核對搜尋改寫有否保留比較對象

訂明本例驗收任務

假設搜尋改寫先由同事標出比較對象的固定編號。改寫須保留全部原對象，不能加入新對象；欠項與新增項分開列。本例只核對已標出的清單，不聲稱可自動理解任意文字。

保存原對象編號清單
雙向核對改寫對象
修正欠項或額外範圍

對象未能標明，先由同事確認。

使用已填個案

原創假設輸入及確切答案鍵
個案	已提供輸入	預期結構化結果	判斷及原因
漏一個比較對象	{"original":["induction","refresher"],"rewritten":["induction"]}	{"missing":["refresher"],"added":[],"preserved":false}	改寫後沒有問重溫程序。
對象相同而次序不同	{"original":["induction","refresher"],"rewritten":["refresher","induction"]}	{"missing":[],"added":[],"preserved":true}	本例對象用無序集合，兩個仍齊備。
加入無關對象	{"original":["induction"],"rewritten":["induction","payroll"]}	{"missing":[],"added":["payroll"],"preserved":false}	保留入職程序也不夠；新增 payroll 改了搜尋範圍。

下載可編輯 CSV ↓

逐步處理 new-target 個案，須得到甚麼結果？原因是甚麼？

查看參考答案

{"missing":[],"added":["payroll"],"preserved":false}。保留入職程序也不夠；新增 payroll 改了搜尋範圍。

執行完整本機驗算

另存為 .mjs，以 Node.js 執行；只驗提供的個案

import assert from 'node:assert/strict';

function evaluate(x){
 const a=[...new Set(x.original)].sort(),b=[...new Set(x.rewritten)].sort();
 const missing=a.filter(id=>!b.includes(id)),added=b.filter(id=>!a.includes(id));return {missing,added,preserved:missing.length+added.length===0};
}

const cases = [
  {
    "id": "lost-comparison",
    "input": {
      "original": [
        "induction",
        "refresher"
      ],
      "rewritten": [
        "induction"
      ]
    },
    "expected": {
      "missing": [
        "refresher"
      ],
      "added": [],
      "preserved": false
    }
  },
  {
    "id": "reordered-targets",
    "input": {
      "original": [
        "induction",
        "refresher"
      ],
      "rewritten": [
        "refresher",
        "induction"
      ]
    },
    "expected": {
      "missing": [],
      "added": [],
      "preserved": true
    }
  },
  {
    "id": "new-target",
    "input": {
      "original": [
        "induction"
      ],
      "rewritten": [
        "induction",
        "payroll"
      ]
    },
    "expected": {
      "missing": [],
      "added": [
        "payroll"
      ],
      "preserved": false
    }
  }
];
for (const c of cases) assert.deepEqual(evaluate(c.input), c.expected, c.id);
console.log(JSON.stringify({checked:cases.length,caseIds:cases.map(c=>c.id)}));

下載可執行例子 ↓

原問法與比較對象標籤跟改寫一併保留。集合通過只確認這些對象仍在，沒有驗證整句自然語言。

訂明本例驗收任務

使用已填個案

執行完整本機驗算

相關資源