app.py

import gradio as gr
import os
from upstash_vector import Index
from openai import OpenAI
import dotenv

# Load environment variables
dotenv.load_dotenv()

# Services (keep the same as in your original code)
class ConfigService:
    @staticmethod
    def load_config():
        return {
            "OPENAI_API_KEY": os.getenv("OPENAI_API_KEY"),
            "UPSTASH_VECTOR_REST_URL": os.getenv("UPSTASH_VECTOR_REST_URL"),
            "UPSTASH_VECTOR_REST_TOKEN": os.getenv("UPSTASH_VECTOR_REST_TOKEN"),
        }

class ClientService:
    @staticmethod
    def initialize_clients(config):
        return {
            "index": Index(url=config["UPSTASH_VECTOR_REST_URL"], token=config["UPSTASH_VECTOR_REST_TOKEN"]),
            "openai": OpenAI(api_key=config["OPENAI_API_KEY"]),
        }

class EmbeddingService:
    @staticmethod
    def get_embedding(content, client):
        response = client.embeddings.create(
            input=content,
            model="text-embedding-3-large"
        )
        return response.data[0].embedding

class VectorSearchService:
    @staticmethod
    def search(index, vector, top_k=2):
        return index.query(vector=vector, top_k=top_k, include_metadata=True)

class PromptService:
    @staticmethod
    def load_system_prompt(file_path):
        with open(file_path, 'r') as file:
            return file.read().strip()

    @staticmethod
    def create_chat_prompt(question, context):
        return f"Question: {question}\n\nContext: {context}"

class ChatService:
    def __init__(self, clients, system_prompt):
        self.clients = clients
        self.messages = [{"role": "system", "content": system_prompt}]

    def ask_question(self, question):
        question_embedding = EmbeddingService.get_embedding(question, self.clients["openai"])
        search_results = VectorSearchService.search(self.clients["index"], question_embedding)
        
        context = []
        for r in search_results:
            if r.score > 0.7:
                context.append(r.metadata['content'])
        
        context = "\n".join(context)
        final_prompt = PromptService.create_chat_prompt(question, context)
        self.messages.append({"role": "user", "content": final_prompt})
        
        chat_completion = self.clients["openai"].chat.completions.create(
            model="gpt-4o-mini",
            messages=self.messages
        )
        response_text = chat_completion.choices[0].message.content
        self.messages.append({"role": "assistant", "content": response_text})
        return response_text

# Initialize services
config = ConfigService.load_config()
clients = ClientService.initialize_clients(config)
system_prompt = PromptService.load_system_prompt("prompts/sys.md")
chat_service = ChatService(clients, system_prompt)

# Gradio interface
def chatbot(message, history):
    response = chat_service.ask_question(message)
    return response

demo = gr.ChatInterface(
    fn=chatbot,
    title="Vector Database Cloud Chatbot",
    description="Ask questions about Vector Database Cloud",
    theme="default",
    examples=[
        "What is Vector Database Cloud?",
        "How does one-click deployment work?",
        "What vector databases are supported?",
    ],
)

if __name__ == "__main__":
    demo.launch()