Add Axiom V2 exact subset solver

[Dospacite]

Summary

Adds Axiom V2, an exact subset-reachability skill point solver by Dospacite.

The solver models each non-free equipment item as a candidate bit. For every candidate subset, skill totals and cascade thresholds are deterministic, so the algorithm explores reachable subsets instead of equip-order permutations. AxiomPlayer preprocesses equipment into packed requirement, bonus, threshold, and weight arrays.

README checklist

• Implemented under com.wynncraft.algorithms.
• Registered in AlgorithmRegistry as Axiom V2.
• Uses a custom IPlayer implementation: AxiomPlayer.
• Implements clearCache() and self-invalidates cached state by equipment array, candidate count, and assigned base totals.
• Does not mutate IEquipment instances.
• Tie-breaks by item count, then total skill point bonus weight.

Validation

./gradlew test -Palgo='Axiom V2' --rerun-tasks
Axiom V2 PASS: 120 FAIL: 0

Benchmark

Short local JMH run on OpenJDK 21.0.10:

./gradlew jmh -Palgo='Axiom V2' -PjmhArgs='-wi 1 -i 1 -f 1'

Result summary:

FullEquipBenchmark mean: 63.992 ns/op
OneByOneBenchmark mean: 2891.091 ns/op