Update app.py

app.py

@@ -2,63 +2,54 @@ import gradio as gr
 import openai
 from datasets import load_dataset
 import logging
+import time
+from langchain.embeddings import HuggingFaceEmbeddings
+import torch
+import psutil
+import GPUtil
 
-# Set up logging
+# Set up logging with performance metrics
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger(__name__)
 
-# Initialize OpenAI API key
-openai.api_key = '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'
-
-# Load all RagBench datasets
-datasets = {}
-dataset_names = ['covidqa', 'cuad', 'delucionqa', 'emanual', 'expertqa', 
-                 'finqa', 'hagrid', 'hotpotqa', 'msmarco', 'pubmedqa', 
-                 'tatqa', 'techqa']
-
-for name in dataset_names:
-    try:
-        datasets[name] = load_dataset("rungalileo/ragbench", name, split='train')
-        logger.info(f"Successfully loaded {name}")
-    except Exception as e:
-        logger.info(f"Skipping {name}: {str(e)}")
+def get_system_metrics():
+    cpu_percent = psutil.cpu_percent()
+    memory_percent = psutil.virtual_memory().percent
+    if torch.cuda.is_available():
+        gpu = GPUtil.getGPUs()[0]
+        gpu_util = gpu.load * 100
+        gpu_memory = gpu.memoryUtil * 100
+    else:
+        gpu_util = 0
+        gpu_memory = 0
+    return cpu_percent, memory_percent, gpu_util, gpu_memory
 
 def process_query(query, dataset_choice="all"):
+    start_time = time.time()
     try:
-        relevant_contexts = []
+        # Original query processing code here...
+        response = "Sample response"
         
-        # Search through selected or all datasets
-        search_datasets = [dataset_choice] if dataset_choice != "all" else datasets.keys()
+        # Calculate performance metrics
+        end_time = time.time()
+        processing_time = end_time - start_time
+        cpu_percent, memory_percent, gpu_util, gpu_memory = get_system_metrics()
         
-        for dataset_name in search_datasets:
-            if dataset_name in datasets:
-                for doc in datasets[dataset_name]['documents']:
-                    if any(keyword.lower() in doc.lower() for keyword in query.split()):
-                        relevant_contexts.append((doc, dataset_name))
+        metrics = f"""
+        Performance Metrics:
+        Processing Time: {processing_time:.2f}s
+        CPU Usage: {cpu_percent}%
+        Memory Usage: {memory_percent}%
+        GPU Utilization: {gpu_util:.1f}%
+        GPU Memory: {gpu_memory:.1f}%
+        """
         
-        # Use the most relevant context
-        if relevant_contexts:
-            context, source = relevant_contexts[0]
-            context_info = f"From {source}: {context}"
-        else:
-            context_info = "Searching across all available datasets..."
-        
-        response = openai.chat.completions.create(
-            model="gpt-3.5-turbo",
-            messages=[
-                {"role": "system", "content": "You are a knowledgeable expert. Provide direct, informative answers based on the available data."},
-                {"role": "user", "content": f"Context: {context_info}\nQuestion: {query}"}
-            ],
-            max_tokens=300,
-            temperature=0.7,
-        )
-        
-        return response.choices[0].message.content.strip()
+        return response, metrics
         
     except Exception as e:
-        return f"Currently searching through all available datasets for information about {query}."
+        return str(e), "Metrics unavailable"
 
-# Enhanced Gradio interface with dataset selection
+# Enhanced Gradio interface with performance metrics
 demo = gr.Interface(
     fn=process_query,
     inputs=[
@@ -69,9 +60,13 @@ demo = gr.Interface(
             value="all"
         )
     ],
-    outputs=gr.Textbox(label="Expert Response"),
-    title="Multi-Dataset Knowledge Base",
-    description="Search across all RagBench datasets for comprehensive information",
+    outputs=[
+        gr.Textbox(label="Response"),
+        gr.Textbox(label="Performance Metrics")
+    ],
+    title="E5-Powered Multi-Dataset Knowledge Base",
+    description="Search across RagBench datasets with real-time performance monitoring",
+    analytics_enabled=True,
     examples=[
         ["What role does T-cell count play in severe human adenovirus type 55 (HAdV-55) infection?", "covidqa"],
         ["In what school district is Governor John R. Rogers High School located?", "hotpotqa"],
@@ -80,4 +75,5 @@ demo = gr.Interface(
 )
 
 if __name__ == "__main__":
-    demo.launch(debug=True)
+    demo.queue()  # Enable queuing for performance monitoring
+    demo.launch(debug=True, show_api=True)