feat: complete quality/observability/KB port batch

2026-10-23 11:42:51 +08:00
parent b778570045
commit f343bf1832
12 changed files with 202 additions and 2 deletions
@@ -0,0 +1,42 @@
+#!/usr/bin/env node
+'use strict';
+
+// KB 质量审计：数量/父子结构、空答案、重复问题、按来源统计。
+const fs = require('fs');
+const path = require('path');
+const { Pool } = require('../bridge/node_modules/pg');
+
+const ROOT = path.resolve(__dirname, '..');
+const env = {};
+for (const line of fs.readFileSync(path.join(ROOT, '.env'), 'utf8').split('\n')) {
+  if (!line.includes('=') || line.trim().startsWith('#')) continue;
+  const i = line.indexOf('=');
+  env[line.slice(0, i).trim()] = line.slice(i + 1).trim();
+}
+const url = process.env.RAG_PG_URL || env.RAG_PG_URL;
+if (!url) throw new Error('RAG_PG_URL is required');
+
+(async () => {
+  const pool = new Pool({ connectionString: url, max: 1 });
+  try {
+    const summary = await pool.query(`
+      SELECT count(*)::int AS total,
+             count(*) FILTER (WHERE chunk_type='parent')::int AS parents,
+             count(*) FILTER (WHERE chunk_type='child')::int AS children,
+             count(*) FILTER (WHERE answer IS NULL OR answer='')::int AS empty_answers,
+             count(*) FILTER (WHERE text_tsv IS NULL)::int AS missing_tsv
+      FROM kb_docs`);
+    const dups = await pool.query(`
+      SELECT lower(regexp_replace(question, '[^a-zA-Z0-9\\u4e00-\\u9fff]+', ' ', 'g')) AS q, count(*)::int AS n
+      FROM kb_docs WHERE chunk_type='parent'
+      GROUP BY 1 HAVING count(*) > 1
+      ORDER BY n DESC LIMIT 10`);
+    const sources = await pool.query(`
+      SELECT source_file, count(*)::int AS n
+      FROM kb_docs WHERE chunk_type='parent'
+      GROUP BY source_file ORDER BY n DESC`);
+    console.log(JSON.stringify({ summary: summary.rows[0], duplicates: dups.rows, sources: sources.rows }, null, 2));
+  } finally {
+    await pool.end();
+  }
+})().catch((err) => { console.error('[kb-audit]', err.message); process.exit(1); });