How should an IterBridge be batched?

[tgross35]

For types that only implement ParallelIterator but none of the other traits, is there a good recipe to chunk the items before starting parallelism?

I have a function that generates an opaque -> impl Iterator<Item = T> so I am only able to use .par_bridge(). It is a case of cheap operations on a high number of items, so ideally a few thousand iterator items get collected to a Vec and then those chunks get run in parallel.

It is easy enough to do this manually but it seems like there may be a cleaner builtin way that I am not seeing. As an added benefit, if Rayon handles the chunking then it can clear reuse the vectors.