feat(keys): add trade execution simulation endpoint

[TS-mfon]

Summary

Implements the trade execution simulation endpoint, allowing prospective buyers and sellers to preview bonding curve pricing, average price per unit, and percentage price impact prior to submitting on-chain transactions.

Closes #870

---

Implementation Details

• Endpoint: GET /api/v1/keys/:keyId/simulate?quantity=N&side=buy|sell
• Multi-Quantity Support: Accepts comma-separated quantities (e.g. quantity=1,5,10,25) to preview tiered execution scenarios in a single round-trip.
• Constant-Time Integration: Uses closed-form integral evaluation over milestone curve intervals ($\int_{s_0}^{s_1} P(s) ds$) to achieve $O(1)$ runtime and eliminate event-loop blocking from discrete step loops.
• Safety & Validation:
  • Enforces batch size limits (maximum 20 quantities per request).
  • Enforces quantity ceiling (maximum 1,000,000 keys per entry).
  • Rejects sell orders where quantity exceeds circulating supply with an explicit 422 error.
• Caching: Previews cached in Redis for 15 seconds (cache:keys:simulate:{keyId}:{side}:{quantities}).

---

Testing

• Unit test suite added in src/modules/keys/__tests__/key-simulate.test.ts verifying:
  • 422 for invalid/missing side or non-positive quantities.
  • 404 for nonexistent key.
  • 422 rejection when sell quantity exceeds circulating supply.
  • Buy simulation calculations across single and batch quantities.
  • Sell simulation calculations and Redis 15s caching.
• All 6 unit tests pass; TypeScript build passes cleanly.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[TS-mfon]

Superseded by #1007, which carries all four issues (#866, #869, #870, #872) on one branch rebased on latest main.

Splitting these four across separate PRs meant they collided with each other: all four touch src/modules/keys/keys.routes.ts, and #869/#872 both touch src/modules/creator/creator.routes.ts and the prisma schema, so whichever merged first would break the rest. #1007 also addresses the Copilot findings raised here — the migration moved to prisma/schema/migrations and non-finite milestone values are now rejected.

Closing this one in favour of #1007.