Prefetch parquet metadata in the rapidsmpf streaming network

https://github.com/rapidsai/cudf/pull/22700 adds metadata prefetching to cudf-polars. It implements it as a distinct stage that runs after prefetching and before IR execution in a rapidsmpf streaming network.

While metadata prefetching should be relatively fast compared to IR execution, the barrier created by prefetching isn't optimal. It waits for *all* metadata prefetching to *complete* before *any* of the Scan nodes can *begin*. But a Scan node really only requires the prefetched metadata for the `paths` it reads, not all the paths in the entire query.

We can fix this by removing the prefetching barrier and doing the

https://github.com/TomAugspurger/pygdf/commit/18913f5ce5155e45643124bfba3901aa90b15890 is out of date now, but has a POC for how this might be done. At a high level:

1. traverse the IR graph, grouping the Scan nodes by `paths` (we want to read each parquet metadata exactly once)
2. Create actors responsible for each group of parquet metadata
3. Feed the metadata into the Actors performing the Scans