How to import pytorch_geometric.data.batch.to_data_list properly?

[Kevin-shihello-world]

I want to write a training process like a prototypical network to analyze how the GNN learning process goes. I've tried another way to separate single data from batched data by using batch data, which is complex and slow. And I learned the PYG has to_data_list function here for datasets transferred from data list and as the source code says the Tu_datasets are allocated from a data list I wonder if it would satisfy my request.
I tried but found I can't import this function and even after I updated my PYG package(I used on a Linux GPU machine with 11.3 version CUDA and 1.11 pytorch)
I've used the following code:

python
from torch.geometric.Data.batch import to_data_list

I wonder where the problem lies. Does the to_data_list function can't be imported in this way and can be used directly?And whether can I use it on the Tu_dataset?
I would be appreciated it if someone can offer me an idea. Thank you!

[Kevin-shihello-world]

Ok, I've just found out I asked a silly question, I was hurrying to my college and the project was almost done so I asked this question without thoughtful thinking (cover my face.jpg). In fact I can use data[i] to do that job and to_data_list is a function in a class so it can't be imported.

[Kevin-shihello-world]

Ok, I've just found out I asked a silly question, I was hurrying to my college and the project was almost done so I asked this question without thoughtful thinking (cover my face.jpg). In fact I can use data[i] to do that job and to_data_list is a function in a class so it can't be imported.
Sorry for your time.