[JonKuqi]

Replaces the standalone d-voting Pedersen DKG with a thin adapter that wraps pedersen, so d-voting gets dela's DKG features (resharing, reencryption, verifiable encryption) and we stop maintaining two divergent copies.

The old duplicated dkg on d-voting has 4,532 lines of code. The adapter wrapps dela's dkg and adds d-voting specific functionality that reduces it to 1,393 lines of code.

⚠ IMPROTANT: For now this is a Proof of Concept only. Dela PR 45 needs to be merged into main and decide which functions are needed, then this branch needs to be updated by fixing import paths directly to dela instead of having stubs. Therefore the goal of this PR is just to show a maintainable solution to the duplicate DKGs.

What d-voting-specific functionality the adapter adds

• Form binding — one DKG actor per formID, stored in a dkgMap bucket (Listen(formID, txmngr) / GetActor(formID)).
• On-chain pubshare publication — ComputePubshares() computes each node's partial decryptions and submits them as transactions to the evoting smart contract, retrying until accepted or enough nodes have submitted.
• Status reporting — Status() with 7 status codes and a Prometheus metric, surfaced to the frontend.
• Persistence — HandlerData + MarshalJSON + ReadActors so actors survive process restarts.
• Encrypt signature shim — maps dela's Encrypt(K, Cs) to d-voting's (K, C, remainder).

Benefits

• Less code to maintain
• dela's features for free — resharing, reencryption, and verifiable encryption come automatically from dela, with no re-porting.
• One source of truth — a single DKG implementation in dela, so future dela DKG work benefits d-voting automatically and the two copies can't drift apart.