Create a NEAT AI model using Python.

I need two scripts: one for NEAT that trains and provides the best model, and another script that connects with MT5 and utilizes this best model. Please make it simple, with three input parameters: close price, moving average of period 9, and moving average of period 14. The output should be a loop for buy-sell actions. The fitness logic should be based on profit and loss (PnL) or profit-loss per trade. The code should be flexible and compatible with any time frame I provide. The data will be in MT5 CSV format with headers, sourced from Tickstory. This is just a basic model; I plan to enhance it later with support and resistance levels, trendlines, Fibonacci levels, etc.