add function for generating TrajectoryContextDataset from a list of trajectories

[mitchellostrow]

I am working with multiple trajectories and want to create a TrajectoryContextDataset from a list of them, so I created a function to simply loop through all the trajectories and add them into a TCD class. I updated the idx_map variable to now have 2 components [traj_index,time_index]. Then concatenate all data together in the first dimension so that it is of shape (total sum of windows, window_length, dim).

[pietronvll]

Hi @mitchellostrow, thanks for this pull request. We have some similar functionality in the development branch, see

kooplearn/kooplearn/data.py

Line 261 in 7ddbdc0

class MultiTrajectoryContextDataset(TensorContextDataset):

The pre-release version implementing the multi-trajectory context windows are installable from pip by typing

pip install kooplearn==1.2.0b1

Let us know your feedbacks!