]*>\d+<\/a>\./);
assert.doesNotMatch(html, /\s* {
const { renderAssistantMarkdown } = await loadCitationModule();
const html = renderAssistantMarkdown('Assess for lethargy [1] and dehydration [2].', sources);
- assert.match(html, /lethargy ]*>src<\/a> and dehydration ]*>src<\/a>/);
+ assert.match(html, /lethargy ]*>\d+<\/a> and dehydration ]*>\d+<\/a>/);
assert.doesNotMatch(html, /<\/a>
\s*and dehydration/);
});
diff --git a/test/assistant-image-attachments.test.js b/test/assistant-image-attachments.test.js
index b9c1f78b..896aefcf 100644
--- a/test/assistant-image-attachments.test.js
+++ b/test/assistant-image-attachments.test.js
@@ -291,6 +291,9 @@ function browserUI(options = {}) {
console: quiet, AbortController, TextDecoder, TextEncoder, URL, Blob, crypto: require('node:crypto').webcrypto,
FileReader: dom.window.FileReader, File: dom.window.File,
setTimeout() {}, showToast(text, kind) { toasts.push([String(text), kind || '']); }, escapeHtml, escapeAttr: escapeHtml,
+ // Renumbering is exercised in test/citation-ordering.test.js against the
+ // real implementation; here it only has to exist and pass things through.
+ orderSourcesByCitation: (text, sources) => ({ text: text, sources: sources || [] }),
renderAssistantMarkdown: text => escapeHtml(text), renderSourcesList: () => '', EMPTY_PROMPT_SETS: [[]],
createAssistantExporter: () => ({ invalidate() {}, exportAnswerPdf() {} }),
createAssistantImageStore: () => ({ renderGeneratedImage: src => '
', clear() {} }),
diff --git a/test/assistant-saved-tables.test.js b/test/assistant-saved-tables.test.js
index 800b299a..8c63ec50 100644
--- a/test/assistant-saved-tables.test.js
+++ b/test/assistant-saved-tables.test.js
@@ -52,7 +52,9 @@ test('raw v2 actual save/load and SSE/fallback/append/export share all table cel
c.fetchAssistantChat = async () => ({ success: true, answer: table, sources });
const loading = c.appendLoadingMessage();
await c.streamAssistantResponse({}, loading);
- assert.deepEqual(rows(bubble(app)), [['Alpha', '1.25', 'A - B', 'src'], ['Beta', '2-4', 'unchanged', 'src']]);
+ // The chip shows its source number now rather than the word "src", so a
+ // reader can match a marker to the numbered list under the answer.
+ assert.deepEqual(rows(bubble(app)), [['Alpha', '1.25', 'A - B', '2'], ['Beta', '2-4', 'unchanged', '1']]);
assert.match(bubble(app).querySelector('.assistant-cite').title, /Source 2: Synthetic B, page 19/);
assert.equal(bubble(app).querySelectorAll('td')[1].getAttribute('align'), 'right');
assert.equal(app.window.getComputedStyle(bubble(app).querySelectorAll('td')[1]).textAlign, 'right');
@@ -110,7 +112,7 @@ test('code, math, escaped pipes and URLs cannot become lists or steal source-col
assert.equal(rows(bubble(app)).length, 2);
assert.equal(rows(bubble(app))[0][1], 'A - B [1] and $x - y$ and \\(a - b\\)');
assert.equal(rows(bubble(app))[1][1], 'a|b and URL');
- assert.equal(rows(bubble(app))[1][2], 'src');
+ assert.equal(rows(bubble(app))[1][2], '1');
assert.equal(bubble(app).querySelector('code .assistant-cite'), null);
assert.equal(bubble(app).querySelector('a[href^="https:"]').getAttribute('href'), 'https://example.test/a-b?q=1%7C2');
for (const literal of ['`' + collapse(table) + '`', '~~~md\n' + collapse(table) + '\n~~~', ' ' + collapse(table), '$$' + collapse(table) + '$$', '\\[' + collapse(table) + '\\]', 'https://example.test/' + collapse(table).replace(/ /g, '%20')]) {
@@ -177,7 +179,7 @@ test('legacy recovery refuses missing boundaries; supports independent lines and
reopen(app, collapse(escaped), 1);
assert.equal(rows(bubble(app)).length, 2);
assert.equal(rows(bubble(app))[0][1], 'a|b and x|y and $x - y$');
- assert.equal(rows(bubble(app))[0][2], 'src');
+ assert.equal(rows(bubble(app))[0][2], '2');
reopen(app, collapse(table) + '\n\nCaption two\n\n' + collapse(table), 1);
assert.equal(bubble(app).querySelectorAll('table').length, 2);
for (const raw of ['| A | B | --- | --- | x | y |', '| A | B | | --- | --- | | x | y', '| A | B | | --- | --- | | x | y | trailing prose']) {
@@ -198,7 +200,7 @@ test('math/code literals and sentinel-shaped input survive postprocessing withou
assert.equal(bubble(app).querySelector('code').textContent, 'a|b [2][1] $notmath$');
assert.equal(bubble(app).querySelector('code .katex, .katex .assistant-cite'), null);
assert.deepEqual(expressions, ['x \\mid y [1]', 'a - b']);
- assert.equal(rows(bubble(app))[1][2], 'src');
+ assert.equal(rows(bubble(app))[1][2], '1');
for (const raw of ['$$\n' + table + '\n$$', '~~~md\n' + table + '\n[1][2] $code$\n~~~', '\uE000html:0\uE001 and `\uE000markdown:0\uE001 [2][1]`']) {
reopen(app, raw);
assert.equal(bubble(app).querySelectorAll('table, .assistant-cite').length, 0);
@@ -408,7 +410,12 @@ test('durable jobs preserve legacy provenance, provisional/clicked turn sources
assert.match(app.document.querySelector('#assistant-source-2').textContent, /Synthetic B.*19/is);
await c.performAutosave(); const saved = app.saves.at(-1);
assert.equal(saved.messages[0].content, raw); assert.equal(saved.messages[0].retainedAnswer, retained); assert.equal(saved.messages[0].legacyClipped, true);
- assert.deepEqual(saved.messages[0].sources, sources); assert.deepEqual(saved.sources, secondSources);
+ assert.deepEqual(saved.messages[0].sources, sources);
+ // Sources are reordered into citation order on the final render, and anything
+ // retrieved but never cited is marked so the panel can say so. This answer
+ // cites only the first, so the second is carried along as "not cited".
+ assert.deepEqual(saved.sources,
+ [secondSources[0], Object.assign({}, secondSources[1], { uncited: true })]);
assert.deepEqual(saved.messages[0].imageJobs, [{ jobId: id }]); assert.equal(saved.messages.at(-1).content, second);
c.restoreSavedChat(saved); await tick();
const before = JSON.stringify(c.messages);
diff --git a/test/citation-ordering.test.js b/test/citation-ordering.test.js
new file mode 100644
index 00000000..b3a2ad1f
--- /dev/null
+++ b/test/citation-ordering.test.js
@@ -0,0 +1,111 @@
+// 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], so matching a marker to a source meant hunting for it.
+//
+// Borrowed from the quiz app, where validating citations and ordering them fall
+// out of the same pass: it collects the sources the answer actually used into an
+// insertion-ordered map, so the list comes back in first-citation order for
+// free. Reference lists in published writing are numbered by first appearance
+// for the same reason.
+const test = require('node:test');
+const assert = require('node:assert/strict');
+const fs = require('node:fs');
+const path = require('node:path');
+const vm = require('node:vm');
+
+function load() {
+ const src = fs.readFileSync(path.join(__dirname, '..', 'public/js/assistant/citations.js'), 'utf8')
+ .replace(/^import[\s\S]*?from ['"][^'"]+['"];\s*/gm, '')
+ .replace(/^export /gm, '');
+ const ctx = { window: {}, document: undefined, console };
+ vm.createContext(ctx);
+ vm.runInContext(src + '\nthis.orderSourcesByCitation = orderSourcesByCitation;', ctx);
+ return ctx;
+}
+
+const four = [
+ { number: 1, title: 'Alpha' }, { number: 2, title: 'Beta' },
+ { number: 3, title: 'Gamma' }, { number: 4, title: 'Delta' }
+];
+
+test('sources come back in the order the answer cites them', () => {
+ const { orderSourcesByCitation } = load();
+ const out = orderSourcesByCitation('Third first [3]. Then the first [1].', four);
+ assert.deepEqual(Array.from(out.sources.slice(0, 2), s => s.title), ['Gamma', 'Alpha']);
+ assert.equal(out.text, 'Third first [1]. Then the first [2].');
+});
+
+test('a cluster is renumbered as a whole, in its own order', () => {
+ const { orderSourcesByCitation } = load();
+ const out = orderSourcesByCitation('Both of these [4, 2] agree.', four);
+ assert.equal(out.text, 'Both of these [1, 2] agree.');
+ assert.deepEqual(Array.from(out.sources.slice(0, 2), s => s.title), ['Delta', 'Beta']);
+});
+
+test('renumbering happens in one pass, so nothing is renumbered twice', () => {
+ // Rewriting number by number turns 2 into 1, then that 1 into whatever 1
+ // maps to. The whole text is rewritten once instead.
+ const { orderSourcesByCitation } = load();
+ const out = orderSourcesByCitation('[2] then [1] then [2] again.', four);
+ assert.equal(out.text, '[1] then [2] then [1] again.');
+});
+
+test('a source cited twice keeps its first position', () => {
+ const { orderSourcesByCitation } = load();
+ const out = orderSourcesByCitation('[3] ... [1] ... [3] again.', four);
+ assert.deepEqual(Array.from(out.sources.slice(0, 2), s => s.title), ['Gamma', 'Alpha']);
+});
+
+test('retrieved but uncited sources follow, marked and still numbered', () => {
+ // The panel is also a view of what the search returned, so they stay — just
+ // no longer mixed in among the numbers the answer used.
+ const { orderSourcesByCitation } = load();
+ const out = orderSourcesByCitation('Only this one [2].', four);
+ assert.equal(out.sources[0].title, 'Beta');
+ assert.equal(out.sources[0].uncited, undefined);
+ 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 and is left alone', () => {
+ // It is not turned into a link either; the audit records it. What matters
+ // here is that it cannot push a real source down the list.
+ const { orderSourcesByCitation } = load();
+ const out = orderSourcesByCitation('Invented [9]. Real [2].', four);
+ assert.equal(out.sources[0].title, 'Beta');
+ assert.match(out.text, /Invented \[9\]/, 'the unresolved marker is untouched');
+ assert.match(out.text, /Real \[1\]/);
+});
+
+test('a cluster containing an invented number is left whole', () => {
+ // Renumbering half of it would silently change which source the good half
+ // points at.
+ const { orderSourcesByCitation } = load();
+ const out = orderSourcesByCitation('Mixed [2, 9].', four);
+ assert.equal(out.text, 'Mixed [2, 9].');
+});
+
+test('an answer that cites nothing is returned untouched', () => {
+ const { orderSourcesByCitation } = load();
+ const out = orderSourcesByCitation('No citations here.', four);
+ assert.equal(out.text, 'No citations here.');
+ assert.deepEqual(Array.from(out.sources), four);
+});
+
+test('the originals are not mutated, so a stored answer stays readable', () => {
+ // Its text still holds the original markers; renumbering in place would make
+ // the two disagree.
+ const { orderSourcesByCitation } = load();
+ const before = JSON.parse(JSON.stringify(four));
+ orderSourcesByCitation('[3] [1]', four);
+ assert.deepEqual(four, before);
+});
+
+test('sources with no number field fall back to position', () => {
+ const { orderSourcesByCitation } = load();
+ const bare = [{ title: 'One' }, { title: 'Two' }, { title: 'Three' }];
+ const out = orderSourcesByCitation('Cite the third [3].', bare);
+ assert.equal(out.sources[0].title, 'Three');
+ assert.equal(out.text, 'Cite the third [1].');
+});
diff --git a/test/clinical-conversation.test.js b/test/clinical-conversation.test.js
index 36a84c7d..5914c458 100644
--- a/test/clinical-conversation.test.js
+++ b/test/clinical-conversation.test.js
@@ -175,6 +175,9 @@ function browserUI(options = {}) {
window: dom.window, document: dom.window.document, navigator: dom.window.navigator,
console: quiet, AbortController, TextDecoder, TextEncoder, URL, Blob, crypto: require("node:crypto").webcrypto,
setTimeout() {}, showToast() {}, escapeHtml, escapeAttr: escapeHtml,
+ // Renumbering is exercised in test/citation-ordering.test.js against the
+ // real implementation; here it only has to exist and pass things through.
+ orderSourcesByCitation: (text, sources) => ({ text: text, sources: sources || [] }),
renderAssistantMarkdown: text => escapeHtml(text), renderSourcesList: () => '', ...options.renderers, EMPTY_PROMPT_SETS: [[]],
createAssistantExporter: () => ({ invalidate() {}, exportAnswerPdf() {} }),
createAssistantImageStore: () => ({ renderGeneratedImage: src => '', clear() {} }),