Update app.py

app.py

@@ -1,60 +1,34 @@
 import os
 import pandas as pd
 import gradio as gr
-from huggingface_hub import InferenceClient
 from transformers import pipeline
 
-# Install necessary dependencies
-os.system('pip install transformers gradio requests pandas')
-
-# Inference client for chat completion using GPT-4
-client = InferenceClient("openai/gpt-4")
-
-# Different pipelines for different tasks
+# Initialize the question-answering pipeline with the 'deepset/roberta-base-squad2' model
 qa_pipeline = pipeline("question-answering", model="deepset/roberta-base-squad2")
 
-def respond(message, system_message, max_tokens, temperature, top_p):
-    messages = [
-        {"role": "system", "content": system_message},
-        {"role": "user", "content": message}
-    ]
-    
-    response = ""
-    try:
-        for message in client.chat_completion(
-            messages=messages,
-            max_tokens=max_tokens,
-            stream=True,
-            temperature=temperature,
-            top_p=top_p
-        ):
-            token = message.choices[0].delta.get("content", "")
-            if token:
-                response += token
-    except Exception as e:
-        return f"Error during chat completion: {str(e)}"
-    
-    return response
-
 # GDPR Compliance Expert
 def evaluate_gdpr_compliance(audit_data):
-    system_message = (
-        "You are an expert GDPR compliance officer. Assess the audit data for compliance with GDPR regulations. "
-        "Provide an analysis that identifies any compliance issues and suggestions for remediation. "
-        "Ensure a thorough evaluation of data processing, storage, and protection practices in line with GDPR requirements."
+    # Example question for GDPR compliance
+    question = (
+        "Based on the provided audit data, are there any compliance issues related to the GDPR regulations? "
+        "Evaluate the data processing, storage, and protection practices."
     )
-    compliance_analysis = respond(audit_data, system_message, max_tokens=1024, temperature=0.7, top_p=0.95)
-    return compliance_analysis
+    
+    # Apply the question-answering pipeline
+    response = qa_pipeline(question=question, context=audit_data)
+    return response['answer']
 
 # PCI Compliance Expert
 def evaluate_pci_compliance(audit_data):
-    system_message = (
-        "You are an expert PCI compliance officer. Assess the audit data for compliance with PCI DSS regulations. "
-        "Provide an analysis that identifies any compliance issues and suggestions for remediation. "
-        "Ensure a thorough evaluation of payment card data security, storage, and processing practices in line with PCI requirements."
+    # Example question for PCI DSS compliance
+    question = (
+        "Based on the provided audit data, are there any compliance issues related to PCI DSS regulations? "
+        "Evaluate the payment card data security, storage, and processing practices."
     )
-    compliance_analysis = respond(audit_data, system_message, max_tokens=1024, temperature=0.7, top_p=0.95)
-    return compliance_analysis
+    
+    # Apply the question-answering pipeline
+    response = qa_pipeline(question=question, context=audit_data)
+    return response['answer']
 
 # Analyze CSV file input
 def analyze_csv_file(file_obj):
@@ -66,7 +40,13 @@ def analyze_csv_file(file_obj):
     
     # Convert DataFrame to dictionary for processing
     audit_data = df.to_dict(orient='records')
-    return audit_data
+    
+    # Convert the dictionary to a string format suitable for the QA model
+    audit_data_str = ""
+    for record in audit_data:
+        audit_data_str += " ".join([f"{key}: {value}" for key, value in record.items()]) + "\n"
+    
+    return audit_data_str
 
 # Custom CSS for the specified theme
 custom_css = """