src/lib/agent/RouteSelector.ts

/**
 * RouteSelector.ts — Adaptive A/B routing between GraphOrchestrator and ReAct (FASE 5)
 *
 * Selects the best execution path based on:
 *   1. Deterministic rules (low complexity → react, high+3tools → graph)
 *   2. Historical quality scores per task signature (stored in agentMemory)
 *   3. 10% epsilon-greedy exploration to avoid local optima
 *
 * S344: RouteStats ora traccia qualityScore (0–1 da goalVerifier) invece di binario
 *       success/failure — il RouteSelector sceglie per qualità reale, non lunghezza testo.
 *
 * 100% local, no API calls, synchronous selection, async persistence.
 */

import { agentMemory } from "../agentMemory";
import type { PlanLike } from "./GraphOrchestrator";

export type Route = "graph" | "react";

export interface RouteOutcome {
  route:         Route;
  success:       boolean;
  latencyMs:     number;
  qualityScore?: number;  // S344: 0–1 da goalVerifier (sostituisce heuristica lunghezza)
}

interface RouteStats {
  gs:  number; // graph successes (binario — backward compat)
  gt:  number; // graph total
  rs:  number; // react successes (binario — backward compat)
  rt:  number; // react total
  gqs: number; // graph quality score sum (S344 — goalVerifier scores)
  rqs: number; // react quality score sum  (S344 — goalVerifier scores)
}

const STORAGE_KEY = "route_selector:v2"; // S344: bump version — nuovo schema stats
const MIN_SAMPLES_FOR_STATS = 2; // FIX-5a: 2 sample già sufficienti su progetti piccoli
const EXPLORE_RATE = 0.12; // FIX-5b: 12% — più esplorazione con pochi task totali

function planSignature(plan: PlanLike): string {
  const tools = [...new Set(plan.subtasks.filter(s => s.tool !== "respond").map(s => s.tool))].sort().join(",");
  return `${plan.complexity}|${tools}`;
}

/**
 * S344: usa qualityScore medio se disponibile (goalVerifier), altrimenti successRate binario.
 * qualitySum > 0 → almeno una sessione con goalVerifier attivo → score medio più accurato.
 */
function winRate(success: number, total: number, qualitySum: number): number {
  if (total === 0) return 0.5; // prior ottimistico al primo run
  if (qualitySum > 0) return qualitySum / total; // goalVerifier score — il più accurato
  // FIX-5c: Laplace smoothing — (success+1)/(total+2) evita 0.0/1.0 con 1-2 sample
  return (success + 1) / (total + 2);
}

export class RouteSelector {
  private stats = new Map<string, RouteStats>();
  private loaded = false;

  // Lazy-load from agentMemory on first use
  async ensureLoaded(): Promise<void> {
    if (this.loaded) return;
    this.loaded = true;
    try {
      // Tenta prima la chiave v2 (S344), poi v1 per backward compat
      const raw = agentMemory.get(STORAGE_KEY) ?? agentMemory.get("route_selector:v1");
      if (raw) {
        const obj = JSON.parse(raw) as Record<string, RouteStats>;
        for (const [k, v] of Object.entries(obj)) {
          // S344: backward compat — old v1 entries non hanno gqs/rqs
          this.stats.set(k, { ...v, gqs: v.gqs ?? 0, rqs: v.rqs ?? 0 });
        }
      }
    } catch { /* first run — empty stats */ }
  }

  private persist(): void {
    try {
      const obj: Record<string, RouteStats> = {};
      for (const [k, v] of this.stats.entries()) {
        obj[k] = v;
      }
      agentMemory.set(STORAGE_KEY, JSON.stringify(obj), "routing");
    } catch { /* non-blocking */ }
  }

  /** Select the route for a given plan. Must call ensureLoaded() first. */
  select(plan: PlanLike): Route {
    const toolCount = plan.subtasks.filter(s => s.tool !== "respond").length;

    // ── Hard rules (deterministic) ──────────────────────────────────────────
    if (plan.complexity === "low" || toolCount <= 1) return "react";
    if (plan.complexity === "high" && toolCount >= 3)  return "graph";

    // ── Epsilon-greedy exploration ───────────────────────────────────────────
    if (Math.random() < EXPLORE_RATE) {
      return Math.random() < 0.5 ? "graph" : "react";
    }

    // ── Stats-based selection (S344: qualità-pesata) ─────────────────────────
    const sig = planSignature(plan);
    const s = this.stats.get(sig);

    if (!s || (s.gt + s.rt) < MIN_SAMPLES_FOR_STATS) {
      // Default for medium complexity: graph if 2+ tools, react otherwise
      return toolCount >= 2 ? "graph" : "react";
    }

    const graphWin = winRate(s.gs, s.gt, s.gqs);
    const reactWin = winRate(s.rs, s.rt, s.rqs);
    return graphWin >= reactWin ? "graph" : "react";
  }

  /**
   * Record the outcome of a completed execution for future routing decisions.
   * S344: qualityScore (da goalVerifier) sostituisce l'heuristica binaria length > 80.
   * Fallback: se qualityScore non disponibile (task senza formalGoal), usa 0.8/0.2.
   */
  record(plan: PlanLike, outcome: RouteOutcome): void {
    const sig = planSignature(plan);
    const s = this.stats.get(sig) ?? { gs: 0, gt: 0, rs: 0, rt: 0, gqs: 0, rqs: 0 };

    // S344: quality score da usare — goalVerifier se disponibile, fallback binario
    const quality = outcome.qualityScore ?? (outcome.success ? 0.8 : 0.2);

    if (outcome.route === "graph") {
      s.gt++;
      if (outcome.success) s.gs++;
      s.gqs += quality;
    } else {
      s.rt++;
      if (outcome.success) s.rs++;
      s.rqs += quality;
    }

    this.stats.set(sig, s);
    this.persist();
  }

  /** Debug: return stats for all known signatures */
  dump(): Record<string, RouteStats> {
    const out: Record<string, RouteStats> = {};
    for (const [k, v] of this.stats.entries()) out[k] = v;
    return out;
  }
}

export const routeSelector = new RouteSelector();