[RL Unplugged] How to recover the sequential data

Hi, I am interested in testing my algorithm on `dm_control` dataset as benchmark and I am trapped in data loading part.

Since `dm_control` is a set of sequential control tasks and I have seen in `dm_control_suite_crr.ipynb` the CRR algorithm is using LSTM, I assume the dataset must have structure to store the time order of the data. But, when I try the demos in the repository, the dataset seem to be shuffled and I only get tuples like (s, a, r, s', a').

I am not familiar with `tf.data.Dataset`, so could anyone hint me out about how to recover the sequential structure of the data? Or, the sequential structure is not stored by the dataset at all? Thanks in advance!

Start with dm_control_suite_crr.ipynb and the repository demos; inspect how tf.data.Dataset is built and whether shuffling occurs. Determine whether sequential ordering is preserved or discarded, then document how users can recover it or state that it is unavailable.

Written by the indexing model from the issue text.