wikipedia.mjs

const endpoint = "https://zh.wikipedia.org/w/api.php";
const userAgent = "QuickLearn/1.0 (https://github.com/fly1d/quicklearn-agent)";

function cleanText(value = "") {
  return String(value).replace(/\s+/g, " ").trim();
}

function sentenceList(value, limit = 6) {
  return cleanText(value)
    .split(/(?<=[。！？.!?])\s*/)
    .map((item) => item.trim())
    .filter((item) => item.length >= 12)
    .slice(0, limit);
}

function normalizeTerm(value = "") {
  return value.toLowerCase().replace(/[\s_（）()·.-]+/g, "");
}

function isDisambiguation(page) {
  return Object.hasOwn(page.pageprops || {}, "disambiguation") || /消歧[义義][页頁]/.test(page.description || "");
}

function sourceUrl(title) {
  return `https://zh.wikipedia.org/wiki/${encodeURIComponent(title.replace(/\s+/g, "_"))}`;
}

function clarificationResult(input, pages) {
  const top = pages[0];
  const disambiguation = pages.find(isDisambiguation);
  if (!top || !disambiguation) return null;

  const exactDisambiguation = normalizeTerm(disambiguation.title) === normalizeTerm(input);
  const topIsExact = normalizeTerm(top.title) === normalizeTerm(input);
  const likelyAmbiguous = top === disambiguation || exactDisambiguation || (!topIsExact && disambiguation.index <= 2);
  if (!likelyAmbiguous) return null;

  const options = [...new Set(pages
    .filter((page) => !isDisambiguation(page) && !/列表条目|列表條目/.test(page.description || ""))
    .map((page) => page.title))]
    .slice(0, 3);
  if (options.length < 2) return null;

  return {
    needsClarification: true,
    question: `“${input}”可能有几种意思，你想了解哪一个？`,
    options
  };
}

function learningResult(page) {
  const facts = sentenceList(page.extract);
  const description = cleanText(page.description);
  const summary = (facts.slice(0, 2).join(" ") || description).slice(0, 360);
  const details = [
    facts[0] || description || `这是关于 ${page.title} 的百科条目。`,
    facts[1] || `${page.title} 的背景和适用语境值得结合原文继续确认。`,
    facts[2] || `可以通过相近概念和实际例子进一步理解 ${page.title}。`
  ];
  const takeaways = facts.slice(0, 3).map((item) => item.slice(0, 92));
  while (takeaways.length < 3) {
    takeaways.push(["先掌握准确定义", "再确认背景与边界", "用例子验证理解"][takeaways.length]);
  }

  return {
    title: page.title,
    category: "维基百科速览",
    summary,
    definition: details[0],
    why: description
      ? `${page.title}通常被概括为“${description}”。理解它有助于建立相关主题的基础背景和概念边界。`
      : `理解 ${page.title} 的定义、背景与边界，可以为继续阅读专业资料建立稳定起点。`,
    concepts: [["基本定义", details[0]], ["关键背景", details[1]], ["延伸理解", details[2]]],
    takeaways,
    misconception: "百科摘要适合建立第一层认识，但条目可能持续更新；涉及专业判断时仍应核对原始资料。",
    steps: [`用一句话复述 ${page.title}`, `区分 ${page.title} 与相近概念`, "打开来源，重点查看定义、背景和示例"],
    source: {
      url: sourceUrl(page.title),
      host: "zh.wikipedia.org",
      title: page.title,
      provider: "维基百科",
      license: "CC BY-SA 4.0",
      excerpts: facts.map((item) => item.slice(0, 280)).slice(0, 6)
    }
  };
}

export async function lookupWikipedia(input, options = {}) {
  const query = input.trim();
  if (!query || query.length > 120) return null;

  const fetchImpl = options.fetchImpl || fetch;
  const controller = new AbortController();
  const timeout = setTimeout(() => controller.abort(), options.timeoutMs ?? 3500);
  timeout.unref?.();

  try {
    const url = new URL(endpoint);
    url.search = new URLSearchParams({
      action: "query",
      generator: "search",
      gsrsearch: query,
      gsrnamespace: "0",
      gsrlimit: "6",
      prop: "extracts|description|pageprops",
      exintro: "1",
      explaintext: "1",
      exsentences: "6",
      ppprop: "disambiguation",
      redirects: "1",
      format: "json",
      formatversion: "2"
    });
    const response = await fetchImpl(url, {
      signal: controller.signal,
      headers: { Accept: "application/json", "User-Agent": userAgent }
    });
    if (!response.ok) return null;
    const data = await response.json();
    const pages = Array.isArray(data?.query?.pages)
      ? data.query.pages
        .filter((page) => page?.ns === 0 && typeof page.title === "string")
        .sort((a, b) => (a.index ?? Number.MAX_SAFE_INTEGER) - (b.index ?? Number.MAX_SAFE_INTEGER))
      : [];
    if (!pages.length) return null;

    const clarification = clarificationResult(query, pages);
    if (clarification) return clarification;
    const page = pages.find((item) => !isDisambiguation(item) && cleanText(item.extract).length >= 30);
    return page ? learningResult(page) : null;
  } catch {
    return null;
  } finally {
    clearTimeout(timeout);
  }
}