import gradio as gr
import os
from langchain.agents import initialize_agent
from langchain.llms import OpenAI
from gradio_tools import BaseTool
from langchain.memory import ConversationBufferMemory

# Set up OpenAI API Key
if not os.getenv("OPENAI_API_KEY"):
    os.environ["OPENAI_API_KEY"] = "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"

# Simulated method to fetch a review from Google Maps (You'll replace this with an actual implementation)
def fetch_gmaps_review(location):
    # Simulated review
    review = "The location was outstanding with beautiful views and amazing food. Service could be better though."
    return review

class GMapsReviewSummarizationTool(BaseTool):

    def __init__(self) -> None:
        super().__init__(name="GMapsReviewSummarizer",
                         description="A tool that fetches a Google Maps review based on a location and then summarizes it.",
                         src="your_gradio_app_url_or_id_here")

    def create_job(self, query: str) -> str:
        review = fetch_gmaps_review(query)
        return review

    def postprocess(self, output: str) -> str:
        llm = OpenAI(temperature=0.5)
        response = llm.query(f"Summarize the following Google Maps review: '{output}'")
        return response

def main():
    llm = OpenAI(temperature=0.5)
    memory = ConversationBufferMemory(memory_key="review_history")
    tools = [GMapsReviewSummarizationTool().langchain]

    agent = initialize_agent(tools, llm, memory=memory, agent="conversational-review-summarizer", verbose=True)

    def gr_interface(location_input):
        return agent.run(input=f"Fetch and summarize a review for the location: {location_input}")

    interface = gr.Interface(fn=gr_interface, 
                             inputs="textbox", 
                             outputs="textbox",
                             live=True,
                             title="GMaps Review Summarizer")
    
    interface.launch()

if __name__ == "__main__":
    main()