feat(admin): add SEO/GEO publication catalog with crawler stats

Expose a DB-backed catalog of all online publications with SEO, GEO, total, and bot view counts from h5_publication_views for the memind_adm dashboard. Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-12 16:41:39 +08:00
parent 6542a43495
commit fdc234e5d0
3 changed files with 372 additions and 0 deletions
@@ -0,0 +1,113 @@
+import assert from 'node:assert/strict';
+import test from 'node:test';
+
+import { classifyReferrerHost } from './mindspace-analytics-discovery.mjs';
+import {
+  listAdminSeoGeoPublicationCatalog,
+  resolveAdminCatalogPublicUrl,
+} from './mindspace-seo-geo-admin-catalog.mjs';
+
+test('classifyReferrerHost maps search and ai referrers', () => {
+  assert.deepEqual(classifyReferrerHost('www.google.com'), {
+    discovery_channel: 'seo',
+    discovery_source: 'google',
+  });
+  assert.deepEqual(classifyReferrerHost('chatgpt.com'), {
+    discovery_channel: 'geo',
+    discovery_source: 'chatgpt',
+  });
+});
+
+test('resolveAdminCatalogPublicUrl normalizes relative public urls', () => {
+  assert.equal(
+    resolveAdminCatalogPublicUrl('/u/john/pages/demo'),
+    'https://m.tkmind.cn/u/john/pages/demo',
+  );
+});
+
+test('listAdminSeoGeoPublicationCatalog merges zero-traffic pages with view stats', async () => {
+  const calls = [];
+  const pool = {
+    async query(sql, params = []) {
+      calls.push({ sql, params });
+      const normalized = String(sql).replace(/\s+/g, ' ').trim();
+      if (normalized.startsWith('SELECT COUNT(*) AS total')) {
+        return [[{ total: 2 }]];
+      }
+      if (normalized.includes('FROM h5_publish_records pr') && normalized.includes('JOIN h5_users')) {
+        return [[
+          {
+            publicationId: 'pub-1',
+            pageId: 'page-1',
+            userId: 'user-1',
+            publicUrl: '/u/john/pages/demo',
+            accessMode: 'public',
+            status: 'online',
+            userConfirmedAt: 1000,
+            expiresAt: null,
+            publishedAt: 2000,
+            lifetimeViewCount: 12,
+            pageTitle: 'Demo',
+            pageSummary: 'Summary',
+            username: 'john',
+            displayName: 'John',
+          },
+          {
+            publicationId: 'pub-2',
+            pageId: 'page-2',
+            userId: 'user-2',
+            publicUrl: '/u/jane/pages/empty',
+            accessMode: 'public',
+            status: 'online',
+            userConfirmedAt: null,
+            expiresAt: null,
+            publishedAt: 1500,
+            lifetimeViewCount: 0,
+            pageTitle: 'Empty',
+            pageSummary: '',
+            username: 'jane',
+            displayName: 'Jane',
+          },
+        ]];
+      }
+      if (normalized.includes('FROM h5_publication_views')) {
+        return [[
+          {
+            publicationId: 'pub-1',
+            device_type: 'desktop',
+            referrer_host: 'www.google.com',
+            views: 3,
+          },
+          {
+            publicationId: 'pub-1',
+            device_type: 'bot',
+            referrer_host: null,
+            views: 2,
+          },
+        ]];
+      }
+      return [[]];
+    },
+  };
+
+  const result = await listAdminSeoGeoPublicationCatalog(pool, {
+    page: 1,
+    pageSize: 10,
+    startAt: 1,
+    endAt: 9_999_999_999_999,
+    sortBy: 'totalViews',
+    sortOrder: 'desc',
+  });
+
+  assert.equal(result.total, 2);
+  assert.equal(result.data.length, 2);
+  assert.equal(result.data[0].publicationId, 'pub-1');
+  assert.equal(result.data[0].seoViews, 3);
+  assert.equal(result.data[0].crawlerViews, 2);
+  assert.equal(result.data[0].totalViews, 5);
+  assert.equal(result.data[0].indexable, true);
+  assert.equal(result.data[1].publicationId, 'pub-2');
+  assert.equal(result.data[1].totalViews, 0);
+  assert.equal(result.data[1].indexable, false);
+  assert.match(result.data[1].publicUrl, /^https:\/\/m\.tkmind\.cn\//);
+});