In batch cell verification, take commitment for each cell

[jtraglia]

In verify_cell_kzg_proof_batch(), we used to take as input a list of deduplicated commitments. This PR changes it so that it takes as input all the commitments. This is done for UI reasons since it's more natural for clients who do batch verification across multiple sidecars, and also for security reasons because it forces us to validate all the received sidecar commitments.

Specifically, to verify all cells from column 0 and column 1 we would pass something like this to verify_cell_kzg_proof_batch():

Inputs:
row_commitments = {commitment1, commitment2, commitment3}
row_indices = {0,1,2,0,1,2}
column_indices = {0,0,0,1,1,1}
cells = {cell1, cell2, cell3, cell4, cell5, cell6}
proofs = {proof1, proof2, proof3, proof4, proof5, proof6}

whereas after this PR we would give inputs:

Inputs:
row_commitments = {commitment1, commitment2, commitment3, commitment1, commitment2, commitment3}
column_indices = {0,0,0,1,1,1}
cells = {cell1, cell2, cell3, cell4, cell5, cell6}
proofs = {proof1, proof2, proof3, proof4, proof5, proof6}

Keep in mind that the helper function verify_cell_kzg_proof_batch_impl() stays the same as before (using the old input format), and we have the public function verify_cell_kzg_proof_batch() move from the new input format to the old input format, by checking the commitments provided and recreating row_indices.

[dankrad]

This will make verification slower, right? If you know which commitments are duplicates, then the commitment MSM will be smaller.

[jtraglia]

This will make verification slower, right?

It shouldn't make much of a difference. The public method will create a set of commitments and indices to each. The core verification method (verify_cell_kzg_proof_batch_impl) still accepts row indices like before. So the MSM will be the exact same and potentially better as there's no longer the opportunity for unused commitments.

[asn-d6]

Instead of doing a review, I pushed two commits as you suggested.
Please check them out and let me know what you think.

[asn-d6]

LGTM! Thanks!
Maybe we should see if @b-wagn can also review this?

[b-wagn]

LGTM, API change seems reasonable to me.

[jimmygchen]

Thanks @jtraglia @asn-d6 the change looks good to me!