Progressively merge `das` to `unstable`

[jimmygchen]

Description

Progressively merge das branch into unstable. Full das branch PR for reference 👇

• PeerDAS implementation #5683

The following proposed by @dapplion :

PeerDAS to unstable (topological sort)

• Add DataColumn type - Add PeerDAS DataColumn type #5913
• Store changes to persist data columns - Store changes to persist data columns #6073
• da_checker is_available update - Update is_available check to support PeerDAS. #6076
• Add gossip data_column topic - Add DataColumnSidecar gossip topic and message handling #6147
• Add gossip validation - Data column gossip validation and error handling #6181
• Subscribe to data_column topic on peerdas epoch
• Import gossip data_column to da_checker Import gossip data column into data availability checker #6197
• Add KZG lib integration Add PeerDAS KZG lib integration (construction & KZG verification) #6212
• Add supernode functionality Add plumbing for PeerDAS supernodes (#5050, #5409, #5570, #5966) #6216

Conga line from here

1. Implement data columns by network boilerplate #6224
2. Add PeerDAS RPC import boilerplate #6238
3. Implement PeerDAS RPC handlers #6237
4. Persist data columns to store #6255

• Block production + publishing updates <<< Here e2e test a supernode only network
• Reconstruction
• Lookup sync (custody) updates <<< Here e2e test a N regular node + 1 supernode network
• Range sync (custody) updates <<< Here e2e test adding new node and sync to head

(here have parity with das branch w.r.t. custody)

• Re-process early sampling requests
• Checkpoint Sync data columns
• Respond with first byte to extend timeout for early sampling requests
• Batch verify KZG proofs where possible

(now tackle sampling, which should be purely additive)

• Add sampling
• Network service update to maintain peers on all columns
• Forkchoice updates w.r.t. sampling result

... plus improvements, metrics, etc

[jimmygchen]

Completed in #5683 🎉