h2oGPTe Banking RAG Demo - Intelligent Document Processing for Financial Services This demo showcases how to leverage h2oGPTe's powerful RAG capabilities for banking and financial services use cases. It demonstrates: - Automated knowledge base creation from banking documentation - Intelligent Q&A using RAG - Best practices for financial documen…

h2ogpte_banking_rag_demo.py

"""
h2oGPTe Banking RAG Demo
Demonstrates using h2oGPTe for intelligent document processing in financial services

This demo shows how to:
1. Crawl relevant banking documentation
2. Create a knowledge base
3. Perform intelligent Q&A using RAG
"""

import h2ogpte
from h2ogpte import H2OGPTE
import requests
from typing import List
import logging

# Setup logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

class BankingKnowledgeBase:
    def __init__(self, api_key: str):
        """Initialize the banking knowledge base with h2oGPTe"""
        self.client = H2OGPTE(api_key=api_key)
        self.sources = []
        
    def add_wikipedia_sources(self) -> None:
        """Add relevant banking Wikipedia pages to the knowledge base"""
        banking_pages = [
            "Federal_Reserve",
            "Investment_banking",
            "Retail_banking",
            "Financial_technology"
        ]
        
        for page in banking_pages:
            url = f"https://en.wikipedia.org/wiki/{page}"
            try:
                # Add document to h2oGPTe knowledge base
                self.client.add_document(
                    url,
                    metadata={"source": "wikipedia", "topic": page}
                )
                self.sources.append(url)
                logger.info(f"Added {url} to knowledge base")
            except Exception as e:
                logger.error(f"Error adding {url}: {e}")
                
    def ask_question(self, question: str) -> str:
        """Query the knowledge base using RAG"""
        try:
            response = self.client.query(
                question,
                search_options={"sources": self.sources}
            )
            return response.text
        except Exception as e:
            logger.error(f"Error querying knowledge base: {e}")
            return None

def main():
    """Main demo function"""
    # Initialize with your API key
    api_key = "YOUR_H2OGPTE_API_KEY"
    kb = BankingKnowledgeBase(api_key)
    
    # Build knowledge base
    logger.info("Building banking knowledge base...")
    kb.add_wikipedia_sources()
    
    # Example questions
    questions = [
        "What are the key differences between retail and investment banking?",
        "How is fintech impacting traditional banking services?",
        "What role does the Federal Reserve play in the US banking system?"
    ]
    
    # Demonstrate RAG capabilities
    logger.info("Demonstrating RAG capabilities...")
    for question in questions:
        logger.info(f"\nQuestion: {question}")
        answer = kb.ask_question(question)
        logger.info(f"Answer: {answer}")

if __name__ == "__main__":
    main()