Hankan/test/domain/words.test.ts

/* Looking a word up: what its underline says, what adding it adds, and
   English → 한국어 — against the shipped dictionary. */

import { describe, it, expect, beforeAll, afterAll, vi } from "vitest";
import type { Db } from "@app/db/types.js";
import { editAddCustomWord, editCard } from "@app/db/writes.js";
import { wordInfo, wordToAdd } from "@app/domain/words.js";
import { buildEnIndex, enLookup, isLookupWord, loadEnglishIndex } from "@app/domain/english.js";
import { cardLemmaFor } from "@app/domain/evidence.js";
import { deck } from "@app/domain/cards.js";
import { SECURE_INTERVAL, newCard } from "@lib/srs.js";
import { dictionaryDb } from "../helpers/dict-db.js";

let db: Db;

beforeAll(async () => {
  db = await dictionaryDb(1);
});
afterAll(async () => {
  vi.unstubAllGlobals();
  await db.close();
});

describe("the underline", () => {
  it("marks a deck word he has not studied as new, and a secure one as known", async () => {
    let info = (await wordInfo(db, ["바다"])).get("바다")!;
    expect(info).toMatchObject({ state: "new", inDeck: true, head: "바다" });
    expect(info.gloss?.gloss).toBeTruthy();

    const id = (await cardLemmaFor(db, "바다"))!;
    await editCard(db, id, { ...newCard(), state: 2, interval: SECURE_INTERVAL, due: 99_999, reps: 6 });
    info = (await wordInfo(db, ["바다"])).get("바다")!;
    expect(info.state).toBe("known");
  });

  it("gives a form its dictionary word's state", async () => {
    const id = (await cardLemmaFor(db, "가다"))!;
    await editCard(db, id, { ...newCard(), state: 1, due: 0, reps: 1 });
    const info = (await wordInfo(db, ["갔어"])).get("갔어")!;
    expect(info.state).toBe("learning");
    expect(info.head).toBe("가다");
  });

  it("calls a word only the dictionary knows explainable, and nothing at all unknown", async () => {
    const row = await db.get<{ headword: string }>(
      `SELECT l.headword FROM lemma l
        WHERE l.source NOT IN ('curated','grammar','sfx','curriculum','custom','sentence')
          AND l.gloss_en <> '' AND l.headword NOT IN
            (SELECT headword FROM lemma WHERE source IN ('curated','grammar','sfx','curriculum','custom'))
        ORDER BY l.freq_rank LIMIT 1`,
    );
    const word = row!.headword;
    const infos = await wordInfo(db, [word, "뷁뷁"]);
    expect(infos.get(word)).toMatchObject({ state: "gloss", inDeck: false });
    expect(infos.get("뷁뷁")).toMatchObject({ state: "unknown", inDeck: false, gloss: null });
  });
});

describe("adding a looked-up word", () => {
  it("adds the dictionary word, and it then counts as his — in the deck and underlined new", async () => {
    const row = await db.get<{ headword: string }>(
      `SELECT l.headword FROM lemma l
        WHERE l.source NOT IN ('curated','grammar','sfx','curriculum','custom','sentence')
          AND l.gloss_en <> '' AND l.headword NOT IN
            (SELECT headword FROM lemma WHERE source IN ('curated','grammar','sfx','curriculum','custom'))
        ORDER BY l.freq_rank LIMIT 1 OFFSET 3`,
    );
    const word = row!.headword;
    const before = (await wordInfo(db, [word])).get(word)!;
    const add = wordToAdd(before);
    expect(add.headword).toBe(word);

    await editAddCustomWord(db, add);
    const after = (await wordInfo(db, [word])).get(word)!;
    expect(after).toMatchObject({ state: "new", inDeck: true });
    expect((await deck(db)).some((e) => e.headword === word)).toBe(true);
  });

  it("adds a form nothing knows as he met it", () => {
    expect(
      wordToAdd({ token: "뷁다", state: "unknown", gloss: null, head: null, inDeck: false, lemma: null }),
    ).toEqual({ headword: "뷁다", pos: "verb", gloss: "—" });
  });
});

describe("English → 한국어", () => {
  const index = buildEnIndex([
    { ko: "형", gloss: "older brother (said by a man)", note: "noun", pos: "noun", inDeck: true },
    { ko: "크다", gloss: "to be big", note: "adj", pos: "adj", inDeck: true },
    { ko: "단어", gloss: "word", note: "noun", pos: "noun", inDeck: true },
    { ko: "책상", gloss: "desk; table", note: "noun", pos: "noun", inDeck: true },
  ]);

  it("tries the longest phrase around the word first", () => {
    const words = "my older brother is tall".split(" ");
    expect(enLookup(index, words, 2).map((e) => e.ko)).toEqual(["형"]);
  });

  it("finds a word inside a gloss, a clause of one, and a plural made singular", () => {
    expect(enLookup(index, ["big"], 0).map((e) => e.ko)).toEqual(["크다"]);
    expect(enLookup(index, ["table"], 0).map((e) => e.ko)).toEqual(["책상"]);
    expect(enLookup(index, ["the", "words"], 1).map((e) => e.ko)).toEqual(["단어"]);
  });

  it("never offers the small words", () => {
    expect(isLookupWord("the")).toBe(false);
    expect(isLookupWord("a")).toBe(false);
    expect(isLookupWord("table")).toBe(true);
  });

  it("is built from the course's own material, never the frequency bands", async () => {
    const shipped = await loadEnglishIndex(db);
    expect(enLookup(shipped, ["water"], 0).map((e) => e.ko)).toContain("물");

    const course = new Set(
      (
        await db.all<{ headword: string }>(
          `SELECT headword FROM lemma
            WHERE source IN ('curated', 'grammar', 'sentence', 'sfx', 'curriculum', 'custom')`,
        )
      ).map((r) => r.headword),
    );
    const indexed = new Set([...shipped.values()].flat().map((e) => e.ko));
    expect(indexed.size).toBeGreaterThan(500);
    expect([...indexed].filter((ko) => !course.has(ko))).toEqual([]);
  });
});