import { mkdirSync, mkdtempSync, rmSync, writeFileSync } from "node:fs";
import { tmpdir } from "node:os";
import path from "node:path";
import { describe, it, expect, afterAll } from "vitest";
import { discoverSourceLists, normalizeWords } from "../sourceLists.js";

describe("normalizeWords", () => {
  it("trims whitespace and drops empty lines", () => {
    expect(normalizeWords(["  Haus ", "", "   ", "Tür"])).toEqual([
      "Haus",
      "Tür",
    ]);
  });

  it("dedups while preserving first-occurrence order", () => {
    expect(normalizeWords(["centro", "borde", "centro", "suelo"])).toEqual([
      "centro",
      "borde",
      "suelo",
    ]);
  });
});

describe("discoverSourceLists", () => {
  const root = mkdtempSync(path.join(tmpdir(), "lila-source-data-"));
  afterAll(() => rmSync(root, { recursive: true, force: true }));

  it("finds only supported language/pos paths and normalizes their words", () => {
    mkdirSync(path.join(root, "de"));
    mkdirSync(path.join(root, "es"));
    mkdirSync(path.join(root, "german")); // unsupported dir name — ignored
    writeFileSync(
      path.join(root, "de", "noun"),
      "Haus\nTür\nHaus\n\n  Tisch \n",
    );
    writeFileSync(path.join(root, "es", "noun"), "casa\npared\n");
    writeFileSync(path.join(root, "es", "nouns"), "ignored\n"); // unsupported pos name
    writeFileSync(path.join(root, "german", "noun"), "ignored\n");
    writeFileSync(path.join(root, "stray.txt"), "ignored\n");

    const lists = discoverSourceLists(root);
    expect(lists).toHaveLength(2);
    expect(lists[0]).toMatchObject({
      sourceLanguage: "de",
      pos: "noun",
      words: ["Haus", "Tür", "Tisch"],
    });
    expect(lists[1]).toMatchObject({
      sourceLanguage: "es",
      pos: "noun",
      words: ["casa", "pared"],
    });
  });
});