}")
             self.graph = None
             self.system = None
 
     def __call__(self, question: str) -> str:
         print(f"Agent received question: {question[:100]}...")
         
-        if self.graph is None or self.system is None:
+        if self.system is None:
             return "Error: Agent not properly initialized"
         
         try:
@@ -43,8 +65,15 @@ class EnhancedMultiLLMAgent:
             # Additional validation
             if not answer or answer == question or len(answer.strip()) == 0:
                 return "Information not available"
-                
-            return answer.strip()
+            
+            # Clean up the answer
+            answer = answer.strip()
+            
+            # Ensure proper formatting for evaluation
+            if "FINAL ANSWER:" in answer:
+                answer = answer.split("FINAL ANSWER:")[-1].strip()
+            
+            return answer
                 
         except Exception as e:
             error_msg = f"Error: {str(e)}"
@@ -52,7 +81,7 @@ class EnhancedMultiLLMAgent:
             return error_msg
 
 def run_and_submit_all(profile: gr.OAuthProfile | None):
-    """Fetch questions, run enhanced agent, and submit answers."""
+    """Fetch questions, run enhanced Agno agent, and submit answers."""
     space_id = os.getenv("SPACE_ID")
     
     if profile:
@@ -66,11 +95,11 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     questions_url = f"{api_url}/questions"
     submit_url = f"{api_url}/submit"
 
-    # 1. Instantiate Agent
+    # 1. Instantiate Enhanced Agent
     try:
         agent = EnhancedMultiLLMAgent()
-        if agent.graph is None:
-            return "Error: Failed to initialize agent properly", None
+        if agent.system is None:
+            return "Error: Failed to initialize enhanced agent properly", None
     except Exception as e:
         print(f"Error instantiating agent: {e}")
         return f"Error initializing agent: {e}", None
@@ -92,10 +121,10 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         print(f"Error fetching questions: {e}")
         return f"Error fetching questions: {e}", None
 
-    # 3. Run Enhanced Agent
+    # 3. Run Enhanced Agno Agent
     results_log = []
     answers_payload = []
-    print(f"Running Enhanced Multi-LLM agent with vector database on {len(questions_data)} questions...")
+    print(f"Running Enhanced Agno Multi-LLM agent on {len(questions_data)} questions...")
     
     for i, item in enumerate(questions_data):
         task_id = item.get("task_id")
@@ -136,7 +165,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
 
     # 4. Prepare Submission 
     submission_data = {"username": username.strip(), "agent_code": agent_code, "answers": answers_payload}
-    status_update = f"Enhanced Multi-LLM Agent finished. Submitting {len(answers_payload)} answers for user '{username}'..."
+    status_update = f"Enhanced Agno Multi-LLM Agent finished. Submitting {len(answers_payload)} answers for user '{username}'..."
     print(status_update)
 
     # 5. Submit
@@ -163,7 +192,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
 
 # --- Build Gradio Interface ---
 with gr.Blocks() as demo:
-    gr.Markdown("# Enhanced Multi-LLM Agent with Vector Database Integration")
+    gr.Markdown("# Enhanced Multi-LLM Agent with Agno + NVIDIA Integration")
     gr.Markdown(
         """
         **Instructions:**
@@ -171,25 +200,32 @@ with gr.Blocks() as demo:
         2. Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.
         
         **Enhanced Agent Features:**
-        - **Multi-LLM Support**: Groq (Llama-3 8B/70B, DeepSeek)
-        - **Vector Database Integration**: FAISS + Supabase for similar question retrieval
-        - **Intelligent Routing**: Automatically selects best provider based on query complexity
-        - **Enhanced Tools**: Mathematical operations, web search, Wikipedia integration
-        - **Question-Answering**: Optimized for evaluation tasks with proper formatting
-        - **Similar Questions Context**: Uses vector similarity to provide relevant context
-        - **Error Handling**: Robust fallback mechanisms and comprehensive logging
+        - **NVIDIA NIM Models**: Enterprise-grade optimized models for maximum accuracy
+        - **Open-Source Models**: Groq, Ollama, Together AI, Anyscale, Hugging Face
+        - **Specialized Agents**: Enterprise research, advanced math, coding, fast response
+        - **Intelligent Routing**: Automatically selects best model/agent for each task
+        - **Advanced Tools**: DuckDuckGo search, Wikipedia, calculator, reasoning tools
+        - **Agno Framework**: Professional agent framework with memory and tool integration
+        
+        **Available Model Providers:**
+        - **NVIDIA NIM**: meta/llama3-70b-instruct, meta/codellama-70b-instruct, etc.
+        - **Groq (Free)**: llama3-70b-8192, llama3-8b-8192, mixtral-8x7b-32768
+        - **Ollama (Local)**: llama3, mistral, phi3, codellama, gemma, qwen
+        - **Together AI**: Meta-Llama models, Mistral, Qwen
+        - **Anyscale**: Enterprise hosting for open-source models
+        - **Hugging Face**: Direct model access
         
         **Routing Examples:**
-        - Math: "What is 25 multiplied by 17?" → Llama-3 70B
-        - Search: "Find information about Mercedes Sosa" → Search-Enhanced
-        - Complex: "Analyze quantum computing principles" → DeepSeek
-        - Simple: "What is the capital of France?" → Llama-3 8B
+        - Enterprise: "Enterprise analysis of quantum computing" → NVIDIA NIM
+        - Math: "Calculate 25 × 17" → Advanced Math Agent
+        - Code: "Write Python factorial function" → Advanced Coding Agent
+        - Research: "Find Mercedes Sosa discography" → Enterprise Research Agent
+        - Quick: "Capital of France?" → Fast Response Agent
         
-        **Vector Database Features:**
-        - Automatic loading of metadata.jsonl if present
-        - Similar question retrieval for enhanced context
-        - Supabase integration for persistent storage
-        - FAISS for fast vector similarity search
+        **Setup Requirements:**
+        - NVIDIA_API_KEY for enterprise models (optional)
+        - GROQ_API_KEY for free tier models
+        - Other API keys optional for additional providers
         """
     )
 
@@ -206,5 +242,20 @@ with gr.Blocks() as demo:
     )
 
 if __name__ == "__main__":
-    print("\n" + "-"*30 + " Enhanced Multi-LLM Agent with Vector DB Starting " + "-"*30)
+    print("\n" + "-"*30 + " Enhanced Agno Multi-LLM Agent Starting " + "-"*30)
+    
+    # Display system status
+    if VERYFINAL_AVAILABLE:
+        try:
+            test_system = UnifiedAgnoEnhancedSystem()
+            info = test_system.get_system_info()
+            print(f"✅ System ready with {info.get('total_models', 0)} models")
+            print(f"📊 Model breakdown: {len(info.get('model_breakdown', {}).get('nvidia_models', []))} NVIDIA, "
+                  f"{len(info.get('model_breakdown', {}).get('groq_models', []))} Groq, "
+                  f"{len(info.get('model_breakdown', {}).get('ollama_models', []))} Ollama")
+        except Exception as e:
+            print(f"⚠️  System initialization warning: {e}")
+    else:
+        print("❌ veryfinal.py not properly imported")
+    
     demo.launch(debug=True, share=False)