from langchain_openai import ChatOpenAI from langchain.schema import AIMessage, HumanMessage import openai import gradio as gr import datetime import json from pathlib import Path llm = ChatOpenAI(model_name='gpt-4-turbo-preview') counter_file_path = Path("process_counter.json") daily_API_limit=50 def read_counter_file(): if counter_file_path.exists(): with open(counter_file_path, "r") as file: return json.load(file) else: return {"count": 0, "last_run_date": str(datetime.date.today())} def write_counter_file(data): with open(counter_file_path, "w") as file: json.dump(data, file) def reset_counter_if_needed(data): last_run_date = datetime.datetime.strptime(data["last_run_date"], "%Y-%m-%d").date() today = datetime.date.today() if last_run_date < today: data["count"] = 0 data["last_run_date"] = str(today) return data def check_daily_limit(data, daily_limit=daily_API_limit): if data["count"] >= daily_limit: return True # Daily limit reached else: return False # Daily limit not reached def check_daily_limit_reached(): data = read_counter_file() data = reset_counter_if_needed(data) daily_limit_reached = check_daily_limit(data) if not daily_limit_reached: data["count"] += 1 # Increment the process run count daily_limit_reached = check_daily_limit(data) # Re-check after incrementing write_counter_file(data) # Save the updated count and date return daily_limit_reached def predict(message, history): history_langchain_format = [] for human, ai in history: history_langchain_format.append(HumanMessage(content=human)) history_langchain_format.append(AIMessage(content=ai)) history_langchain_format.append(HumanMessage(content=message)) daily_limit_reached=check_daily_limit_reached() if daily_limit_reached: return("Daily limit reached. Please try again tomorrow.") else: gpt_response = llm(history_langchain_format) return gpt_response.content gr.ChatInterface(predict).launch()