新增内容判断引擎文章发布接口

2026-09-11 16:55:32 +08:00
parent dd34718c8c
commit dd0326e30e
12 changed files with 860 additions and 21 deletions
@@ -0,0 +1,189 @@
+import assert from 'node:assert/strict';
+import { after, before, test } from 'node:test';
+import express from 'express';
+import { ApiError, createContentIngestRouter, plainTextToParagraphHtml } from '../src/content-ingest.js';
+
+const TOKEN = 'test-content-ingest-token';
+const HASH_V1 = 'a'.repeat(64);
+const HASH_V2 = 'b'.repeat(64);
+let server;
+let baseUrl;
+let repository;
+
+class MemoryRepository {
+  constructor() {
+    this.articles = new Map();
+    this.requests = new Map();
+  }
+
+  async upsert(input, publicSiteUrl) {
+    const request = this.requests.get(input.idempotencyKey);
+    if (request) {
+      if (request.contentHash !== input.contentHash) {
+        throw new ApiError(409, 'CONTENT_CONFLICT', '相同Idempotency-Key对应了不同正文哈希');
+      }
+      return { created: false, result: { ...request.result } };
+    }
+    const identity = `${input.sourceSystem}:${input.sourceTaskId}:${input.siteProfile}`;
+    const existing = this.articles.get(identity);
+    const article = existing || { id: `art-${this.articles.size + 1}`, slug: input.requestedSlug, revision: 0 };
+    article.revision += 1;
+    article.title = input.title;
+    article.bodyPlain = input.bodyPlain;
+    article.contentHtml = input.contentHtml;
+    article.status = input.publishMode === 'publish' ? 'published' : 'draft';
+    article.contentHash = input.contentHash;
+    article.publishedAt = new Date().toISOString();
+    this.articles.set(identity, article);
+    const result = {
+      publication_id: article.id,
+      status: article.status,
+      url: `${publicSiteUrl}/insights/${article.slug}`,
+      revision: article.revision,
+      content_hash: article.contentHash,
+      published_at: article.publishedAt,
+    };
+    this.requests.set(input.idempotencyKey, { contentHash: input.contentHash, result });
+    return { created: !existing, result };
+  }
+}
+
+function payload(overrides = {}) {
+  return {
+    request_id: 'request-1',
+    source_system: 'amaba-content-core',
+    source_task_id: 'task-1',
+    source_version_id: 'version-1',
+    source_version_number: 1,
+    content_hash: HASH_V1,
+    site_profile: 'human',
+    publish_mode: 'publish',
+    article: {
+      title: '测试文章',
+      body_plain: '第一自然段。\n第二自然段。',
+      summary: '测试摘要',
+      slug: 'test-article',
+      primary_keyword: '企业进化',
+      keywords: ['企业进化'],
+      content_type: 'industry',
+      direction: '行业观察',
+      target_audience: '企业经营者',
+      author_name: '阿美巴',
+      cover_image_url: null,
+      canonical_url: null,
+    },
+    quality: { human_confirmed: true, confirmed_at: '2026-09-11T10:00:00+08:00', remaining_issue_levels: ['P3'] },
+    provenance: { knowledge_refs: [] },
+    semantic: { entities: [], questions: [], facts: [], suggested_schema_type: 'Article' },
+    ...overrides,
+  };
+}
+
+async function request(path, { token = TOKEN, body, key = 'task-1:version-1:human', method = 'POST' } = {}) {
+  const headers = {};
+  if (token !== null) headers.Authorization = `Bearer ${token}`;
+  if (body !== undefined) headers['Content-Type'] = 'application/json';
+  if (key !== null) headers['Idempotency-Key'] = key;
+  return fetch(`${baseUrl}${path}`, { method, headers, body: body === undefined ? undefined : JSON.stringify(body) });
+}
+
+before(async () => {
+  repository = new MemoryRepository();
+  const app = express();
+  app.use(express.json({ limit: '5mb' }));
+  app.use('/api/internal/content/v1', createContentIngestRouter({ repository, token: TOKEN, publicSiteUrl: 'https://www.ameiba.cn', rateLimitMax: 100 }));
+  server = app.listen(0, '127.0.0.1');
+  await new Promise((resolve) => server.once('listening', resolve));
+  baseUrl = `http://127.0.0.1:${server.address().port}`;
+});
+
+after(() => new Promise((resolve) => server.close(resolve)));
+
+test('健康检查需要正确Token并返回human站点身份', async () => {
+  const unauthorized = await request('/api/internal/content/v1/health', { method: 'GET', token: null, key: null });
+  assert.equal(unauthorized.status, 401);
+  const response = await request('/api/internal/content/v1/health', { method: 'GET', key: null });
+  assert.equal(response.status, 200);
+  assert.deepEqual(await response.json(), { status: 'ok', service: 'amaba-human-website', site_profile: 'human', api_version: 'v1' });
+});
+
+test('未配置发布Token时接口返回503', async () => {
+  const app = express();
+  app.use(express.json());
+  app.use('/api/internal/content/v1', createContentIngestRouter({ repository: new MemoryRepository(), token: '' }));
+  const localServer = app.listen(0, '127.0.0.1');
+  await new Promise((resolve) => localServer.once('listening', resolve));
+  const response = await fetch(`http://127.0.0.1:${localServer.address().port}/api/internal/content/v1/health`);
+  assert.equal(response.status, 503);
+  await new Promise((resolve) => localServer.close(resolve));
+});
+
+test('错误Token返回401', async () => {
+  const response = await request('/api/internal/content/v1/articles/upsert', { token: 'wrong', body: payload() });
+  assert.equal(response.status, 401);
+});
+
+test('缺少标题返回400', async () => {
+  const data = payload();
+  data.article.title = '';
+  assert.equal((await request('/api/internal/content/v1/articles/upsert', { body: data })).status, 400);
+});
+
+test('缺少正文返回400', async () => {
+  const data = payload();
+  data.article.body_plain = '';
+  assert.equal((await request('/api/internal/content/v1/articles/upsert', { body: data })).status, 400);
+});
+
+test('未人工确认时拒绝发布', async () => {
+  const data = payload();
+  data.quality.human_confirmed = false;
+  assert.equal((await request('/api/internal/content/v1/articles/upsert', { body: data })).status, 403);
+});
+
+test('错误site_profile被拒绝', async () => {
+  const data = payload({ site_profile: 'ai' });
+  assert.equal((await request('/api/internal/content/v1/articles/upsert', { body: data })).status, 403);
+});
+
+test('首次发布返回201并生成公开详情URL', async () => {
+  const response = await request('/api/internal/content/v1/articles/upsert', { body: payload() });
+  const result = await response.json();
+  assert.equal(response.status, 201);
+  assert.equal(result.success, true);
+  assert.equal(result.revision, 1);
+  assert.equal(result.url, 'https://www.ameiba.cn/insights/test-article');
+  assert.equal(repository.articles.size, 1);
+});
+
+test('相同幂等键重复请求不创建重复文章', async () => {
+  const response = await request('/api/internal/content/v1/articles/upsert', { body: payload() });
+  assert.equal(response.status, 200);
+  assert.equal((await response.json()).publication_id, 'art-1');
+  assert.equal(repository.articles.size, 1);
+});
+
+test('相同幂等键与不同哈希返回409', async () => {
+  const response = await request('/api/internal/content/v1/articles/upsert', { body: payload({ content_hash: HASH_V2 }) });
+  assert.equal(response.status, 409);
+});
+
+test('新版本更新同一文章、保留URL并增加revision', async () => {
+  const data = payload({ source_version_id: 'version-2', source_version_number: 2, content_hash: HASH_V2 });
+  data.article.slug = 'should-not-replace-original-url';
+  data.article.body_plain = '新版本正文。';
+  const response = await request('/api/internal/content/v1/articles/upsert', { body: data, key: 'task-1:version-2:human' });
+  const result = await response.json();
+  assert.equal(response.status, 200);
+  assert.equal(result.publication_id, 'art-1');
+  assert.equal(result.url, 'https://www.ameiba.cn/insights/test-article');
+  assert.equal(result.revision, 2);
+  assert.equal(repository.articles.size, 1);
+});
+
+test('纯文本按自然段转义，HTML和脚本不会执行', () => {
+  assert.equal(
+    plainTextToParagraphHtml('第一段\n\n<script>alert(1)</script>\n<b>文字</b>'),
+    '<p>第一段</p>\n<p>&lt;script&gt;alert(1)&lt;/script&gt;</p>\n<p>&lt;b&gt;文字&lt;/b&gt;</p>',
+  );
+});