bookworm-smart-assistant/scripts/patches/patch-c3-dirichlet-hardening.js

#!/usr/bin/env node
/**
 * Patch C3: adaptive-disambiguator Dirichlet 根治
 * - 懒初始化补 _initialAlphas 快照 (修 updateFromFeedback 路径)
 * - 竞争技能软衰减 (避免单调累积 → 权重塌陷)
 * - EMA 上限 (Σα ≤ 200，饱和后按比例回缩)
 * - adaptiveDisambiguate 归一化改 softmax-lite (不再用 maxAdj 线性缩放)
 * 幂等: sentinel C3_DIRICHLET_HARDENING_v1
 */
const fs = require('fs');
const path = require('path');

const TARGET = path.join(__dirname, '..', 'adaptive-disambiguator.js');
const SENTINEL = 'C3_DIRICHLET_HARDENING_v1';

function main() {
  const src = fs.readFileSync(TARGET, 'utf8');
  if (src.includes(SENTINEL)) { console.error('[patch-c3] already applied'); process.exit(0); }

  let patched = src;
  let hits = 0;

  // (1) 懒初始化补 _initialAlphas 快照 (L258-267)
  const old1 =
    "      if (!state.pairs[key]) {\r\n" +
    "        state.pairs[key] = {\r\n" +
    "          alphas: {\r\n" +
    "            [actualSkill]: CONFIG.weakPrior,\r\n" +
    "            [competing]: CONFIG.weakPrior,\r\n" +
    "          },\r\n" +
    "          totalSamples: 0,\r\n" +
    "          lastUpdated: null,\r\n" +
    "        };\r\n" +
    "      }";
  // 兼容 LF 情况
  const old1LF = old1.replace(/\r\n/g, '\n');
  let use1 = null;
  if (patched.includes(old1)) use1 = old1;
  else if (patched.includes(old1LF)) use1 = old1LF;
  if (!use1) { console.error('[patch-c3] anchor1 miss'); process.exit(3); }

  const NL = use1.includes('\r\n') ? '\r\n' : '\n';
  const new1 =
    "      if (!state.pairs[key]) {" + NL +
    "        // " + SENTINEL + ": 懒初始化必须保存 _initialAlphas 快照，否则 drift 报警失效" + NL +
    "        const _alphas = {" + NL +
    "          [actualSkill]: CONFIG.weakPrior," + NL +
    "          [competing]: CONFIG.weakPrior," + NL +
    "        };" + NL +
    "        state.pairs[key] = {" + NL +
    "          alphas: _alphas," + NL +
    "          _initialAlphas: { ..._alphas }," + NL +
    "          totalSamples: 0," + NL +
    "          lastUpdated: null," + NL +
    "        };" + NL +
    "      }";
  patched = patched.replace(use1, new1);
  hits++;

  // (2) updateFromFeedback α 更新: 加竞争软衰减 + EMA 上限
  // 锚点: `pair.alphas[actualSkill] = (pair.alphas[actualSkill] || CONFIG.weakPrior) + 1;`
  const old2 = "      pair.alphas[actualSkill] = (pair.alphas[actualSkill] || CONFIG.weakPrior) + 1;";
  if (!patched.includes(old2)) { console.error('[patch-c3] anchor2 miss'); process.exit(4); }
  const new2 =
    "      // " + SENTINEL + ": 正样本 +1; 竞争技能软衰减 (1%) 防止单调累积; EMA 上限 Σα ≤ 200" + NL +
    "      pair.alphas[actualSkill] = (pair.alphas[actualSkill] || CONFIG.weakPrior) + 1;" + NL +
    "      if (pair.alphas[competing] !== undefined && pair.alphas[competing] > CONFIG.weakPrior) {" + NL +
    "        pair.alphas[competing] = Math.max(CONFIG.weakPrior," + NL +
    "          pair.alphas[competing] - 0.01 * (pair.alphas[competing] - CONFIG.weakPrior));" + NL +
    "      }" + NL +
    "      // EMA 上限: 总样本量超过 200 时按比例回缩 (保持相对比例)" + NL +
    "      const _sumAlpha = Object.values(pair.alphas).reduce(function (s, a) { return s + a; }, 0);" + NL +
    "      if (_sumAlpha > 200) {" + NL +
    "        const _scale = 200 / _sumAlpha;" + NL +
    "        for (const _k of Object.keys(pair.alphas)) {" + NL +
    "          pair.alphas[_k] = Math.max(CONFIG.weakPrior, pair.alphas[_k] * _scale);" + NL +
    "        }" + NL +
    "      }";
  patched = patched.replace(old2, new2);
  hits++;

  // (3) adaptiveDisambiguate: softmax-lite 归一化
  const old3 =
    "    // 归一化 Bayesian 调整量到 [0, 1] 范围";
  if (!patched.includes(old3)) { console.error('[patch-c3] anchor3 miss'); process.exit(5); }
  // 精确锚点: 替换从"归一化"注释到"let maxAdj = 0" 块后两行，但为安全只改注释标签，配合下方 result.map 内逻辑替换
  // 更稳妥: 直接替换 map 里 `const normalizedAdj = maxAdj > 0 ? adj / maxAdj : 0;`
  const old3b = "      const normalizedAdj = maxAdj > 0 ? adj / maxAdj : 0;";
  if (!patched.includes(old3b)) { console.error('[patch-c3] anchor3b miss'); process.exit(6); }
  // 引入 softmax 归一化: 在函数内替换 maxAdj 计算块 + normalizedAdj 计算
  const old3Block =
    "    // 归一化 Bayesian 调整量到 [0, 1] 范围" + NL +
    "    let maxAdj = 0;" + NL +
    "    for (const adj of adjustments.values()) {" + NL +
    "      if (Math.abs(adj) > maxAdj) maxAdj = Math.abs(adj);" + NL +
    "    }";
  const old3BlockLF = old3Block.replace(/\r\n/g, '\n');
  let use3 = null;
  if (patched.includes(old3Block)) use3 = old3Block;
  else if (patched.includes(old3BlockLF)) use3 = old3BlockLF;
  if (!use3) { console.error('[patch-c3] anchor3 block miss'); process.exit(7); }
  const new3Block =
    "    // " + SENTINEL + ": softmax-lite 归一化，避免 maxAdj 线性缩放导致的 ±1 饱和" + NL +
    "    // Σ_i softmax(adj_i) = 1; 映射到 [-1, +1] 区间做 boost 基准" + NL +
    "    const _expAdj = new Map();" + NL +
    "    let _sumExp = 0;" + NL +
    "    for (const [_k, _v] of adjustments.entries()) {" + NL +
    "      const _e = Math.exp(Math.max(-5, Math.min(5, _v)));" + NL +
    "      _expAdj.set(_k, _e);" + NL +
    "      _sumExp += _e;" + NL +
    "    }" + NL +
    "    const _n = adjustments.size || 1;" + NL +
    "    const _uniform = 1 / _n;" + NL +
    "    // 兼容符号: 保留 maxAdj 供 fallback (若 softmax 退化则退回原算法)" + NL +
    "    let maxAdj = 0;" + NL +
    "    for (const adj of adjustments.values()) {" + NL +
    "      if (Math.abs(adj) > maxAdj) maxAdj = Math.abs(adj);" + NL +
    "    }";
  patched = patched.replace(use3, new3Block);
  hits++;

  // 并替换 normalizedAdj 使用 softmax 结果
  const new3b =
    "      // " + SENTINEL + ": softmax 归一化 (退化时 fallback 到 maxAdj 线性)" + NL +
    "      const _soft = _sumExp > 0 ? (_expAdj.get(c.name) || 0) / _sumExp : _uniform;" + NL +
    "      // 映射 [0,1] → [-1,+1]: (soft - uniform) / uniform，再按 maxAdj 兜底" + NL +
    "      const _softSigned = (_soft - _uniform) / Math.max(_uniform, 1e-6);" + NL +
    "      const normalizedAdj = Math.max(-1, Math.min(1, _softSigned));";
  patched = patched.replace(old3b, new3b);
  hits++;

  if (patched === src) { console.error('[patch-c3] no change'); process.exit(8); }
  if (hits < 4) { console.error('[patch-c3] expected 4 hits, got', hits); process.exit(9); }

  const bakDir = path.join(path.dirname(TARGET), 'patches', 'bak');
  if (!fs.existsSync(bakDir)) fs.mkdirSync(bakDir, { recursive: true });
  fs.writeFileSync(path.join(bakDir, 'adaptive-disambiguator.js.bak.c3.' + Date.now()), src);
  const tmp = TARGET + '.tmp.' + process.pid;
  fs.writeFileSync(tmp, patched);
  fs.renameSync(tmp, TARGET);
  console.error('[patch-c3] applied, hits=' + hits);
}
main();