Update get_data_column_sidecars to take cells/proofs

[jtraglia]

We received feedback from clients that they were unable to re-create data column sidecars after performing recovery because get_data_column_sidecars took blobs, but recover_cells_and_kzg_proofs provided cells/proofs and there was no public method to convert the cells back to a blob. I think it makes sense to change this function to take cells/proofs instead of blobs.

[ppopth]

This makes sense to me.

[agnxsh]

It would be nice to maybe overload this function from an implementation POV. Because this function is still used with blobs getting passed, in the proposer impls. Exactly where proposer extracts blob sidecars from blob bundle and then gets data column sidecars from the blobs, at slot start.

ref - https://github.com/prysmaticlabs/prysm/blob/peerDAS/beacon-chain/rpc/prysm/v1alpha1/validator/proposer.go#L379-L411

cc @nisdas @nalepae correct me if i am wrong

[ppopth]

Because this function is still used with blobs getting passed, in the proposer impls. Exactly where proposer extracts blob sidecars from blob bundle and then gets data column sidecars from the blobs, at slot start.

You can call compute_cells_and_kzg_proofs before calling get_data_column_sidecars. It's quite simple. Just like you call recover_cells_and_kzg_proofs in reconstruction.

[agnxsh]

Because this function is still used with blobs getting passed, in the proposer impls. Exactly where proposer extracts blob sidecars from blob bundle and then gets data column sidecars from the blobs, at slot start.

You can call compute_cells_and_kzg_proofs before calling get_data_column_sidecars. It's quite simple. Just like you call recover_cells_and_kzg_proofs in reconstruction.

OK, thanks

[asn-d6]

LGTM! Thanks!

[Aga92-92]

``