Shuffle buffer for controlling sample correlation

Nobody has claimed this yet.

## 🚀 Feature
Add a sample level shuffle buffer like `webdataset` to StreamingDataloader, to make shuffling more random when datasets are built with some correlation.

### Motivation
I have some data that correlates across samples when building, and I observe that when using a smaller number of workers, data correlation. This is because litdata only shuffles within chunk, so when the chunk is large and many samples in the chunk are correlated, the randomly drawn samples are still correlated with each other. Exact shuffle is, in general, not tractable - but a good heuristic can be to maintain a shuffle buffer explicitly like `webdataset`. As long as the buffer size is larger than a chunk, data correlation might be improved.

### My case
Here's an example:

The logged value is the sparsity of input. You can see that different clusters of sparsity persist for quite a while, likely because they are shuffling from the same chunk.

Start by locating StreamingDataloader and reading how it currently shuffles samples within chunks. Determine where a configurable sample-level shuffle buffer would fit; done means the dataloader provides the requested buffer behavior and improves mixing across correlated samples without breaking existing loading behavior.

Written by the indexing model from the issue text.