Is there a problem with the dataset build?

[LonelyFireFox]

Hello！
I think there is a problem when you build sample in the dataloader.py.
The price and text data are read in reverse order. However, when constructing the time window data, the loop code block in the figure below will use the labels and price features of the predicted target day.
At this time, the first index should be the real label, why use the last one?
Is there a mistake here? Or I misunderstood, please correct me.