Tokens Service Performance Improvements

[stefangutica]

Reasoning

• Token processing had two major performance bottlenecks: repeated full-array filtering passes and repeated linear identifier lookups.
• In high-cardinality token datasets, these patterns increase CPU usage and allocation pressure, especially on frequently used endpoints.
• This PR improves performance by minimizing repeated scans and enabling cached key-based access for reused token arrays.

Proposed Changes

• Refactored getFilteredTokens() to evaluate all active filters in a single Array.filter() pass with early exits.
• Precomputed normalized filter inputs once per request (search/name/identifier/identifiers and mexPairType values) to avoid repeated string operations.
• Introduced ArrayIndexer backed by WeakMap to cache per-array, per-property positions and provide near O(1) lookups after first index build.
• Replaced multiple identifier-based linear searches in TokenService and CacheWarmerService with ArrayIndexer.getItemByKeyValue(...).
• Preserved functional behavior and sorting semantics while reducing redundant work in hot paths.

How to test

• Run existing token and cache warmer test suites and verify no behavioral regressions.
• Validate filter combinations (search, name, identifier, identifiers, includeMetaESDT, mexPairType, priceSource) and confirm output parity.
• Validate token retrieval, token-with-balance, and token-with-roles flows for both existing and missing identifiers.
• Compare before/after latency and CPU consumption using representative load tests on large token lists.

[github-actions]

k6 load testing comparison.
Base Commit Hash: 79b2062
Target Commit Hash: 119b26b

Metric	Base				Target				Diff
	Avg	Max	90	95	Avg	Max	90	95	Avg	Max	90	95
Accounts	94.11	664.90	296.90	331.54	70.45	1377.33	101.79	123.57	-25.14% ✅	+107.15% 🔴	-65.71% ✅	-62.73% ✅
Transactions	694.84	49377.95	106.13	107.97	82.21	2183.12	111.93	122.80	-88.17% ✅	-95.58% ✅	+5.47% 🔴	+13.73% 🔴
Mex	88.40	767.22	212.44	327.33	61.70	339.03	66.31	68.96	-30.20% ✅	-55.81% ✅	-68.78% ✅	-78.93% ✅
Pool	90.82	760.51	265.92	328.77	63.36	1263.58	66.48	69.10	-30.24% ✅	+66.15% 🔴	-75.00% ✅	-78.98% ✅
Blocks	114.34	870.17	371.55	436.12	72.71	1404.57	78.65	90.07	-36.41% ✅	+61.41% 🔴	-78.83% ✅	-79.35% ✅
Nodes	986.50	35748.13	214.82	360.06	61.90	339.07	66.28	69.04	-93.72% ✅	-99.05% ✅	-69.15% ✅	-80.83% ✅
Tokens	91.25	800.49	272.95	329.53	62.49	543.88	66.46	69.02	-31.52% ✅	-32.06% ✅	-75.65% ✅	-79.05% ✅
Test Run Duration	60003.17				60006.11

Legend: Avg - Average Response Time, Max - Maximum Response Time, 90 - 90th Percentile, 95 - 95th Percentile
All times are in milliseconds.