Lumi/plugins/lumi_ai_web_search/tests/verify.js

const assert = require("assert");
const fs = require("fs");
const os = require("os");
const path = require("path");
const { SearchProvider } = require("../backend/provider_adapter");
const { formatResults } = require("../backend/result_formatter");
const { WebSearchTool } = require("../backend/search_tool");
const { readSettings, writeSettings } = require("../backend/settings");
const { evaluateUrl, matchesRule } = require("../backend/url_policy");
const { ToolRegistry } = require("../../lumi_ai/backend/tool_router");
const { ToolInstaller } = require("../../lumi_ai/backend/tool_installer");
const { ToolLoader } = require("../../lumi_ai/backend/tool_loader");
const plugin = require("../index");

const PUBLIC_DNS = async () => ["93.184.216.34"];

async function run() {
  await verifyPolicy();
  await verifyRedirectPolicy();
  verifyFormatting();
  await verifySearchFlow();
  await verifyLoaderLifecycle();
  verifyRegistrationAvailability();
  verifyStaticFiles();
  console.log("Lumi AI Web Search verification passed.");
}

async function verifyLoaderLifecycle() {
  const root = fs.mkdtempSync(path.join(os.tmpdir(), "lumi-web-loader-"));
  const pluginsDir = path.join(root, "plugins");
  const toolDir = path.join(pluginsDir, "lumi_ai_web_search");
  copyDirectory(path.resolve(__dirname, ".."), toolDir, new Set(["data"]));
  fs.mkdirSync(path.join(toolDir, "data"), { recursive: true });
  const installer = new ToolInstaller({
    pluginsDir,
    stagingRoot: path.join(root, "staging"),
    repoClient: {}
  });
  const registry = new ToolRegistry(() => {});
  const loader = new ToolLoader({
    registry,
    installer,
    settings: { getSetting: (_key, fallback) => fallback },
    stateFile: path.join(root, "enabled.json"),
    lumiAiVersion: "0.8.0",
    lumiVersion: "0.1.0"
  });
  const unavailable = await loader.enable("lumi_ai_web_search");
  assert.equal(unavailable.unavailable, true);
  assert.equal(registry.tools.has("lumi_ai_web_search.search"), false);
  writeSettings(path.join(toolDir, "data"), providerSettings());
  const enabled = await loader.enable("lumi_ai_web_search");
  assert.equal(enabled.loaded, true);
  assert.equal(registry.tools.has("lumi_ai_web_search.search"), true);
  await loader.disable("lumi_ai_web_search");
  assert.equal(registry.tools.has("lumi_ai_web_search.search"), false);
  fs.rmSync(root, { recursive: true, force: true });
}

async function verifyPolicy() {
  let result = await evaluateUrl("https://docs.example.com/guide", {
    mode: "whitelist",
    rules: ["*.example.com/*"],
    resolveHost: PUBLIC_DNS
  });
  assert.equal(result.allowed, true);
  result = await evaluateUrl("https://unrelated.test/guide", {
    mode: "whitelist",
    rules: ["*.example.com/*"],
    resolveHost: PUBLIC_DNS
  });
  assert.equal(result.allowed, false);
  assert.equal(result.reason, "not_whitelisted");

  result = await evaluateUrl("https://ads.example.com/tracker", {
    mode: "blacklist",
    rules: ["*.example.com/tracker*"],
    resolveHost: PUBLIC_DNS
  });
  assert.equal(result.allowed, false);
  result = await evaluateUrl("https://docs.example.org/", {
    mode: "blacklist",
    rules: ["*.example.com/tracker*"],
    resolveHost: PUBLIC_DNS
  });
  assert.equal(result.allowed, true);

  for (const target of [
    "http://127.0.0.1/",
    "http://10.1.2.3/",
    "http://169.254.169.254/latest/meta-data/",
    "http://localhost/",
    "file:///etc/passwd"
  ]) {
    result = await evaluateUrl(target, {
      mode: "blacklist",
      rules: [],
      resolveHost: PUBLIC_DNS
    });
    assert.equal(result.allowed, false, target);
  }
  result = await evaluateUrl("https://dns-rebind.example/", {
    mode: "blacklist",
    rules: [],
    resolveHost: async () => ["10.0.0.8"]
  });
  assert.equal(result.allowed, false);
  assert.equal(matchesRule(new URL("https://docs.example.com/guide/start"), "example.com/guide"), true);
  assert.equal(matchesRule(new URL("https://example.com/"), "http://example.com/"), false);
}

async function verifyRedirectPolicy() {
  const provider = new SearchProvider({
    resolveHost: PUBLIC_DNS,
    fetch: async () => response({
      status: 302,
      headers: { location: "http://127.0.0.1/private" }
    })
  });
  await assert.rejects(
    () => provider.search("test", providerSettings()),
    /blocked by policy/i
  );
  let calls = 0;
  const crossOrigin = new SearchProvider({
    resolveHost: PUBLIC_DNS,
    fetch: async () => {
      calls += 1;
      return response({
        status: 302,
        headers: { location: "https://other-provider.example/search" }
      });
    }
  });
  await assert.rejects(
    () => crossOrigin.search("test", { ...providerSettings(), provider_api_key: "secret" }),
    /cross_origin_provider_redirect/i
  );
  assert.equal(calls, 1);
}

function verifyFormatting() {
  const rows = [
    result("Official docs", "https://docs.example.com/guide", "A detailed official answer for the requested subject.", "documentation"),
    result("Community post", "https://community.example.com/post", "A secondary explanation with useful context.", "web"),
    result("Recent update", "https://news.example.com/update", "A recently published update.", "news", "2026-06-12")
  ];
  const settings = {
    max_results: 5,
    show_source_links: true,
    twitch_output_chars: 180,
    discord_output_chars: 700,
    webui_output_chars: 3000,
    other_output_chars: 500
  };
  const fact = formatResults(rows, { reason: "fact_lookup", origin: "twitch", settings });
  const resource = formatResults(rows, { reason: "resource_lookup", origin: "discord", settings });
  const webui = formatResults(rows, { reason: "documentation_lookup", origin: "webui", settings });
  assert(fact.condensed_text.length <= 180);
  assert(fact.results.length <= 2);
  assert(resource.condensed_text.length <= 700);
  assert(webui.condensed_text.length > fact.condensed_text.length);
  assert.equal(webui.results[0].source_type, "documentation");
}

async function verifySearchFlow() {
  const root = fs.mkdtempSync(path.join(os.tmpdir(), "lumi-web-search-"));
  const settings = {
    ...providerSettings(),
    enabled: true,
    policy_mode: "whitelist",
    url_rules: ["*.example.com/*"],
    allowed_origins: ["webui", "discord", "twitch"],
    cache_ttl_seconds: 60
  };
  writeSettings(root, settings);
  let calls = 0;
  const provider = {
    resolveHost: PUBLIC_DNS,
    async search() {
      calls += 1;
      return [
        result("<b>Verified fact</b>", "https://docs.example.com/fact", "The <em>answer</em> is current.", "documentation"),
        result("Blocked local", "http://127.0.0.1/private", "Must never be returned.", "web")
      ];
    },
    async fetchPage() {
      return { url: "https://docs.example.com/fact", text: "Expanded public page text." };
    }
  };
  const tool = new WebSearchTool({ dataDir: root, provider });
  const first = await tool.run({
    query: "current fact",
    reason: "fact_lookup",
    user: { id: "user-1" },
    ctx: { origin: "webui", server_id: "server-1" }
  });
  assert.equal(first.status, "ok");
  assert.equal(first.result_count, 1);
  assert.equal(first.results[0].title, "Verified fact");
  assert.equal(first.results.some((entry) => entry.url?.includes("127.0.0.1")), false);
  const cached = await tool.run({
    query: "current fact",
    reason: "fact_lookup",
    user: { id: "user-1" },
    ctx: { origin: "webui", server_id: "server-1" }
  });
  assert.equal(cached.cache_hit, true);
  assert.equal(calls, 1);

  const twitch = await tool.run({
    query: "current fact twitch",
    reason: "resource_lookup",
    user: { id: "user-1" },
    ctx: { origin: "twitch", channel_id: "channel-1" }
  });
  assert(twitch.condensed_text.length <= readSettings(root).twitch_output_chars);
  assert.equal(calls, 2);

  writeSettings(root, { ...readSettings(root), allowed_origins: ["webui"] });
  const blockedOrigin = await tool.run({
    query: "current fact",
    reason: "fact_lookup",
    origin: "webui",
    user: { id: "user-1" },
    ctx: { origin: "discord" }
  });
  assert.equal(blockedOrigin.status, "blocked");
  assert.equal(blockedOrigin.blocked_reason, "origin_not_allowed");

  const failing = new WebSearchTool({
    dataDir: root,
    provider: {
      resolveHost: PUBLIC_DNS,
      async search() { throw new Error("provider secret https://provider.example/api?token=secret"); }
    }
  });
  const failed = await failing.run({
    query: "failure",
    reason: "general_lookup",
    user: { id: "user-2" },
    ctx: { origin: "webui" }
  });
  assert.equal(failed.status, "unavailable");
  assert.equal(failed.error.includes("provider.example"), false);

  const audit = fs.readFileSync(path.join(root, "audit.jsonl"), "utf8")
    .trim().split(/\r?\n/).map(JSON.parse);
  assert(audit.some((entry) =>
    entry.query === "current fact" &&
    entry.actor === "user-1" &&
    entry.origin === "webui" &&
    typeof entry.timing_ms === "number"
  ));
  fs.rmSync(root, { recursive: true, force: true });
}

function verifyRegistrationAvailability() {
  const root = fs.mkdtempSync(path.join(os.tmpdir(), "lumi-web-register-"));
  assert.equal(plugin.checkAvailability({ paths: { data: root } }).available, false);
  writeSettings(root, { ...providerSettings(), enabled: true });
  assert.equal(plugin.checkAvailability({ paths: { data: root } }).available, true);
  const definitions = [];
  plugin.register({
    paths: { data: root },
    registerTool: (definition) => definitions.push(definition)
  });
  assert.equal(definitions.length, 1);
  assert.equal(definitions[0].tool_id, "lumi_ai_web_search.search");
  assert.equal(definitions[0].read_only, true);
  assert.equal(definitions[0].origin_check({
    context: { origin: "discord", permission_context: { webui_actions_allowed: false } }
  }), true);
  assert.equal(definitions[0].permission_check({
    user: { id: "user" },
    context: { origin: "discord", permission_context: { webui_actions_allowed: false } }
  }), true);
  fs.rmSync(root, { recursive: true, force: true });
}

function verifyStaticFiles() {
  const root = path.resolve(__dirname, "..");
  const metadata = JSON.parse(fs.readFileSync(path.join(root, "tool_info.json"), "utf8"));
  assert.equal(metadata.tool_id, "lumi_ai_web_search");
  assert.equal(metadata.settings_schema.policy_mode.default, "whitelist");
  assert(fs.existsSync(path.join(root, "readme.md")));
  assert(fs.readFileSync(path.join(root, "views", "settings-modal.ejs"), "utf8").includes("settings_schema"));
}

function providerSettings() {
  return {
    ...Object.fromEntries(
      Object.entries(require("../tool_info.json").settings_schema).map(([key, field]) => [key, structuredClone(field.default)])
    ),
    provider_endpoint: "https://search.example.net/search",
    enabled: true,
    allowed_origins: ["webui", "discord", "twitch"],
    url_rules: ["*.example.com/*"]
  };
}

function result(title, url, snippet, sourceType, date = null) {
  return { title, url, snippet, source_type: sourceType, date, relevance_score: 0.9 };
}

function response({ status = 200, headers = {}, body = "" }) {
  const normalized = Object.fromEntries(Object.entries(headers).map(([key, value]) => [key.toLowerCase(), value]));
  return {
    ok: status >= 200 && status < 300,
    status,
    headers: {
      get(name) { return normalized[String(name).toLowerCase()] || null; }
    },
    async arrayBuffer() { return Buffer.from(body); }
  };
}

function copyDirectory(source, destination, ignored = new Set()) {
  fs.mkdirSync(destination, { recursive: true });
  for (const entry of fs.readdirSync(source, { withFileTypes: true })) {
    if (ignored.has(entry.name)) continue;
    const from = path.join(source, entry.name);
    const to = path.join(destination, entry.name);
    if (entry.isDirectory()) copyDirectory(from, to);
    else if (entry.isFile()) fs.copyFileSync(from, to);
  }
}

run().catch((error) => {
  console.error(error);
  process.exitCode = 1;
});