collabSupport/tests/markdownToAdf.test.js

// Unit tests for utils/markdownToAdf.js — uses Node's built-in test
// runner (`node --test tests/markdownToAdf.test.js`) so we don't need
// a separate test framework in package.json for this scope.

import test from 'node:test';
import assert from 'node:assert/strict';

import {
  markdownToAdf,
  markdownToAdfContent,
  tokenizeLine,
} from '../utils/markdownToAdf.js';

test('empty string produces an empty doc', () => {
  assert.deepEqual(markdownToAdf(''), { version: 1, type: 'doc', content: [] });
  assert.deepEqual(markdownToAdfContent(''), []);
});

test('whitespace-only input produces an empty doc', () => {
  assert.deepEqual(markdownToAdfContent('   \n\n   '), []);
});

test('non-string input is treated as empty', () => {
  assert.deepEqual(markdownToAdfContent(null), []);
  assert.deepEqual(markdownToAdfContent(undefined), []);
  assert.deepEqual(markdownToAdfContent(42), []);
});

test('plain single-paragraph text becomes one paragraph', () => {
  const out = markdownToAdfContent('hello world');
  assert.deepEqual(out, [
    { type: 'paragraph', content: [{ type: 'text', text: 'hello world' }] },
  ]);
});

test('bold: **text**', () => {
  assert.deepEqual(tokenizeLine('a **bold** b'), [
    { type: 'text', text: 'a ' },
    { type: 'text', text: 'bold', marks: [{ type: 'strong' }] },
    { type: 'text', text: ' b' },
  ]);
});

test('italic: *text*', () => {
  assert.deepEqual(tokenizeLine('a *italic* b'), [
    { type: 'text', text: 'a ' },
    { type: 'text', text: 'italic', marks: [{ type: 'em' }] },
    { type: 'text', text: ' b' },
  ]);
});

test('link: [label](url)', () => {
  assert.deepEqual(tokenizeLine('see [here](https://x.y)'), [
    { type: 'text', text: 'see ' },
    {
      type: 'text',
      text: 'here',
      marks: [{ type: 'link', attrs: { href: 'https://x.y' } }],
    },
  ]);
});

test('all three marks combined in one line', () => {
  const out = tokenizeLine('**bold** and *em* and [click](https://x.y) done');
  assert.deepEqual(out, [
    { type: 'text', text: 'bold', marks: [{ type: 'strong' }] },
    { type: 'text', text: ' and ' },
    { type: 'text', text: 'em', marks: [{ type: 'em' }] },
    { type: 'text', text: ' and ' },
    {
      type: 'text',
      text: 'click',
      marks: [{ type: 'link', attrs: { href: 'https://x.y' } }],
    },
    { type: 'text', text: ' done' },
  ]);
});

test('emoji at line start survives as plain text', () => {
  assert.deepEqual(tokenizeLine('✅ ok'), [
    { type: 'text', text: '✅ ok' },
  ]);
});

test('arrow → and bullet • characters are plain text', () => {
  assert.deepEqual(tokenizeLine(' → foo • bar'), [
    { type: 'text', text: ' → foo • bar' },
  ]);
});

test('blank line splits into two paragraphs', () => {
  const out = markdownToAdfContent('one\n\ntwo');
  assert.equal(out.length, 2);
  assert.equal(out[0].type, 'paragraph');
  assert.deepEqual(out[0].content, [{ type: 'text', text: 'one' }]);
  assert.deepEqual(out[1].content, [{ type: 'text', text: 'two' }]);
});

test('single newline becomes hardBreak inside one paragraph', () => {
  const out = markdownToAdfContent('line1\nline2');
  assert.equal(out.length, 1);
  assert.equal(out[0].type, 'paragraph');
  assert.deepEqual(out[0].content, [
    { type: 'text', text: 'line1' },
    { type: 'hardBreak' },
    { type: 'text', text: 'line2' },
  ]);
});

test('multiple blank lines collapse to a single paragraph split', () => {
  const out = markdownToAdfContent('a\n\n\n\nb');
  assert.equal(out.length, 2);
  assert.deepEqual(out[0].content, [{ type: 'text', text: 'a' }]);
  assert.deepEqual(out[1].content, [{ type: 'text', text: 'b' }]);
});

test('malformed markdown does not throw', () => {
  // Never throws is the only hard guarantee — the exact token shape
  // for ambiguous fragments is best-effort. The parser scans strong
  // before em, so an unclosed `**` frequently gets rescued as an em
  // starting at the second `*`. That's fine, just verify the concrete
  // behavior stays stable.
  assert.doesNotThrow(() => tokenizeLine('**unclosed and *also* here'));
  const out = tokenizeLine('**unclosed and *also* here');
  // Something got marked em — the middle `*...*` window matches.
  assert.ok(out.some((n) => n.marks?.[0]?.type === 'em'));
});

test('empty link body [](url) is treated as plain text', () => {
  // LINK_RE requires at least one char inside brackets; the empty case
  // won't match, so the raw text passes through unmarked.
  const out = tokenizeLine('a [](https://x.y) b');
  const rendered = out.map((n) => n.text).join('');
  assert.equal(rendered, 'a [](https://x.y) b');
  assert.ok(!out.some((n) => n.marks?.[0]?.type === 'link'));
});

test('*a**b*c does not throw and produces plausible tokens', () => {
  // Nasty edge case — verify we don't crash. Exact shape is not
  // contract, just "well-formed ADF inline nodes with no exception".
  assert.doesNotThrow(() => tokenizeLine('*a**b*c'));
});

test('markdownToAdf full doc wraps content correctly', () => {
  const doc = markdownToAdf('**hi**\n\n[go](https://x)');
  assert.equal(doc.version, 1);
  assert.equal(doc.type, 'doc');
  assert.equal(doc.content.length, 2);
  assert.deepEqual(doc.content[0].content, [
    { type: 'text', text: 'hi', marks: [{ type: 'strong' }] },
  ]);
  assert.deepEqual(doc.content[1].content, [
    {
      type: 'text',
      text: 'go',
      marks: [{ type: 'link', attrs: { href: 'https://x' } }],
    },
  ]);
});

test('adjacent plain-text runs collapse into one text node', () => {
  // Verifies the pushText coalescing behavior — cleaner ADF output.
  const out = tokenizeLine('abc def ghi');
  assert.equal(out.length, 1);
  assert.deepEqual(out[0], { type: 'text', text: 'abc def ghi' });
});