WikitTeam · umouyoumou-del · Aug 21, 2026
diff --git a/auto-attribution.js b/auto-attribution.js
@@ -12,6 +12,7 @@ const path = require('path');
 const prisma = require('./lib/prisma');
 const config = require('./wikitdb.config.js');
 const { fetchAttributionPage, aggregateAuthorScores, normalizePageKey } = require('./utils/attribution');
+const { fetchCromSiteRanking, fetchCromSitePages, toCromHttpBase } = require('./utils/crom');
 
 const httpAgent = new http.Agent({ keepAlive: true, maxSockets: 10 });
 const httpsAgent = new https.Agent({ keepAlive: true, maxSockets: 10 });
@@ -80,7 +81,7 @@ async function processSite(siteConfig) {
         return;
     }
 
-    // 1. 全量同步归属记录
+    // 1. 全量同步归属记录（用于详情页展示作者归属页面列表）
     await prisma.$transaction([
         prisma.authorAttribution.deleteMany({ where: { siteParam: wikiParam } }),
         prisma.authorAttribution.createMany({
@@ -89,31 +90,106 @@ async function processSite(siteConfig) {
         })
     ]);
 
-    // 2. 获取页面评分（listpages）
-    const allPages = await fetchAllPages(siteConfig);
+    const baseUrl = siteConfig.URL.replace(/\/$/, '');
+    const cromEnabled = !!siteConfig.CROM_API;
+    const cromHttpBase = cromEnabled ? toCromHttpBase(baseUrl) : '';
+
+    // 2. 对开启 CROM_API 的站点：统一使用 crom 作为权威源
+    //      - fetchCromSiteRanking: 作者排行榜 (author_score)
+    //      - fetchCromSitePages : 页面评分 + 搜索索引 (page_scores / pages_index)
+    //    未开启或 crom 失败：fallback 到 listpages + 归属页聚合
+    let cromScores = null;
+    let pageScoreRows = null;
+    let pagesIndexRows = null;
+    let allPages = [];
+    let scoreSource = 'attribution';
+    let pagesSource = 'listpages';
+
+    if (cromEnabled) {
+        // 串行（先排行再页面）+ 保守并发，避免同时触发 crom 服务器限流
+        // （crom complexity + 频率检查叠加后，并发跑两边经常被 ban 几十秒）
+        try {
+            const data = await fetchCromSiteRanking(baseUrl, {
+                request,
+                endpoint: siteConfig.CROM_API,
+                concurrency: 3,
+                batchSleepMs: 500,
+                onPage: (p) => { if (p.rank % 100 === 1) logLine(`[归属] ${wikiParam} crom 排行 rank=${p.rank} 已得 ${p.fetched} 作者`); }
+            });
+            if (data && Object.keys(data).length > 0) {
+                cromScores = data;
+                scoreSource = 'crom';
+            }
+        } catch (e) {
+            logLine(`[归属] ${wikiParam} crom 作者排行失败: ${e.message}`);
+        }
+        try {
+            const data = await fetchCromSitePages(cromHttpBase, {
+                request,
+                endpoint: siteConfig.CROM_API,
+                pageSize: 50,
+                batchSleepMs: 300,
+                onProgress: (p) => { if (p.pageNum % 10 === 1) logLine(`[归属] ${wikiParam} crom 页面已拉 ${p.pages} 条 (第${p.pageNum}批)`); }
+            });
+            if (data && data.scores && data.scores.length > 0) {
+                pageScoreRows = data.scores;
+                pagesIndexRows = data.index;
+                pagesSource = 'crom';
+            }
+        } catch (e) {
+            logLine(`[归属] ${wikiParam} crom 页面抓取失败: ${e.message}`);
+        }
+    }
 
-    // 3. 保存页面评分
-    const pageScoreRows = allPages.map(p => ({
-        page: p.page, title: p.title, rating: p.rating,
-        upvotes: p.upvotes, downvotes: p.downvotes
-    }));
-    await prisma.setting.upsert({
-        where: { key: `page_scores:${wikiParam}` },
-        update: { value: JSON.stringify(pageScoreRows) },
-        create: { key: `page_scores:${wikiParam}`, value: JSON.stringify(pageScoreRows) }
-    });
+    // 3. crom 未启用或失败时，fallback 到 listpages 拿页面评分
+    if (!pageScoreRows) {
+        allPages = await fetchAllPages(siteConfig).catch(e => {
+            logLine(`[归属] ${wikiParam} listpages 抓取失败: ${e.message}`);
+            return [];
+        });
+        pageScoreRows = allPages.map(p => ({
+            page: p.page, title: p.title, rating: p.rating,
+            upvotes: p.upvotes, downvotes: p.downvotes
+        }));
+    }
+
+    // 4. 保存页面评分（page_scores 用于详情页页面列表）
+    if (pageScoreRows && pageScoreRows.length > 0) {
+        await prisma.setting.upsert({
+            where: { key: `page_scores:${wikiParam}` },
+            update: { value: JSON.stringify(pageScoreRows) },
+            create: { key: `page_scores:${wikiParam}`, value: JSON.stringify(pageScoreRows) }
+        });
+    }
+
+    // 4b. 保存搜索索引（crom 源才保存，因为 listpages 不含 component/theme/art 等页，搜索本来就不全；
+    //      非 crom 站点仍走 Wikit GraphQL articles 接口）
+    if (pagesIndexRows && pagesIndexRows.length > 0) {
+        await prisma.setting.upsert({
+            where: { key: `pages_index:${wikiParam}` },
+            update: { value: JSON.stringify(pagesIndexRows) },
+            create: { key: `pages_index:${wikiParam}`, value: JSON.stringify(pagesIndexRows) }
+        });
+    }
 
-    // 4. 按归属聚合作者分数（每个归属用户获得页面全分）
-    const pageRatings = new Map();
-    for (const pg of allPages) pageRatings.set(normalizePageKey(pg.page), { rating: pg.rating });
-    const authorScores = aggregateAuthorScores(attrRecords, pageRatings);
+    // 5. 生成 author_score：
+    //    优先 crom（含组件/版式/艺术等所有归属页面，与作者页/官方排行一致）；
+    //    crom 不可用时 fallback 到归属页 + listpages 聚合。
+    let authorScores;
+    if (cromScores) {
+        authorScores = cromScores;
+    } else {
+        const pageRatings = new Map();
+        for (const pg of (pageScoreRows || [])) pageRatings.set(normalizePageKey(pg.page), { rating: pg.rating });
+        authorScores = aggregateAuthorScores(attrRecords, pageRatings);
+    }
     await prisma.setting.upsert({
         where: { key: `author_score:${wikiParam}` },
         update: { value: JSON.stringify(authorScores) },
         create: { key: `author_score:${wikiParam}`, value: JSON.stringify(authorScores) }
     });
 
-    logLine(`[归属] ${wikiParam} 归属 ${attrRecords.length} 条 | 页面评分 ${allPages.length} | 作者 ${Object.keys(authorScores).length} 位`);
+    logLine(`[归属] ${wikiParam} 归属 ${attrRecords.length} 条 | 页面评分 ${pageScoreRows.length}（来源: ${pagesSource}） | 作者 ${Object.keys(authorScores).length} 位（来源: ${scoreSource}）`);
 }
 
 let isRunning = false;

diff --git a/auto-crawler.js b/auto-crawler.js
@@ -10,8 +10,8 @@ const { buildBackupArgs, getWikiName } = require('./utils/wikitBackup');
 const fs = require('fs');
 const path = require('path');
 
-const httpAgent = new http.Agent({ keepAlive: true, maxSockets: 10 });
-const httpsAgent = new https.Agent({ keepAlive: true, maxSockets: 10 });
+const httpAgent = new http.Agent({ keepAlive: true, maxSockets: 20 });
+const httpsAgent = new https.Agent({ keepAlive: true, maxSockets: 20 });
 const request = axios.create({
     httpAgent,
     httpsAgent,
@@ -20,6 +20,9 @@ const request = axios.create({
 
 const sleep = ms => new Promise(r => setTimeout(r, ms));
 
+// 跨页面/跨站点复用 UserLookup 结果，避免对同一作者反复查询 wikidot
+let userIdCache = {};
+
 const CRAWLER_LOG_FILE = path.join(process.cwd(), 'crawler.log');
 
 /** 同时输出到控制台和 crawler.log，方便管理后台查看 */
@@ -130,13 +133,22 @@ async function runCrawler() {
             finishedAt: null,
             currentSite: null,
             currentStage: '',
-            overall: { totalSites: config.SUPPORT_WIKI.length, doneSites: 0 },
+            overall: { totalSites: config.SUPPORT_WIKI.filter(s => !s.CROM_API).length, doneSites: 0 },
             sites: config.SUPPORT_WIKI.map(buildSiteStatus),
             lastRun: crawlStatus.lastRun
         };
         await persistCrawlStatus();
         for (const siteConfig of config.SUPPORT_WIKI) {
             const wikiParam = siteConfig.PARAM;
+
+            // crom 站点由 auto-attribution.js 独立处理（crom API 拿作者分数 + 页面评分），
+            // auto-crawler 逐页爬取对 8k+ 页面的站点效率极低且会与 attribution 争抢 DB 连接，
+            // 因此跳过已配置 CROM_API 的站点，仅保留非 crom 站点的讨论区/投票抓取
+            if (siteConfig.CROM_API) {
+                logLine(`[${new Date().toLocaleString()}] 跳过 ${wikiParam}（crom 站点，由 attribution 服务处理）`);
+                continue;
+            }
+
             const actualWikiName = siteConfig.URL.replace(/^https?:\/\//i, '').split('.')[0];
             const baseUrl = siteConfig.URL.replace(/\/$/, '');
 
@@ -186,7 +198,7 @@ async function runCrawler() {
                 } catch (e) {
                     await sleep(3000);
                 }
-                await sleep(1000);
+                await sleep(300);
             }
 
             if (siteStatus) {
@@ -197,7 +209,7 @@ async function runCrawler() {
 
             let userVotesMap = {};
             let count = 0;
-            const CONCURRENCY = 3;
+            const CONCURRENCY = 5;
 
             for (let i = 0; i < allPages.length; i += CONCURRENCY) {
                 const batch = allPages.slice(i, i + CONCURRENCY);
@@ -234,7 +246,6 @@ async function runCrawler() {
                                     const { data: forumHtml } = await request.get(forumUrl, { headers: baseHeaders });
                                     const $forum = cheerio.load(forumHtml);
                                     const posts = [];
-                                    const userIdCache = {};
 
                                     const postElements = $forum('.post').toArray();
                                     for (const el of postElements) {
@@ -276,7 +287,7 @@ async function runCrawler() {
                                                 userid = userIdCache[author];
                                             } else {
                                                 try {
-                                                    const lookupRes = await axios.get(`https://www.wikidot.com/quickmodule.php?module=UserLookupQModule&q=${encodeURIComponent(author)}`, { timeout: 5000 });
+                                                    const lookupRes = await axios.get(`https://www.wikidot.com/quickmodule.php?module=UserLookupQModule&q=${encodeURIComponent(author)}`, { timeout: 3000 });
                                                     if (lookupRes.data && lookupRes.data.users && lookupRes.data.users.length > 0) {
                                                         userid = lookupRes.data.users[0].user_id;
                                                         userIdCache[author] = userid;
@@ -390,7 +401,7 @@ async function runCrawler() {
                         siteStatus.errors = siteErrors;
                         await persistCrawlStatus();
                     }
-                    for (const [user, newVotes] of Object.entries(userVotesMap)) {
+                    await Promise.all(Object.entries(userVotesMap).map(async ([user, newVotes]) => {
                         const key = `user_votes_${user.toLowerCase().replace(/_/g, '-').replace(/ /g, '-')}`;
                         const record = await prisma.setting.findUnique({ where: { key } });
                         let existingMap = new Map();
@@ -417,10 +428,10 @@ async function runCrawler() {
                             update: { value: JSON.stringify(truncatedVotes) },
                             create: { key, value: JSON.stringify(truncatedVotes) }
                         });
-                    }
+                    }));
                     userVotesMap = {}; 
                 }
-                await sleep(2500);
+                await sleep(600);
             }
 
             if (siteStatus) {
@@ -459,7 +470,12 @@ async function runCrawler() {
 }
 
 cron.schedule('0 */3 * * *', () => runCrawler());
-runCrawler();
+// 启动时是否立即执行一轮全量抓取。
+// 默认开启（与原行为一致）；低内存服务器建议在 systemd 里设 RUN_CRAWLER_ON_START=false，
+// 避免「抓取 → OOM 被杀 → 重启 → 再抓取」的循环把整个服务器拖垮。
+if (String(process.env.RUN_CRAWLER_ON_START).toLowerCase() !== 'false') {
+    runCrawler();
+}
 
 let isBackupRunning = false;
 

diff --git a/components/AuthorActivityChart.js b/components/AuthorActivityChart.js
@@ -2,28 +2,95 @@ import React, { useEffect, useRef, useState } from 'react';
 // 核心修复：直接使用 auto 全自动注册，彻底解决漏引组件导致的致命闪退
 import Chart from 'chart.js/auto';
 
-export default function AuthorActivityChart({ data = [] }) {
+/**
+ * 把「单页面」数组（每篇一页，含 created_at + rating）按月聚合成
+ * 组件绘图所需的 { date: YYYY-MM, pages: 发布数, rating: 当月评分和 } 结构。
+ */
+function aggregateByMonth(pages) {
+  if (!Array.isArray(pages) || pages.length === 0) return [];
+  const bucket = new Map(); // YYYY-MM -> { pages, rating }
+  for (const p of pages) {
+    const rawDate = p && (p.created_at || p.createdAt || p.date);
+    let d;
+    if (!rawDate) continue;
+    // created_at 可能是 "YYYY-MM-DDTHH:mm:ss" 或 "YYYY-MM-DD"
+    const s = String(rawDate);
+    const first = s.split(/[T\s]/)[0]; // YYYY-MM-DD
+    const parts = first.split('-');
+    if (parts.length < 2) {
+      // 尝试 Date 兜底解析
+      d = new Date(s);
+      if (isNaN(d.getTime())) continue;
+      parts = [d.getFullYear(), d.getMonth() + 1];
+    }
+    const yyyy = Number(parts[0]);
+    const mm = Number(parts[1]);
+    if (!yyyy || !mm) continue;
+    const key = `${yyyy}-${String(mm).padStart(2, '0')}`;
+    if (!bucket.has(key)) bucket.set(key, { pages: 0, rating: 0 });
+    const b = bucket.get(key);
+    b.pages += 1;
+    const r = Number(p && p.rating != null ? p.rating : 0) || 0;
+    b.rating += r;
+  }
+  const arr = Array.from(bucket.entries()).map(([date, v]) => ({
+    date,
+    pages: v.pages,
+    rating: Math.round(v.rating * 100) / 100
+  }));
+  arr.sort((a, b) => String(a.date).localeCompare(String(b.date)));
+  return arr;
+}
+
+export default function AuthorActivityChart({ pages = [], data /* 兼容老调用 */ }) {
   const canvasRef = useRef(null);
   const chartInstance = useRef(null);
   const [chartError, setChartError] = useState(null);
 
+  // 优先使用 pages（调用方 pages/authors.js 传 pages={data.pages}），
+  // 若外部传了已聚合好的 data 则直接用（向前兼容）
+  const rawInput = Array.isArray(pages) && pages.length > 0 ? pages : (data || []);
+
   useEffect(() => {
     setChartError(null);
-    if (!canvasRef.current || !data || data.length === 0) return;
+    if (!canvasRef.current) return;
+    if (!rawInput || rawInput.length === 0) return;
 
     if (chartInstance.current) {
       chartInstance.current.destroy();
+      chartInstance.current = null;
     }
 
     try {
-      // 格式必须为 { date: "YYYY-MM", pages: num, rating: num }
-      const sortedData = [...data].sort((a, b) => String(a.date).localeCompare(String(b.date)));
-
+      // 先判断输入是否已经是「按月聚合」的 {date, pages, rating}，
+      // 还是「单页列表」（需要聚合）
+      const firstItem = rawInput[0] || {};
+      const alreadyAggregated =
+        typeof firstItem === 'object' &&
+        firstItem !== null &&
+        typeof (firstItem.date || firstItem.month) === 'string' &&
+        /\d{4}-\d{2}/.test(firstItem.date || firstItem.month) &&
+        ('pages' in firstItem || 'pageCount' in firstItem || 'rating' in firstItem);
+
+      // 格式要求：{ date: "YYYY-MM", pages: num, rating: num }
+      const monthly = alreadyAggregated
+        ? rawInput.map(x => ({
+            date: String(x.date || x.month).slice(0, 7),
+            pages: Number(x.pages != null ? x.pages : x.pageCount) || 0,
+            rating: Number(x.rating != null ? x.rating : 0) || 0
+          }))
+        : aggregateByMonth(rawInput);
+
+      if (!monthly.length) return;
+
+      const sortedData = [...monthly].sort((a, b) => String(a.date).localeCompare(String(b.date)));
+
       const minDateStr = sortedData[0].date;
       const maxDateStr = sortedData[sortedData.length - 1].date;
 
-      let [minY, minM] = minDateStr.split('-').map(Number);
-      let [maxY, maxM] = maxDateStr.split('-').map(Number);
+      let [minY, minM] = (minDateStr || '').split('-').map(Number);
+      let [maxY, maxM] = (maxDateStr || '').split('-').map(Number);
+      if (!minY || !minM || !maxY || !maxM) return;
 
       const labels = [];
       const pagesData = [];
@@ -103,7 +170,13 @@ export default function AuthorActivityChart({ data = [] }) {
             y: {
               beginAtZero: true,
               grid: { color: 'rgba(255, 255, 255, 0.05)' },
-              ticks: { color: 'rgb(110, 118, 129)', stepSize: 1 }
+              ticks: {
+                color: 'rgb(110, 118, 129)',
+                stepSize: (() => {
+                  const maxP = Math.max(1, ...pagesData);
+                  return maxP <= 10 ? 1 : undefined;
+                })()
+              }
             }
           }
         }
@@ -113,7 +186,7 @@ export default function AuthorActivityChart({ data = [] }) {
       setChartError(err.message);
     }
 
-  }, [data]);
+  }, [rawInput]);
 
   return (
     <div className="w-full h-full relative min-h-[260px]">

diff --git a/content/about.md b/content/about.md
@@ -58,6 +58,7 @@ WikitDB 是一个 **面向 Wikidot 社区生态的非营利性同人项目**，
 | 世界机构联合体 | pin | 世界观共创 |
 | SCP 基金会 Minecraft 分部 | scp-wiki-mc | SCP + MC 融合 |
 | The Backrooms 中文维基 | brcn | Backrooms 主分社区 |
+| The Nationarea | na | 架空国家地区世界观创作 |
 
 > 如需申请收录更多站点，请通过社区渠道联系管理员。