feat: govern knowledge assets and source citations

2026-08-02 01:43:43 +08:00
parent cafb836cda
commit 699cc08050
144 changed files with 17205 additions and 453 deletions
@@ -0,0 +1,77 @@
+import test from 'node:test';
+import assert from 'node:assert/strict';
+import fs from 'node:fs/promises';
+import os from 'node:os';
+import path from 'node:path';
+import { parseArgs, run } from './stage-legacy-knowledge.mjs';
+
+test('dry-run is the default and does not call the mutation endpoint', async () => {
+  const directory = await fs.mkdtemp(path.join(os.tmpdir(), 'aihr-legacy-dry-'));
+  const options = parseArgs(['--manifest', path.join(directory, 'manifest.json')]);
+  const result = await run(options, { fetchImpl: async () => assert.fail('fetch must not be called') });
+  assert.equal(result.mode, 'DRY_RUN');
+  assert.equal(result.preview, null);
+  await assert.rejects(fs.access(options.manifest));
+});
+
+test('authenticated dry-run calls only the read-only preview endpoint', async () => {
+  const directory = await fs.mkdtemp(path.join(os.tmpdir(), 'aihr-legacy-preview-'));
+  const options = parseArgs([
+    '--token', 'preview-secret', '--manifest', path.join(directory, 'manifest.json'), '--batch-size', '20'
+  ]);
+  let requestedUrl;
+  const result = await run(options, { fetchImpl: async (url, request) => {
+    requestedUrl = String(url);
+    assert.equal(request.method, 'GET');
+    return { ok: true, status: 200, json: async () => ({ code: 200, data: {
+      discovered: 20, sourceAvailable: 18, sourceUnavailable: 2, nextCursor: 42, moreAvailable: true
+    } }) };
+  } });
+  assert.match(requestedUrl, /\/api\/knowledge\/quality\/legacy\/preview/);
+  assert.equal(result.preview.sourceUnavailable, 2);
+  await assert.rejects(fs.access(options.manifest));
+});
+
+test('execute writes a resumable manifest without credentials', async () => {
+  const directory = await fs.mkdtemp(path.join(os.tmpdir(), 'aihr-legacy-run-'));
+  const options = parseArgs([
+    '--execute', '--token', 'top-secret', '--manifest', path.join(directory, 'manifest.json'),
+    '--max-batches', '2', '--batch-size', '20'
+  ]);
+  const replies = [
+    { discovered: 20, staged: 20, reparsed: 18, quarantined: 2, stagedAssetIds: [11, 12],
+      failedAttachmentIds: [], nextCursor: 100, moreAvailable: true },
+    { discovered: 3, staged: 3, reparsed: 2, quarantined: 1, stagedAssetIds: [13],
+      failedAttachmentIds: [], nextCursor: 103, moreAvailable: false }
+  ];
+  let calls = 0;
+  const result = await run(options, {
+    fetchImpl: async (_url, request) => {
+      assert.match(request.headers.Authorization, /top-secret/);
+      return { ok: true, status: 200, json: async () => ({ code: 200, data: replies[calls++] }) };
+    },
+    now: () => new Date('2026-08-01T00:00:00Z')
+  });
+  const manifestText = await fs.readFile(options.manifest, 'utf8');
+  const manifest = JSON.parse(manifestText);
+  assert.equal(result.status, 'COMPLETE');
+  assert.equal(manifest.cursor, 103);
+  assert.equal(manifest.batches.length, 2);
+  assert.doesNotMatch(manifestText, /top-secret/);
+});
+
+test('failed attachment does not advance the recovery cursor', async () => {
+  const directory = await fs.mkdtemp(path.join(os.tmpdir(), 'aihr-legacy-fail-'));
+  const options = parseArgs([
+    '--execute', '--token', 'secret', '--manifest', path.join(directory, 'manifest.json'), '--batch-size', '20'
+  ]);
+  await assert.rejects(run(options, {
+    fetchImpl: async () => ({ ok: true, status: 200, json: async () => ({ code: 200, data: {
+      discovered: 2, staged: 1, reparsed: 1, quarantined: 0, stagedAssetIds: [11],
+      failedAttachmentIds: [42], nextCursor: 42, moreAvailable: true
+    } }) })
+  }), /cursor was not advanced/);
+  const manifest = JSON.parse(await fs.readFile(options.manifest, 'utf8'));
+  assert.equal(manifest.status, 'FAILED');
+  assert.equal(manifest.cursor, 0);
+});