Cache coin symbol counts and refactor duplicate-symbol detection in pool.js

[MoneroOcean]

Motivation

• Improve performance when resolving duplicate coin symbols by avoiding repeated scans of poolStats.coins for each lookup.
• Ensure stable duplicate detection when poolStats.coins reference hasn't changed by caching computed symbol counts.
• Handle empty/invalid inputs early to avoid unnecessary work.

Description

• Introduced a WeakMap named coinSymbolCountsCache to cache symbol count maps keyed by the poolStats object.
• Refactored hasDuplicateCoinSymbol to early-return when the slug is empty and to consult cached counts via a new getCoinSymbolCounts function.
• Implemented getCoinSymbolCounts to compute a Map of slug -> count from coinStatsRows(poolStats), cache it with the coins reference as coinsRef, and return the counts.
• Updated counting logic to skip empty slugs and to use the cached counts for duplicate checks (counts.get(wanted) > 1).

Testing

• Ran the test suite with npm test, which passed successfully.
• Exercised coin routing and duplicate detection logic through unit tests covering coinRouteId, coinRouteSlug, and hasDuplicateCoinSymbol, all of which passed.

---

Codex Task