Incorporating forecasted input values into a time window forecasting model

Hello, everyone!

I’ve followed Tensorflow’s tutorial on time series forecasting, and I am wondering if forecasted input values can be incorporated into training, based on tensorflow’s “window” approach to solving time-series forecasting problems.

For example:

The day is Jan 1 2022 and we are building a model to predict the price of electricity throughout 7 days, using comprehensive historical weather data as input.

Now, if we only use the last 7 days to base the prediction on, and want to predict 7 days out, that would give us a window length of 14, with an offset of 7. My question is: what if we want to incorporate weather forecasting data for the next 7 days?

I guess what I’m asking is if anyone knows of a way that forecasted weather data can be used alongside historical weather data in order to forecast, lets say electricity price, during the timeframe that we have forecasted weather values.

Does anyone have any thoughts or information?
Thank you!