Description:

Implement the core side-chain packing algorithm over the prepared data models (ContactGraph, Conformations, SelfEnergyTable, PairEnergyTable, JunctionTree), producing a public pack() entry point that drives the full pipeline.

Tasks:

• Phase 1 — contact graph construction
• Phase 2 — conformation expansion (rotamers + H-orientations)
• Phase 3 — self-energy fill + threshold pruning + compaction
• Phase 4 — pair-energy fill
• Phase 5 — Goldstein DEE + Split DEE (iterative)
• Phase 6 — junction-tree DP → winner extraction + write-back
• pack() — public entry point