embed-markdown/index.test.js

const test = require("node:test");
const assert = require("node:assert/strict");

const {
  buildRequestPayload,
  buildStoreText,
  getConfig,
  main,
  sanitizeDocumentId,
} = require("./index");

test("getConfig defaults to the ingress API URL", () => {
  const config = getConfig({
    coreModule: {
      getInput() {
        return "";
      },
    },
    githubModule: {
      context: {
        serverUrl: "https://github.example",
      },
    },
    env: {},
  });

  assert.equal(config.apiUrl, "http://agent-api.k8s.private");
  assert.equal(config.route, "agent-lake");
  assert.equal(config.method, "embeddings_insert");
});

test("sanitizeDocumentId removes special characters and preserves periods", () => {
  assert.equal(
    sanitizeDocumentId("acme.docs-repo.guides/setup v2!.md"),
    "acme.docsrepo.guidessetupv2.md",
  );
});

test("buildStoreText prefixes repository metadata before the markdown content", () => {
  const content = "# Title\n\n## First\nalpha\n\n## Second\nbeta\n";
  const storeText = buildStoreText({
    owner: "acme",
    repo: "docs-repo",
    file: "guides/setup.md",
    content,
  });

  assert.equal(
    storeText,
    "repo_name: docs-repo\nrepo_owner: acme\npath: guides/setup.md\n\n# Title\n\n## First\nalpha\n\n## Second\nbeta\n",
  );
});

test("buildRequestPayload sends metadata-prefixed store_text and an empty embed_text", () => {
  const content = "# Title\n\n## First\nalpha\n\n## Second\nbeta\n";
  const payload = buildRequestPayload({
    route: "agent-lake",
    method: "embeddings_insert",
    segmentId: "docs.acme",
    owner: "acme",
    repo: "docs-repo",
    file: "guides/setup.md",
    content,
  });

  assert.equal(payload.inputs.inputs.embed_text, "");
  assert.equal(
    payload.inputs.inputs.store_text,
    "repo_name: docs-repo\nrepo_owner: acme\npath: guides/setup.md\n\n# Title\n\n## First\nalpha\n\n## Second\nbeta\n",
  );
  assert.equal(
    payload.inputs.inputs.document_id,
    "acme.docsrepo.guidessetup.md",
  );
});

test("main sends one request per file even when the document contains multiple sections", async () => {
  const calls = [];
  const infos = [];
  const failures = [];
  const content = "# Title\n\n## First\nalpha\n\n## Second\nbeta\n";

  await main({
    coreModule: {
      getInput(name) {
        const inputs = {
          api_url: "https://agents.example/api",
          api_token: "",
          route: "agent-lake",
          method: "embeddings_insert",
          debug: "false",
        };
        return inputs[name] || "";
      },
      info(message) {
        infos.push(message);
      },
      setFailed(message) {
        failures.push(message);
      },
    },
    githubModule: {
      context: {
        serverUrl: "https://github.example",
      },
    },
    env: {
      GITHUB_REPOSITORY: "acme/docs-repo",
      GITHUB_SERVER_URL: "https://github.example",
    },
    executor() {
      return "README.md\n";
    },
    fsModule: {
      readFileSync(file, encoding) {
        assert.equal(file, "README.md");
        assert.equal(encoding, "utf8");
        return content;
      },
    },
    fetchFn: async (url, options) => {
      calls.push({
        url,
        options,
        body: JSON.parse(options.body),
      });

      return {
        ok: true,
        text: async () => "ok",
      };
    },
    waitFn: async () => {},
  });

  assert.equal(calls.length, 1);
  assert.equal(calls[0].url, "https://agents.example/api");
  assert.equal(calls[0].body.inputs.inputs.embed_text, "");
  assert.equal(
    calls[0].body.inputs.inputs.store_text,
    "repo_name: docs-repo\nrepo_owner: acme\npath: README.md\n\n# Title\n\n## First\nalpha\n\n## Second\nbeta\n",
  );
  assert.equal(
    calls[0].body.inputs.inputs.document_id,
    "acme.docsrepo.README.md",
  );
  assert.deepEqual(failures, []);
  assert.deepEqual(infos, ["Agent response: ok"]);
});