koc-loop/tests/account-tagging.test.mjs

import assert from "node:assert/strict";
import test from "node:test";
import {
  buildAccountTagText,
  generateAccountTagsFromMcp,
  parseAccountTags,
} from "../lib/mcp-collection-client.ts";
import { needsResourceEnrichment } from "../lib/resource-import.ts";

function sse(payload, headers = {}) {
  return new Response(`event: message\ndata: ${JSON.stringify(payload)}\n\n`, {
    status: 200,
    headers: {
      "content-type": "text/event-stream",
      ...headers,
    },
  });
}

function toolEnvelope(payload, isError = false) {
  return {
    jsonrpc: "2.0",
    id: "tool-call",
    result: {
      isError,
      content: [{ type: "text", text: JSON.stringify(payload) }],
    },
  };
}

function createFakeTaggingMcp(toolResult, isError = false) {
  const calls = [];
  const fetchImpl = async (url, init) => {
    const body = JSON.parse(init.body);
    calls.push({ url: String(url), body });
    if (body.method === "initialize") {
      return sse(
        {
          jsonrpc: "2.0",
          id: body.id,
          result: {
            protocolVersion: "2025-03-26",
            capabilities: {},
            serverInfo: { name: "ai-tool-center", version: "test" },
          },
        },
        { "mcp-session-id": "session-test" },
      );
    }
    if (body.method === "notifications/initialized") {
      return new Response("", { status: 202 });
    }
    if (body.method === "tools/call") {
      return sse(toolEnvelope(toolResult, isError));
    }
    throw new Error(`Unexpected MCP method: ${body.method}`);
  };
  return { fetchImpl, calls };
}

const config = { endpoint: "https://mcp.test/mcp", key: "test-key" };
const taggingEnv = { AI_TAGGING_TOOL: "generate_tags" };
const input = {
  platform: "小红书",
  nickname: "小美的日常",
  bio: "分享护肤和生活",
  gender: "女",
  ipLocation: "上海",
  followers: 1200,
  providerTags: ["美妆"],
  recentNoteTitles: ["平价护肤分享"],
};

test("parseAccountTags 解析纯 JSON 数组", () => {
  assert.deepEqual(parseAccountTags('["美妆","学生党"]'), ["美妆", "学生党"]);
});

test("parseAccountTags 剥离代码块围栏", () => {
  assert.deepEqual(parseAccountTags('```json\n["美食"]\n```'), ["美食"]);
  assert.deepEqual(parseAccountTags('```\n["旅行"]\n```'), ["旅行"]);
});

test("parseAccountTags 从夹杂文字中提取数组", () => {
  const text = '好的，以下是标签：["旅行","摄影"] 希望有帮助';
  assert.deepEqual(parseAccountTags(text), ["旅行", "摄影"]);
});

test("parseAccountTags 清洗与过滤", () => {
  assert.deepEqual(parseAccountTags('["#美食","探店#"]'), ["美食", "探店"]);
  assert.deepEqual(parseAccountTags('["美妆","美妆","穿搭"]'), [
    "美妆",
    "穿搭",
  ]);
  assert.deepEqual(
    parseAccountTags('["这个标签实在太长了超过十二个字","护肤"]'),
    ["护肤"],
  );
  assert.deepEqual(
    parseAccountTags('[1, null, "母婴", true]'),
    ["母婴"],
  );
});

test("parseAccountTags 超过 5 个截断", () => {
  const text = '["一","二","三","四","五","六"]';
  assert.deepEqual(parseAccountTags(text), ["一", "二", "三", "四", "五"]);
});

test("parseAccountTags 无法解析时返回空数组", () => {
  assert.deepEqual(parseAccountTags("无法生成标签"), []);
  assert.deepEqual(parseAccountTags(""), []);
  assert.deepEqual(parseAccountTags('{"tags":"美妆"}'), []);
});

test("buildAccountTagText 包含关键字段", () => {
  const text = buildAccountTagText(input);
  assert.ok(text.includes("小美的日常"));
  assert.ok(text.includes("分享护肤和生活"));
  assert.ok(text.includes("粉丝数：1200"));
  assert.ok(text.includes("美妆"));
  assert.ok(text.includes("平价护肤分享"));
});

test("buildAccountTagText 粉丝数为 0 或空时省略", () => {
  assert.ok(!buildAccountTagText({ ...input, followers: 0 }).includes("粉丝数"));
  assert.ok(
    !buildAccountTagText({ ...input, followers: null }).includes("粉丝数"),
  );
});

test("generateAccountTagsFromMcp 工具名为空时不发起请求", async () => {
  const fetchImpl = async () => {
    throw new Error("should not be called");
  };
  const tags = await generateAccountTagsFromMcp(
    input,
    config,
    {},
    fetchImpl,
  );
  assert.deepEqual(tags, []);
});

test("generateAccountTagsFromMcp 正常返回标签数组", async () => {
  const { fetchImpl, calls } = createFakeTaggingMcp(["美妆", "探店", "好物分享"]);
  const tags = await generateAccountTagsFromMcp(
    input,
    config,
    taggingEnv,
    fetchImpl,
  );
  assert.deepEqual(tags, ["美妆", "探店", "好物分享"]);
  const toolCall = calls.find((call) => call.body.method === "tools/call");
  assert.equal(toolCall.body.params.name, "generate_tags");
  const args = toolCall.body.params.arguments;
  assert.equal(args.objective, "为KOC账号生成运营标签");
  assert.ok(args.text.includes("小美的日常"));
  assert.ok(Array.isArray(args.rules) && args.rules.length > 0);
  assert.equal(args.output_schema.properties.tags.type, "array");
  assert.ok(String(toolCall.url).includes("key=test-key"));
});

test("generateAccountTagsFromMcp 支持对象包裹的返回", async () => {
  const { fetchImpl } = createFakeTaggingMcp({ tags: ["母婴", "育儿"] });
  const tags = await generateAccountTagsFromMcp(
    input,
    config,
    taggingEnv,
    fetchImpl,
  );
  assert.deepEqual(tags, ["母婴", "育儿"]);
});

test("generateAccountTagsFromMcp 工具报错时返回空数组", async () => {
  const { fetchImpl } = createFakeTaggingMcp({ message: "boom" }, true);
  const tags = await generateAccountTagsFromMcp(
    input,
    config,
    taggingEnv,
    fetchImpl,
  );
  assert.deepEqual(tags, []);
});

test("generateAccountTagsFromMcp 网络异常时返回空数组", async () => {
  const fetchImpl = async () => {
    throw new Error("network down");
  };
  const tags = await generateAccountTagsFromMcp(
    input,
    config,
    taggingEnv,
    fetchImpl,
  );
  assert.deepEqual(tags, []);
});

const completeRow = {
  nickname: "小美",
  publicAccountId: "xhs-123",
  ipLocation: "上海",
  followersResolved: true,
  gender: "女",
  bio: "分享生活",
  tags: ["美妆"],
};

test("needsResourceEnrichment 资料齐全且有标签时为 false", () => {
  assert.equal(needsResourceEnrichment(completeRow), false);
});

test("needsResourceEnrichment 资料齐全但无标签时为 true", () => {
  assert.equal(needsResourceEnrichment({ ...completeRow, tags: [] }), true);
});

test("needsResourceEnrichment 缺资料且有标签时为 true", () => {
  assert.equal(
    needsResourceEnrichment({ ...completeRow, bio: "" }),
    true,
  );
});