feat: add configurable account selection strategies

Refactor account selection into a strategy pattern with three options:
- Sticky: cache-optimized, stays on same account until rate-limited
- Round-robin: load-balanced, rotates every request
- Hybrid (default): smart distribution using health scores, token buckets, and LRU

The hybrid strategy uses multiple signals for optimal account selection:
health tracking for reliability, client-side token buckets for rate limiting,
and LRU freshness to prefer rested accounts.

Includes WebUI settings for strategy selection and unit tests.

Co-Authored-By: Claude <noreply@anthropic.com>

src/config.js

@@ -15,7 +15,26 @@ const DEFAULT_CONFIG = {
     persistTokenCache: false,
     defaultCooldownMs: 10000,  // 10 seconds
     maxWaitBeforeErrorMs: 120000, // 2 minutes
-    modelMapping: {}
+    modelMapping: {},
+    // Account selection strategy configuration
+    accountSelection: {
+        strategy: 'hybrid',           // 'sticky' | 'round-robin' | 'hybrid'
+        // Hybrid strategy tuning (optional - sensible defaults)
+        healthScore: {
+            initial: 70,              // Starting score for new accounts
+            successReward: 1,         // Points on successful request
+            rateLimitPenalty: -10,    // Points on rate limit
+            failurePenalty: -20,      // Points on other failures
+            recoveryPerHour: 2,       // Passive recovery rate
+            minUsable: 50,            // Minimum score to be selected
+            maxScore: 100             // Maximum score cap
+        },
+        tokenBucket: {
+            maxTokens: 50,            // Maximum token capacity
+            tokensPerMinute: 6,       // Regeneration rate
+            initialTokens: 50         // Starting tokens
+        }
+    }
 };
 
 // Config locations