pediatric-ai-scribe-v3/test/citation-ordering.test.js

// Sources arrived in retrieval order — an order the reader never sees and
// cannot follow: an answer whose first citation was [7] opened a list that
// began at [1]. Reference lists are numbered by first appearance for exactly
// this reason.
//
// Ordering is derived from the same token walk that renders the chips, so the
// list and the numbers on screen cannot disagree. The text is never rewritten:
// `[7]` stays `[7]`, because that number is the source's identity in the saved
// chat, the export and every chip. Two earlier versions rewrote the text and
// both were wrong — one hit array indexing in code, the other an HTML
// attribute — and a third would have hit the next literal context.
const test = require('node:test');
const assert = require('node:assert/strict');
const MarkdownIt = require('markdown-it');
globalThis.markdownit = MarkdownIt;

let mod;
async function load() {
  if (!mod) {
    const fs = require('node:fs'); const path = require('node:path');
    const src = fs.readFileSync(path.join(__dirname, '..', 'public/js/assistant/citations.js'), 'utf8');
    mod = await import('data:text/javascript;charset=utf-8,' + encodeURIComponent(src));
  }
  return mod;
}
const four = [{ number: 1, title: 'Alpha' }, { number: 2, title: 'Beta' }, { number: 3, title: 'Gamma' }, { number: 4, title: 'Delta' }];
const titles = out => Array.from(out.sources, s => s.title);

test('sources come back in the order the answer cites them', async () => {
  const { orderSourcesByCitation } = await load();
  const out = orderSourcesByCitation('Third first [3]. Then the first [1].', four);
  assert.deepEqual(titles(out).slice(0, 2), ['Gamma', 'Alpha']);
  assert.deepEqual(Array.from(out.sources.slice(0, 2), s => [s.number, s.sourceNumber]), [[1, 3], [2, 1]]);
});

test('the text is returned exactly as given', async () => {
  const { orderSourcesByCitation } = await load();
  const text = 'Third first [3]. Then the first [1].';
  assert.equal(orderSourcesByCitation(text, four).text, text);
});

test('a source cited twice keeps its first position', async () => {
  const { orderSourcesByCitation } = await load();
  assert.deepEqual(titles(orderSourcesByCitation('[3] ... [1] ... [3] again.', four)).slice(0, 2), ['Gamma', 'Alpha']);
});

test('retrieved but uncited sources follow, marked and still numbered', async () => {
  const { orderSourcesByCitation } = await load();
  const out = orderSourcesByCitation('Only this one [2].', four);
  assert.equal(out.sources[0].title, 'Beta');
  assert.equal(out.sources.length, 4, 'nothing is dropped');
  assert.deepEqual(Array.from(out.sources.slice(1), s => s.uncited), [true, true, true]);
  assert.deepEqual(Array.from(out.sources, s => s.number), [1, 2, 3, 4], 'numbering stays contiguous');
});

test('an invented citation reserves no place', async () => {
  const { orderSourcesByCitation } = await load();
  assert.equal(orderSourcesByCitation('Invented [9]. Real [2].', four).sources[0].title, 'Beta');
});

test('a bracket inside code, math or an HTML attribute is not a citation', async () => {
  // By construction, not by pattern: the rule only runs where the parser runs
  // inline rules, and those are places it never does.
  const { orderSourcesByCitation } = await load();
  const text = '```\nx = arr[3][1]\n```\n\n`m[4]` and $f[3]$ and <span title="see [3]">t</span>\n\nReal [2].';
  const out = orderSourcesByCitation(text, four);
  assert.equal(out.sources[0].title, 'Beta', 'the first real citation is first');
  assert.equal(out.sources[0].sourceNumber, 2);
});

test('the rendered chips agree with the ordered list', async () => {
  const { orderSourcesByCitation, renderAssistantMarkdown } = await load();
  const text = 'See [4], then [2], then [4] again.';
  const ordered = orderSourcesByCitation(text, four).sources;
  const html = renderAssistantMarkdown(text, ordered);
  // [4] is display 1, [2] is display 2 — in the chip text and in the list.
  assert.match(html, /data-source-number="4"[^>]*data-display-number="1"[^>]*>1<\/a>/);
  assert.match(html, /data-source-number="2"[^>]*data-display-number="2"[^>]*>2<\/a>/);
  assert.equal(ordered[0].sourceNumber, 4); assert.equal(ordered[0].number, 1);
});

test('the originals are not mutated', async () => {
  const { orderSourcesByCitation } = await load();
  const before = JSON.parse(JSON.stringify(four));
  orderSourcesByCitation('[3] [1]', four);
  assert.deepEqual(four, before);
});