# Citations et statistiques hors des bords de ligne

> Un article clinique à deux colonnes composé deux fois : la première épreuve signale chaque appel ou signe en bord de ligne, la seconde pénalise ces coupures.

- Version HTML: https://postext.dev/fr/cookbook/clinical-paper-line-edges
- Recette Nº 186 · Caractères et texte · Niveau 2 (Intermédiaire) · Sorties: Canvas
- Genres: Articles et publications universitaires
- Nécessite postext ≥ 1.79.0 · testée avec 1.79.0 le 2026-11-12
- Pages: [211](https://postext.dev/cookbook/clinical-paper-line-edges/en/p01.webp?v=56d4126e), [212](https://postext.dev/cookbook/clinical-paper-line-edges/en/p02.webp?v=56d4126e), [213](https://postext.dev/cookbook/clinical-paper-line-edges/en/p03.webp?v=56d4126e), [213](https://postext.dev/cookbook/clinical-paper-line-edges/en/p06.webp?v=56d4126e)
- Ouvrir dans le Sandbox: https://postext.dev/fr/sandbox#recipe=clinical-paper-line-edges&lang=en (.postext: https://postext.dev/cookbook/clinical-paper-line-edges/en/clinical-paper-line-edges.postext)
- Dernière mise à jour: 2026-11-12
- Autres langues: [en](https://postext.dev/en/cookbook/clinical-paper-line-edges.md), [es](https://postext.dev/es/cookbook/clinical-paper-line-edges.md), [ca](https://postext.dev/ca/cookbook/clinical-paper-line-edges.md), [de](https://postext.dev/de/cookbook/clinical-paper-line-edges.md), [it](https://postext.dev/it/cookbook/clinical-paper-line-edges.md), [pt](https://postext.dev/pt/cookbook/clinical-paper-line-edges.md), [zh](https://postext.dev/zh/cookbook/clinical-paper-line-edges.md), [ja](https://postext.dev/ja/cookbook/clinical-paper-line-edges.md), [ar](https://postext.dev/ar/cookbook/clinical-paper-line-edges.md)

## En bref

Dans une colonne justifiée étroite, un appel comme [2, 3] peut ouvrir seul une ligne, et p < 0,05 peut se couper après le signe. L’article est composé deux fois : l’une montre où cela arrive, l’autre l’évite presque partout.

## Ce que vous allez composer

Trois pages d’une revue de pédiatrie de 170 × 240 mm, sur deux colonnes justifiées de 66 mm en STIX Two Text : une étude originale sur l’échographie pulmonaire dans la pneumonie de l’enfant, avec une bibliographie numérotée en style Vancouver. Une colonne aussi étroite laisse peu de choix à l’algorithme de coupure, et à trois endroits le texte anglais se coupe mal. Un appel, *[2, 3]*, ouvre seul une ligne, loin du mot auquel il se rapporte, et deux fois le rapport de vraisemblance *LR− = 0.09* se coupe après le signe égal. La première épreuve est composée avec `citationBreaks` et `operatorBreaks` à `'warn'`. Elle ne change aucune coupure, marque chacune d’une bande ambre pâle et les liste sur un bulletin après la bibliographie. La seconde est composée avec `'avoid'`, et il en reste une sur trois. L’édition espagnole compose la même étude avec la virgule décimale.

**Cette recette répond à:**

- Comment empêcher qu’un appel de citation ou une statistique comme p < 0,05 se coupe en fin de ligne ?
- Comment composer les numéros de citation en exposant et rendre les DOI cliquables, comme dans les revues médicales ?

## La réponse courte

```js
// script.js, lignes 27–42
const bodyText = {
  fontFamily: SERIF, fontSize: pt(9.5), lineHeight: pt(LEAD), color: col('ink'),
  boldColor: col('ink'), italicColor: col('ink'), referenceColor: col('ink'),
  referenceBold: false, textAlign: 'justify', firstLineIndent: mm(3.5),
  indentAfterHeading: false, hyphenation: { enabled: true },
  optimalLineBreaking: true, // the rules price breaks in the Knuth–Plass breaker
  avoidWidows: true, avoidOrphans: true, avoidRunts: true,
  // A line that opens on a citation, "[8, 9]" (citationBreak), or ends or opens on a sign
  // set between spaces, "p <" / "0.05" (operatorBreak). 'warn' reports each one and changes
  // nothing; 'avoid' adds the penalty to the demerits of those breaks and reports the ones
  // it keeps. The first proof sets 'warn' (see proofs).
  citationBreaks: 'avoid', operatorBreaks: 'avoid',
  // 10000 by default, what a line spaced to the limit costs. A 66 mm column offers few other
  // breaks; at 30000 the breaker takes looser lines, never past maxWordSpacing.
  citationBreakPenalty: 30000, operatorBreakPenalty: 30000,
};
```

## Ingrédients

**Enseigne**

- [Appels de référence et signes en bord de ligne](https://postext.dev/fr/docs/configuration-text.md#texte-courant): Repère les lignes qui commencent par un appel de référence numéroté ([12], (12–15)) ou qui finissent ou commencent par un signe entre espaces (p < 0,05, RV+ = 19,19), et les signale (bodyText.citationBreaks / operatorBreaks : 'warn') ou fait composer le paragraphe autrement par la coupure des lignes ('avoid', au coût citationBreakPenalty / operatorBreakPenalty).
- [Coupure optimale des lignes (Knuth–Plass)](https://postext.dev/fr/docs/justification.md#knuth-plass-voir-le-paragraphe-entier): Coupe chaque paragraphe dans son ensemble pour obtenir un espacement des mots régulier, dans les limites de l’espacement des mots, au lieu de remplir ligne par ligne.

**Utilise aussi**

- [Citations dans un style bibliographique](https://postext.dev/fr/docs/document-format-notes-references.md#citations-et-bibliographie)
- [Bibliographie tirée des références](https://postext.dev/fr/docs/document-format-notes-references.md#citations-et-bibliographie)
- [Surlignages](https://postext.dev/fr/docs/document-format-inline.md#surlignages)
- [Césure et langue du document](https://postext.dev/fr/docs/justification.md#langues-prises-en-charge)
- [Une ou deux colonnes](https://postext.dev/fr/docs/configuration-page-layout.md#types-de-mise-en-page)
- [Ouvertures composées](https://postext.dev/fr/docs/configuration-headings.md#étendue-et-maquette-avancée)
- [Attributs de titre](https://postext.dev/fr/docs/document-format.md#attributs-de-titre)
- [Styles de titre](https://postext.dev/fr/docs/configuration-styles.md#styles-de-titre)
- [Titres courants et folios](https://postext.dev/fr/docs/configuration-headers-footers.md#en-têtes-et-pieds-de-page)
- [Titres courants selon le rôle de la page](https://postext.dev/fr/docs/configuration-headers-footers.md#éléments-de-texte)
- [Marges en miroir](https://postext.dev/fr/docs/configuration-page-layout.md#marges-en-miroir)
- [Encadrés](https://postext.dev/fr/docs/configuration-callouts.md#styles-dencadré)
- [Palette de couleurs sémantique](https://postext.dev/fr/docs/configuration-fonts-colors-viewers.md#palette-de-couleurs)
- [Sortir de la grille volontairement](https://postext.dev/fr/docs/architecture.md#éléments-qui-rompent-la-grille)
- [Styles de paragraphe](https://postext.dev/fr/docs/configuration-styles.md#styles-de-paragraphe)
- [Pages liminaires en chiffres romains](https://postext.dev/fr/docs/document-format-directives.md#numbering)
- [Exposants et indices](https://postext.dev/fr/docs/document-format-inline.md#mise-en-forme-en-ligne)
- [Chapitres non numérotés](https://postext.dev/fr/docs/configuration-styles.md#styles-de-titre)

**La configuration en un coup d’œil**

- [`bodyText`](https://postext.dev/fr/docs/configuration-text.md#texte-courant), [`calloutStyles`](https://postext.dev/fr/docs/configuration-callouts.md#styles-dencadré), [`checks`](https://postext.dev/fr/docs/configuration-fonts-colors-viewers.md#contrôle-du-contenu-configuration), [`citations`](https://postext.dev/fr/docs/configuration-notes-references.md#citations), [`colorPalette`](https://postext.dev/fr/docs/configuration-fonts-colors-viewers.md#palette-de-couleurs), [`footer`](https://postext.dev/fr/docs/configuration-headers-footers.md#en-têtes-et-pieds-de-page), [`header`](https://postext.dev/fr/docs/configuration-headers-footers.md#en-têtes-et-pieds-de-page), [`headingStyles`](https://postext.dev/fr/docs/configuration-styles.md#styles-de-titre), [`headings`](https://postext.dev/fr/docs/configuration-headings.md#titres), [`layout`](https://postext.dev/fr/docs/configuration-page-layout.md#mise-en-page), [`locale`](https://postext.dev/fr/docs/configuration-language.md#césure), [`markStyles`](https://postext.dev/fr/docs/configuration-styles.md#styles-de-surlignage), [`page`](https://postext.dev/fr/docs/configuration-page-layout.md#page), [`paragraphStyles`](https://postext.dev/fr/docs/configuration-styles.md#styles-de-paragraphe)

**API**

- [`LOCALES`](https://postext.dev/fr/docs/document-format-notes-references.md#citations-et-bibliographie), [`STYLES`](https://postext.dev/fr/docs/document-format-notes-references.md#citations-et-bibliographie), [`buildDocumentWithFonts`](https://postext.dev/fr/docs/configuration-programmatic-usage.md#charger-les-polices-avant-la-mise-en-page), [`createCiteprocEngine`](https://postext.dev/fr/docs/document-format-notes-references.md#citations-et-bibliographie), [`registerCitationEngine`](https://postext.dev/fr/docs/document-format-notes-references.md#citations-et-bibliographie), [`renderPageToCanvas`](https://postext.dev/fr/docs/configuration-programmatic-usage.md#rendre-une-page-en-bitmap)

**Caractères**

- STIX Two Text (OFL-1.1), IBM Plex Sans (OFL-1.1), IBM Plex Sans Condensed (OFL-1.1)

## Préparation

### 1 · Un prix pour deux sortes de coupure

Le code est [la réponse courte](https://postext.dev/fr/cookbook/clinical-paper-line-edges.md#la-réponse-courte) ci-dessus. Il y a coupure de citation quand une ligne commence par un appel numéroté, ou par la seconde moitié d’un appel, comme *5]* après *[4,*. Il y a coupure d’opérateur quand une ligne finit ou commence par un signe placé entre deux espaces : `=`, `<`, `±`, `×`, le signe moins. `'warn'` signale les deux et laisse les lignes telles quelles. `'avoid'` ajoute une pénalité aux démérites de ces coupures, et l’algorithme de Knuth–Plass cherche une autre façon de couper le paragraphe. Il faut `optimalLineBreaking` : la coupure gloutonne n’a pas de démérites auxquels ajouter quoi que ce soit.

La pénalité par défaut, 10000, est ce que coûte une ligne étirée jusqu’à la limite de l’espacement entre les mots : l’algorithme n’ouvre donc jamais une ligne au-delà de cette limite pour garder un appel avec son mot. Dans cette colonne, 10000 laissait deux coupures dans chaque édition. À 30000, l’algorithme accepte des lignes plus lâches, toujours dans la limite de `maxWordSpacing`, et il en reste une dans chaque édition. Une valeur plus haute ne changeait plus rien : la dernière coupure n’a pas d’autre solution dans la limite.

### 2 · Marquer les coupures dans le texte

```js
// script.js, lignes 254–279
// The word before the citation and the citation, or the sign and the number on its other
// side, wrapped in :mark[…]: a band under ordinary text, so the breaks stay where they were.
const wordEnd = (s, i) => i + s.slice(i).search(/[\s;)]|[,.](?!\d)|$/); // 0.09, not 0
function sides(source, w, a, b) { // a warning names line a, before the break, by its source
  if (w.kind === 'citationBreak') { // "colleagues [@simel1991]"
    const cite = source.lastIndexOf('[@', b.sourceStart + 2);
    return [source.lastIndexOf(' ', cite - 2) + 1, source.indexOf(']', cite) + 1];
  }
  if (w.edge === 'end') { // "LR− = 0.09": the sign ends line a, the number opens line b
    const from = source.lastIndexOf(` ${w.text} `, a.sourceEnd) + 1;
    return [from, wordEnd(source, from + w.text.length + 1)];
  }
  const sign = source.indexOf(` ${w.text} `, b.sourceStart - 1); // the sign opens line b
  return [source.lastIndexOf(' ', sign - 1) + 1, sign + 1 + w.text.length];
}
function markBreaks(source, doc, found) {
  const pairs = doc.blocks.flatMap((b) => b.lines.map((line, i) => [line, b.lines[i + 1]]));
  const spans = new Map();
  for (const w of found) {
    const [a, b] = pairs.find(([line]) => line.sourceStart === w.sourceStart) ?? [];
    const [from, to] = b ? sides(source, w, a, b) : [];
    if (from > a?.sourceStart && to > from && to <= b.sourceEnd) spans.set(from, to);
  }
  return [...spans].sort(([x], [y]) => y - x).reduce((md, [from, to]) => `${md.slice(0, from)}`
    + `:mark[${md.slice(from, to)}]{style="edge"}${md.slice(to)}`, source);
}
```

Chaque avertissement indique sa sorte, l’appel ou le signe, le texte des deux lignes, la position de la première dans la source et la page. Le script retrouve ces deux lignes dans le document composé et enveloppe la coupure dans `:mark[…]` : le mot qui précède l’appel et l’appel, ou le signe et le nombre de l’autre côté. Une marque peint une bande sous le texte et laisse les coupures où elles étaient. La bande n’atteint pas le numéro que produit le style de citation : sur la page, seul le mot qui le précède est marqué.

![Opening page 211: Point-of-care lung ultrasound for community- acquired pneumonia in children: a prospective diagnostic accuracy study.](https://postext.dev/cookbook/clinical-paper-line-edges/en/p01.webp?v=56d4126e)

*Première épreuve, page 211 : LR− = 0.09 coupé dans le résumé, et first en fin de ligne, avec [2, 3] au début de la suivante.*

### 3 · Deux épreuves et un bulletin de questions

```js
// script.js, lignes 283–303
const KIND = { citationBreak: t({ en: 'citation', es: 'cita' }),
  operatorBreak: t({ en: 'sign', es: 'signo' }) };
const edge = (s, end) => (end ? s.trim().split(/\s+/).slice(-2) : s.trim().split(/\s+/)
  .slice(0, 2)).join('\u00a0').replace(/[.,;]$/, ''); // no-break spaces: no breaks of its own
const query = (doc, w) => `${t({ en: 'p.', es: 'pág.' })}\u00a0${doc.pages[w.pageIndex]
  .pageLabel}, ${KIND[w.kind]}: …${edge(w.lines[0], true)}\u00a0/ ${edge(w.lines[1])}…`;
const slip = (mode, doc, found) => ['', `:::callout{type="slip" title="${mode === 'warn'
  ? t({ en: 'First proof: queries', es: 'Primeras pruebas: consultas' })
  : t({ en: 'Second proof: queries', es: 'Segundas pruebas: consultas' })}"}`,
`citationBreaks / operatorBreaks: '${mode}' · ${found.length} ${found.length === 1
  ? t({ en: 'break', es: 'corte' }) : t({ en: 'breaks', es: 'cortes' })}.`,
...found.flatMap((w) => ['', query(doc, w)]), ':::', ''].join('\n');
const build = (md, mode) => buildDocumentWithFonts({ markdown: md }, { ...config,
  bodyText: { ...bodyText, citationBreaks: mode, operatorBreaks: mode } }, kitFonts(FONTS));
async function proof(mode) {
  const plain = await build(markdown, mode); // pass 1: where the breaks are
  const found = (plain.contentWarnings ?? []).filter((w) => /^(citation|operator)Break$/
    .test(w.kind));
  return build(markBreaks(markdown, plain, found) + slip(mode, plain, found), mode); // pass 2
}
const docs = { warn: await proof('warn'), avoid: await proof('avoid') };
```

Chaque épreuve est composée deux fois. La première composition trouve les coupures ; la seconde compose l’article avec les marques et un bulletin après la bibliographie, qui tombe donc sur la dernière page. C’est ainsi qu’un compositeur renvoie ses épreuves avec ses questions. Les lignes du bulletin lient leurs mots par des espaces insécables pour ne pas créer de coupures à leur tour. `capture.doc` publie les deux compositions marquées, `[1, 3]`.

![Page 213: Article information.](https://postext.dev/cookbook/clinical-paper-line-edges/en/p06.webp?v=56d4126e)

*Seconde épreuve, page 213 : le bulletin liste une coupure restante dans l’article.*

La coupure qui reste en anglais est *first* au-dessus de *[2, 3]*, page 211. En espagnol, la première épreuve en trouve quatre, dont *CP+* au-dessus de *= 19,19* et *[4,* au-dessus de *5]*, où l’appel lui-même se coupe, et la seconde en garde une, *anteriores [4,* au-dessus de *5]*.

### 4 · Des appels numérotés entre crochets

```js
// script.js, lignes 46–50
registerCitationEngine(createCiteprocEngine({ styles: STYLES, locales: LOCALES }));
const citations = { style: 'elsevier-vancouver', locale: t({ en: 'en-GB', es: 'es-ES' }),
  marker: 'brackets', collapseRanges: true, // [8, 9] and [4–6], set after a word space
  bibliography: { fontSize: em(0.8), lineHeight: pt(9.6), entrySpacing: pt(1),
    labelWidth: mm(6), labelAlign: 'right', doi: 'hide' } };
```

Le processeur de citations compose chaque `[@key]` sous la forme d’un numéro entre crochets après une espace et regroupe les suites comme *[4–6]*. L’espace avant le crochet et celle qui suit une virgule à l’intérieur sont deux endroits où une ligne peut se couper : c’est pourquoi une colonne étroite chargée de citations a besoin de la règle.

### 5 · Le bandeau de la revue et le bloc de l’article

```js
// script.js, lignes 54–78
const text = (id, content, family, size, extra) => ({ kind: 'text', id, content, align: 'left',
  fontFamily: family, fontSize: pt(size), color: col('ink'), overflow: 'wrap', ...extra });
const at = (to, edge, x, y, width) => ({ anchor: { to, edge }, offset: { x: mm(x), y: mm(y) },
  ...(width && { size: { width: mm(width), height: 'auto' } }) });
const caps = { fontFamily: COND, fontWeight: 600, letterSpacing: pt(1.2),
  textTransform: 'uppercase' };
const JOURNAL = t({ en: 'Ribera Paediatric Journal', es: 'Revista Pediátrica de la Ribera' });
const white = { ...caps, color: col('paper') };
const titleBlock = { enabled: true, minHeight: mm(58), slot: { elements: [
  { kind: 'box', id: 'band', style: { backgroundColor: col('accent') }, placement: {
    anchor: { to: 'page', edge: 'top-left' }, size: { width: 'fill', height: mm(13) } } },
  text('journal', JOURNAL, COND, 9, { ...white, placement: at('page', 'top-left', INNER, 5) }),
  text('issue', '{attr.issue}', COND, 9, { ...white, align: 'right',
    placement: at('page', 'top-right', -OUTER, 5, 60) }),
  text('kind', '{attr.kind}', COND, 8.5, { ...caps, color: col('accent'),
    placement: at('container', 'top-left', 0, 0, MEASURE) }),
  text('title', '{titleText}', SANS, 16, { fontWeight: 600, lineHeight: 1.18,
    placement: at('#kind', 'below', 0, 2.5, MEASURE - 6) }),
  text('authors', '{attr.authors}', SERIF, 10, { inlineMarks: true, // ^1^ → ¹
    placement: at('#title', 'below', 0, 3.5, MEASURE) }),
  text('affiliations', '{attr.affiliations}', SANS, 7.2, { inlineMarks: true, lineHeight: 1.35,
    color: col('muted'), placement: at('#authors', 'below', 0, 1.6, MEASURE) }),
  text('note', '{attr.note}', SANS, 7.2, { color: col('muted'),
    placement: at('#affiliations', 'below', 0, 2.4, MEASURE) }),
] } };
```

Le style de titre dessine le bandeau de la revue, le type d’article, le titre, les auteurs avec leurs numéros d’affiliation et une note, le tout à partir d’attributs du H1. Les deux éditions partagent ce style, et le texte reste dans le Markdown.

## La recette complète

Un seul fichier, composé à partir du dossier de la recette avec le texte d’exemple et le kit commun des Recettes déjà inclus ; il construit sa propre page. Pour l’exécuter, placez-le dans un `<script type="module">` d’une page vide ou collez-le dans le panneau JS d’un nouveau pen CodePen (en tant que module). Il importe postext depuis esm.sh : il n’y a rien à installer ni à compiler.

- Dossier de la recette: https://github.com/drnachio/postext/tree/main/cookbook/clinical-paper-line-edges

### script.js

```js
// ═══ Postext Cookbook · Nº 186 · Citations and statistics off the line edges ═══════════
// https://postext.dev/en/cookbook/clinical-paper-line-edges
// Code: MIT · Text: original (CC BY 4.0) · Pictures: none
// Fonts: STIX Two Text, IBM Plex Sans and Sans Condensed (SIL OFL 1.1) · Needs postext ≥ 1.79.0
import {
  buildDocumentWithFonts, renderPageToCanvas, registerCitationEngine,
} from 'https://esm.sh/postext';
import { createCiteprocEngine, STYLES, LOCALES } from 'https://esm.sh/postext-citeproc';

const LANG = 'en'; // @lang: the language of the sample document ('en' | 'es')
const RECIPE = 'clinical-paper-line-edges';

// ─── 1 · Design ─────────────────────────────────────────────────────────────
// #region palette: an indigo journal, a pale amber band for the proof marks
const palette = { ink: '#1b1e26', accent: '#2c4a7c', tint: '#e9eef6', mark: '#f7dfa9',
  muted: '#596173', paper: '#ffffff' };
const col = (id) => ({ hex: palette[id], model: 'hex', paletteId: id });
const colorPalette = Object.entries({ ...palette, 'main-color': palette.accent })
  .map(([id, hex]) => ({ id, name: id, value: { hex, model: 'hex' } }));
// #endregion
const [SERIF, SANS, COND] = ['STIX Two Text', 'IBM Plex Sans', 'IBM Plex Sans Condensed'];
const [TRIM_W, TRIM_H, TOP, BOTTOM, INNER, OUTER, GUTTER] = [170, 240, 20, 22, 17, 15, 6];
const MEASURE = TRIM_W - INNER - OUTER; // 138 mm: two columns of 66 mm, about 45 characters
const LEAD = 12.8; // pt, for 9.5 pt type

// #region answer: no line opens on a citation, no line ends or opens on a lone sign
const bodyText = {
  fontFamily: SERIF, fontSize: pt(9.5), lineHeight: pt(LEAD), color: col('ink'),
  boldColor: col('ink'), italicColor: col('ink'), referenceColor: col('ink'),
  referenceBold: false, textAlign: 'justify', firstLineIndent: mm(3.5),
  indentAfterHeading: false, hyphenation: { enabled: true },
  optimalLineBreaking: true, // the rules price breaks in the Knuth–Plass breaker
  avoidWidows: true, avoidOrphans: true, avoidRunts: true,
  // A line that opens on a citation, "[8, 9]" (citationBreak), or ends or opens on a sign
  // set between spaces, "p <" / "0.05" (operatorBreak). 'warn' reports each one and changes
  // nothing; 'avoid' adds the penalty to the demerits of those breaks and reports the ones
  // it keeps. The first proof sets 'warn' (see proofs).
  citationBreaks: 'avoid', operatorBreaks: 'avoid',
  // 10000 by default, what a line spaced to the limit costs. A 66 mm column offers few other
  // breaks; at 30000 the breaker takes looser lines, never past maxWordSpacing.
  citationBreakPenalty: 30000, operatorBreakPenalty: 30000,
};
// #endregion

// #region citations: numbered references in brackets, the list in the Vancouver order
registerCitationEngine(createCiteprocEngine({ styles: STYLES, locales: LOCALES }));
const citations = { style: 'elsevier-vancouver', locale: t({ en: 'en-GB', es: 'es-ES' }),
  marker: 'brackets', collapseRanges: true, // [8, 9] and [4–6], set after a word space
  bibliography: { fontSize: em(0.8), lineHeight: pt(9.6), entrySpacing: pt(1),
    labelWidth: mm(6), labelAlign: 'right', doi: 'hide' } };
// #endregion

// #region title: the masthead band, article type, title, authors and the note
const text = (id, content, family, size, extra) => ({ kind: 'text', id, content, align: 'left',
  fontFamily: family, fontSize: pt(size), color: col('ink'), overflow: 'wrap', ...extra });
const at = (to, edge, x, y, width) => ({ anchor: { to, edge }, offset: { x: mm(x), y: mm(y) },
  ...(width && { size: { width: mm(width), height: 'auto' } }) });
const caps = { fontFamily: COND, fontWeight: 600, letterSpacing: pt(1.2),
  textTransform: 'uppercase' };
const JOURNAL = t({ en: 'Ribera Paediatric Journal', es: 'Revista Pediátrica de la Ribera' });
const white = { ...caps, color: col('paper') };
const titleBlock = { enabled: true, minHeight: mm(58), slot: { elements: [
  { kind: 'box', id: 'band', style: { backgroundColor: col('accent') }, placement: {
    anchor: { to: 'page', edge: 'top-left' }, size: { width: 'fill', height: mm(13) } } },
  text('journal', JOURNAL, COND, 9, { ...white, placement: at('page', 'top-left', INNER, 5) }),
  text('issue', '{attr.issue}', COND, 9, { ...white, align: 'right',
    placement: at('page', 'top-right', -OUTER, 5, 60) }),
  text('kind', '{attr.kind}', COND, 8.5, { ...caps, color: col('accent'),
    placement: at('container', 'top-left', 0, 0, MEASURE) }),
  text('title', '{titleText}', SANS, 16, { fontWeight: 600, lineHeight: 1.18,
    placement: at('#kind', 'below', 0, 2.5, MEASURE - 6) }),
  text('authors', '{attr.authors}', SERIF, 10, { inlineMarks: true, // ^1^ → ¹
    placement: at('#title', 'below', 0, 3.5, MEASURE) }),
  text('affiliations', '{attr.affiliations}', SANS, 7.2, { inlineMarks: true, lineHeight: 1.35,
    color: col('muted'), placement: at('#authors', 'below', 0, 1.6, MEASURE) }),
  text('note', '{attr.note}', SANS, 7.2, { color: col('muted'),
    placement: at('#affiliations', 'below', 0, 2.4, MEASURE) }),
] } };
// #endregion

const small = { fontFamily: SANS, fontSize: pt(7.8), lineHeight: pt(10.6), textAlign: 'left',
  firstLineIndent: pt(0), color: col('ink'), boldColor: col('accent') };
const box = (id, spaced, more) => ({ id, border: { enabled: false }, snapToGrid: false,
  background: col('tint'), padding: mm(3), marginTop: pt(LEAD / 2), marginBottom: pt(LEAD),
  titleStyle: { ...caps, fontSize: pt(8), color: col('accent') },
  body: { ...small, paragraphSpacing: spaced }, ...more });
const calloutStyles = [box('abstract', true, { marginTop: pt(0) }), box('slip', false,
  { stripe: { enabled: true, side: 'top', width: pt(2.5), color: col('accent') } })];
const head = (id, content, parity, edge, x, extra) => text(id, content, COND, 7.8, {
  parity, pages: 'body', fontWeight: 500, letterSpacing: pt(0.4), color: col('muted'),
  overflow: 'clip', placement: at('page', edge, x, 11, 100), ...extra });
const [folio, right] = [{ fontWeight: 600, color: col('accent') }, { align: 'right' }];
const header = { elements: [
  head('v-folio', '{pageNumber}', 'even', 'top-left', OUTER, folio),
  head('v-title', '{attr.verso}', 'even', 'top-left', OUTER + 9), // the H1's attributes
  head('r-title', '{attr.recto}', 'odd', 'top-right', -OUTER - 9, right),
  head('r-folio', '{pageNumber}', 'odd', 'top-right', -OUTER, { ...folio, ...right }),
] };
const footer = { elements: [head('drop-folio', '{pageNumber}', 'all', 'bottom-right', -OUTER,
  { ...folio, ...right, pages: 'opener', placement: at('page', 'bottom-right', -OUTER, -12, 10) },
)] };

const sans = (size) => ({ fontFamily: SANS, fontSize: pt(size), fontWeight: 600 });
const config = {
  locale: t({ en: 'en-gb', es: 'es' }),
  colorPalette, citations, calloutStyles, header, footer, bodyText,
  checks: { contentInConfig: { allow: [JOURNAL] } }, // the journal's name (#849)
  headingStyles: [{ id: 'article', numbered: false, span: 'page', advancedDesign: titleBlock },
    { id: 'back', numbered: false, ...sans(9), color: col('ink') }],
  page: { sizePreset: 'custom', width: mm(TRIM_W), height: mm(TRIM_H), dpi: 150,
    pageNumbering: { startAt: 211 }, // the article opens on page 211 of the issue
    margins: { top: mm(TOP), bottom: mm(BOTTOM), left: mm(INNER), right: mm(OUTER),
      mirror: true } },
  layout: { layoutType: 'double', gutterWidth: mm(GUTTER) },
  headings: { fontFamily: SANS, color: col('accent'), fontWeight: 600, levels: [
    { level: 1, breakBefore: { enabled: true, parity: 'any' } },
    { level: 2, ...sans(10.5), lineHeight: pt(LEAD), marginTop: pt(LEAD), marginBottom: pt(0) },
    { level: 3, ...sans(9), color: col('ink'), lineHeight: pt(LEAD), marginTop: pt(LEAD / 2),
      marginBottom: pt(0) },
  ] },
  paragraphStyles: [{ id: 'info', ...small }], // the article information, as journals set it
  markStyles: [{ id: 'edge', name: t({ en: 'Line edge', es: 'Borde de línea' }), color: col('ink'),
    background: col('mark'), paddingX: em(0.08), paddingTop: em(0.02), paddingBottom: em(0.08) }],
};

// ─── 2 · Content ────────────────────────────────────────────────────────────
const markdown = String.raw`---
title: "Lung ultrasound for pneumonia in children"
author: "Inés Barreiro Soto, Tomasz Wierzbicki, Aisha Mensah and Pablo Lago Fernández"
---

# Point-of-care lung ultrasound for community-acquired pneumonia in children: a prospective diagnostic accuracy study {style="article" issue="2026 · 9(4) · 211–213" verso="Ribera Paediatr J 2026;9(4)" recto="Barreiro Soto et al. · Lung ultrasound for pneumonia" kind="Original research" authors="Inés Barreiro Soto^1^, Tomasz Wierzbicki^2^, Aisha Mensah^1^, Pablo Lago Fernández^3^" affiliations="^1^ Paediatric Emergency Department, Hospital Comarcal de la Ribera\n^2^ Department of Radiology, Hospital Comarcal de la Ribera\n^3^ Clinical Epidemiology Unit, Ribera Health Area" note="Illustrative article written for the Postext Cookbook: the study, its authors and its data are invented. The works cited are real."}

:::callout{type="abstract" title="Abstract"}
**Objective** To find out how accurately emergency paediatricians diagnose community-acquired pneumonia in children with lung ultrasound, against chest radiography read by two radiologists and a review on day 14.

**Methods** We enrolled 412 children aged 3 months to 15 years who came to one emergency department with fever and cough or fast breathing. One of nine paediatricians scanned each child before the radiograph, after a six-hour course.

**Results** Pneumonia was diagnosed in 139 children (33.7%). Ultrasound had a sensitivity of 91.4% (95% CI 85.5 to 95.0) and a specificity of 95.2% (92.0 to 97.2), so LR+ = 19.19 and LR− = 0.09. A median scan took 6 minutes.

**Conclusions** After a short course, emergency paediatricians rule pneumonia in or out with ultrasound about as well as a radiograph does, and without radiation.
:::

## Introduction

Pneumonia is the commonest serious infection of childhood and the reason for many of the chest radiographs taken in children’s emergency departments. British guidelines advise against a routine radiograph in a child well enough to go home [@harris2011], yet the film is often asked for to settle a doubtful examination, and two radiologists reading the same film do not always agree.

Lung ultrasound looks at the same disease through another window. Consolidated lung carries sound and shows as a solid, liver-like area with bright air bronchograms, signs that were described and standardised in adults first [@lichtenstein2008; @volpicelli2012]. In children the thin chest wall makes the scan easier, and a meta-analysis and a large prospective study found it both sensitive and specific [@pereda2015; @shah2013]. Most of those scans, however, were done by sonographers or by a few physicians with long experience, and it is not known whether the figures hold when the scan is one task among many on a night shift.

We measured the accuracy of lung ultrasound done by the emergency paediatricians of a district hospital after a short course, and report it as STARD 2015 asks [@bossuyt2015].

## Methods

### Participants

From 1 October 2024 to 30 November 2025 we approached children aged 3 months to 15 years who came to the emergency department of a district hospital of 280 beds with a temperature of 38 °C or more and cough, fast breathing or chest indrawing. Children with chronic lung disease, congenital heart disease or a radiograph taken elsewhere in the previous week were not included. Parents gave written consent, and the regional ethics committee approved the protocol. Data were entered in REDCap [@harris2009].

### Index test

Nine paediatricians, none of whom had used lung ultrasound before, attended a course of six hours, two of them spent scanning under supervision. Each scan covered six zones on each side of the chest with a linear probe of 8 to 12 MHz and was recorded as positive when a consolidation of 1 cm or more with air bronchograms was seen under the pleura. The scan was done before the radiograph, and the paediatrician entered the result before seeing the film.

### Reference standard

Two radiologists who did not know the ultrasound result read every radiograph, and a third settled disagreements. The final diagnosis joined the agreed reading to a telephone review on day 14, as STARD advises when no single test is a gold standard [@bossuyt2015]. A child with a normal film who was given antibiotics for pneumonia by the family doctor within the 14 days was counted as having pneumonia.

### Statistical analysis

Sensitivity and specificity are given with Wilson intervals [@newcombe1998], and likelihood ratios with the intervals of Simel and colleagues [@simel1991]. We read a positive likelihood ratio above 10, or a negative one below 0.1, as a large change in the probability of disease [@altman1994; @deeks2004]. Agreement was measured with Cohen’s kappa and read on the scale of Landis and Koch [@landis1977]. We planned 400 children, so that with a prevalence of one in three the lower limit of the interval for a sensitivity of 90% would stay above 83%. A value of p < 0.05 was taken as significant.

## Results

Of 451 children screened, 412 were enrolled. Their mean age was 4.1 ± 3.2 years, 221 (53.6%) were boys and 97 (23.5%) were under 2 years. Pneumonia was diagnosed in 139 (33.7%), 14 of them only at the review on day 14.

Ultrasound was positive in 140 children and right in 127 of them. Its sensitivity was 91.4% (95% CI 85.5 to 95.0) and its specificity 95.2% (92.0 to 97.2), with a positive predictive value of 90.7% and a negative predictive value of 95.6%. The likelihood ratios were LR+ = 19.19 and LR− = 0.09, both past the limits set in advance. Sensitivity was lower in children under 2 years, 31 of 35 (88.6%), than in older children, 96 of 104 (92.3%), a difference of the size chance would give (p = 0.51). Sensitivity also rose with practice, from 87.5% in each paediatrician’s first 20 scans to 94.0% after them (p = 0.18).

The 12 false negatives were consolidations under 1 cm on the film (7), shadows behind the heart (3) and two films read as pneumonia by both radiologists in children whose fever settled without antibiotics. In a logistic model, a consolidation under 1 cm on the film was the only factor linked to a missed diagnosis (OR = 6.2, 95% CI 2.1 to 18.4; p < 0.001); an age under 2 years was not (OR = 1.6; p = 0.43). Between the paediatricians and an expert who reviewed 120 stored clips, kappa = 0.84; between the two radiologists, kappa = 0.71. The median scan took 6 minutes (interquartile range 4 to 8), and no child needed sedation.

## Discussion

With six hours of training, emergency paediatricians diagnosed pneumonia with ultrasound about as accurately as the experienced operators of earlier studies [@pereda2015; @shah2013], and agreed with an expert more closely than the radiologists agreed with each other. With LR+ = 19.19, a positive scan in a child whose pre-test probability is one in three raises it above 90%, and with LR− = 0.09 a negative scan brings it under 5%.

The false negatives show where the scan is weak: small consolidations and the lung behind the heart, which the probe cannot reach between the ribs. They also show the weakness of the reference standard, since two of them were probably shadows on the film. The study has other limits. It was done in one hospital, the paediatricians volunteered for the course, and a review by telephone on day 14 will have missed some mild cases.

The scan suits the place where the decision is taken. A child who would otherwise be sent to radiology can be scanned in the cubicle, and the radiograph kept for the child whose scan is doubtful or whose fever does not settle as expected.

## Article information {style="back"}

:::paragraphs{style="info"}
**Funding** None. **Competing interests** None declared. **Note** Set in STIX Two Text, IBM Plex Sans and IBM Plex Sans Condensed (SIL OFL). Text: original, CC BY 4.0. The study, its authors, the hospital and the results are invented for this example; the references are to real published works.
:::

## References {style="back"}

:::bibliography{title=""}

:::references{format=bibtex}
@article{harris2011,
  author = {Harris, Michael and Clark, Julia and Coote, Nicky and Fletcher, Penny and Harnden, Anthony and McKean, Michael and Thomson, Anne},
  title = {British Thoracic Society guidelines for the management of community acquired pneumonia in children: update 2011},
  journal = {Thorax}, year = 2011, volume = 66, number = {Suppl 2}, pages = {ii1--ii23}}
@article{landis1977,
  author = {Landis, J. Richard and Koch, Gary G.},
  title = {The measurement of observer agreement for categorical data},
  journal = {Biometrics}, year = 1977, volume = 33, number = 1, pages = {159--174}}
@article{lichtenstein2008,
  author = {Lichtenstein, Daniel A. and Mezière, Gilbert A.},
  title = {Relevance of lung ultrasound in the diagnosis of acute respiratory failure: the {BLUE} protocol},
  journal = {Chest}, year = 2008, volume = 134, number = 1, pages = {117--125}}
@article{volpicelli2012,
  author = {Volpicelli, Giovanni and Elbarbary, Mahmoud and Blaivas, Michael and Lichtenstein, Daniel A. and Mathis, Gebhard and Kirkpatrick, Andrew W. and Melniker, Lawrence},
  title = {International evidence-based recommendations for point-of-care lung ultrasound},
  journal = {Intensive Care Med}, year = 2012, volume = 38, number = 4, pages = {577--591}}
@article{pereda2015,
  author = {Pereda, Maria A. and Chavez, Miguel A. and Hooper-Miele, Christine C. and Gilman, Robert H. and Steinhoff, Mark C. and Ellington, Laura E. and Gross, Margaret},
  title = {Lung ultrasound for the diagnosis of pneumonia in children: a meta-analysis},
  journal = {Pediatrics}, year = 2015, volume = 135, number = 4, pages = {714--722}}
@article{shah2013,
  author = {Shah, Vaishali P. and Tunik, Michael G. and Tsung, James W.},
  title = {Prospective evaluation of point-of-care ultrasonography for the diagnosis of pneumonia in children and young adults},
  journal = {JAMA Pediatr}, year = 2013, volume = 167, number = 2, pages = {119--125}}
@article{bossuyt2015,
  author = {Bossuyt, Patrick M. and Reitsma, Johannes B. and Bruns, David E. and Gatsonis, Constantine A. and Glasziou, Paul P. and Irwig, Les and Lijmer, Jeroen G.},
  title = {{STARD} 2015: an updated list of essential items for reporting diagnostic accuracy studies},
  journal = {BMJ}, year = 2015, volume = 351, pages = {h5527}}
@article{harris2009,
  author = {Harris, Paul A. and Taylor, Robert and Thielke, Robert and Payne, Jonathon and Gonzalez, Nathaniel and Conde, Jose G.},
  title = {Research electronic data capture ({REDCap}): a metadata-driven methodology and workflow process for providing translational research informatics support},
  journal = {J Biomed Inform}, year = 2009, volume = 42, number = 2, pages = {377--381}}
@article{newcombe1998,
  author = {Newcombe, Robert G.},
  title = {Two-sided confidence intervals for the single proportion: comparison of seven methods},
  journal = {Stat Med}, year = 1998, volume = 17, number = 8, pages = {857--872}}
@article{simel1991,
  author = {Simel, David L. and Samsa, Gregory P. and Matchar, David B.},
  title = {Likelihood ratios with confidence: sample size estimation for diagnostic test studies},
  journal = {J Clin Epidemiol}, year = 1991, volume = 44, number = 8, pages = {763--770}}
@article{altman1994,
  author = {Altman, Douglas G. and Bland, J. Martin},
  title = {Diagnostic tests 1: sensitivity and specificity},
  journal = {BMJ}, year = 1994, volume = 308, number = 6943, pages = {1552}}
@article{deeks2004,
  author = {Deeks, Jonathan J. and Altman, Douglas G.},
  title = {Diagnostic tests 4: likelihood ratios},
  journal = {BMJ}, year = 2004, volume = 329, number = 7458, pages = {168--169}}
:::
`;

// ─── 3 · Fonts ──────────────────────────────────────────────────────────────
const FONTS = { 'STIX Two Text': ['400', '400i', '600'], 'IBM Plex Sans': ['400', '600'],
  'IBM Plex Sans Condensed': ['500', '600'] };

// ─── 4 · Build & show ───────────────────────────────────────────────────────
// #region marks: each break a build reported, marked in the source of the second build
// The word before the citation and the citation, or the sign and the number on its other
// side, wrapped in :mark[…]: a band under ordinary text, so the breaks stay where they were.
const wordEnd = (s, i) => i + s.slice(i).search(/[\s;)]|[,.](?!\d)|$/); // 0.09, not 0
function sides(source, w, a, b) { // a warning names line a, before the break, by its source
  if (w.kind === 'citationBreak') { // "colleagues [@simel1991]"
    const cite = source.lastIndexOf('[@', b.sourceStart + 2);
    return [source.lastIndexOf(' ', cite - 2) + 1, source.indexOf(']', cite) + 1];
  }
  if (w.edge === 'end') { // "LR− = 0.09": the sign ends line a, the number opens line b
    const from = source.lastIndexOf(` ${w.text} `, a.sourceEnd) + 1;
    return [from, wordEnd(source, from + w.text.length + 1)];
  }
  const sign = source.indexOf(` ${w.text} `, b.sourceStart - 1); // the sign opens line b
  return [source.lastIndexOf(' ', sign - 1) + 1, sign + 1 + w.text.length];
}
function markBreaks(source, doc, found) {
  const pairs = doc.blocks.flatMap((b) => b.lines.map((line, i) => [line, b.lines[i + 1]]));
  const spans = new Map();
  for (const w of found) {
    const [a, b] = pairs.find(([line]) => line.sourceStart === w.sourceStart) ?? [];
    const [from, to] = b ? sides(source, w, a, b) : [];
    if (from > a?.sourceStart && to > from && to <= b.sourceEnd) spans.set(from, to);
  }
  return [...spans].sort(([x], [y]) => y - x).reduce((md, [from, to]) => `${md.slice(0, from)}`
    + `:mark[${md.slice(from, to)}]{style="edge"}${md.slice(to)}`, source);
}
// #endregion

// #region proofs: the same article twice, 'warn' then 'avoid', each with a query slip
const KIND = { citationBreak: t({ en: 'citation', es: 'cita' }),
  operatorBreak: t({ en: 'sign', es: 'signo' }) };
const edge = (s, end) => (end ? s.trim().split(/\s+/).slice(-2) : s.trim().split(/\s+/)
  .slice(0, 2)).join('\u00a0').replace(/[.,;]$/, ''); // no-break spaces: no breaks of its own
const query = (doc, w) => `${t({ en: 'p.', es: 'pág.' })}\u00a0${doc.pages[w.pageIndex]
  .pageLabel}, ${KIND[w.kind]}: …${edge(w.lines[0], true)}\u00a0/ ${edge(w.lines[1])}…`;
const slip = (mode, doc, found) => ['', `:::callout{type="slip" title="${mode === 'warn'
  ? t({ en: 'First proof: queries', es: 'Primeras pruebas: consultas' })
  : t({ en: 'Second proof: queries', es: 'Segundas pruebas: consultas' })}"}`,
`citationBreaks / operatorBreaks: '${mode}' · ${found.length} ${found.length === 1
  ? t({ en: 'break', es: 'corte' }) : t({ en: 'breaks', es: 'cortes' })}.`,
...found.flatMap((w) => ['', query(doc, w)]), ':::', ''].join('\n');
const build = (md, mode) => buildDocumentWithFonts({ markdown: md }, { ...config,
  bodyText: { ...bodyText, citationBreaks: mode, operatorBreaks: mode } }, kitFonts(FONTS));
async function proof(mode) {
  const plain = await build(markdown, mode); // pass 1: where the breaks are
  const found = (plain.contentWarnings ?? []).filter((w) => /^(citation|operator)Break$/
    .test(w.kind));
  return build(markBreaks(markdown, plain, found) + slip(mode, plain, found), mode); // pass 2
}
const docs = { warn: await proof('warn'), avoid: await proof('avoid') };
// #endregion

const TITLE = t({ en: 'Citations and statistics off the line edges',
  es: 'Citas y estadísticos fuera de los bordes de línea' });
const show = (key) => showPages(docs[key], { title: `${TITLE} · '${key}'` });
const button = (key) => Object.assign(document.createElement('button'),
  { type: 'button', textContent: key, onclick: () => show(key) });
document.getElementById('pt-actions').prepend(button('warn'), button('avoid'));
show('warn');

// ─── Kit ── helpers shared by every Cookbook recipe · postext.dev/cookbook ─────

// ─── Kit · core v1 ── the same in every recipe · postext.dev/cookbook
function mm(value) { return { value, unit: 'mm' }; }
function pt(value) { return { value, unit: 'pt' }; }
function em(value) { return { value, unit: 'em' }; }
/** The sample language's string: t({ en: 'Figure', es: 'Figura' }). */
function t(strings) { return strings[LANG] ?? Object.values(strings)[0]; }
/** A file in this recipe's assets folder, served from the Postext repo by jsDelivr. */
function asset(file) { return `https://cdn.jsdelivr.net/gh/drnachio/postext@main/cookbook/${RECIPE}/assets/${file}`; }

// ─── Kit · fonts v3 ── the same in every recipe · postext.dev/cookbook
// Postext loads the faces a config asks for before layout and asks the kit
// for Fontsource's files, the ones the PDF embeds too.

/** buildDocumentWithFonts' options: the resolver, and FONTS' faces loaded
 *  with the config's. FONTS = { 'Family Name': ['400', '400i', '700'] }. */
function kitFonts(FONTS = {}) {
  const faces = Object.entries(FONTS).flatMap(([family, specs]) => specs.map((spec) =>
    ({ family, weight: parseInt(spec, 10), style: spec.endsWith('i') ? 'italic' : 'normal' })));
  return { resolve: fontsourceResolver, faces };
}

/** A face's files with their ranges: latin, then latin-ext and greek when
 *  `text` needs them (kitSubsetsFor). Null for a face Fontsource lacks. */
async function fontsourceResolver(family, weight, style, { text = '' } = {}) {
  const id = fontsourceId(family);
  const meta = await fontsourceMeta(family);
  if (meta && !(await fontsourceWeights(family, style)).includes(weight)) return null;
  kitStatus('Loading fonts…');
  const at = await fontsourceAt(id);
  return ['latin', ...kitSubsetsFor(text, meta)].map((subset) => ({
    source: `${at}files/${id}-${subset}-${weight}-${style}.woff2`,
    unicodeRange: kitRanges()[subset],
  }));
}

/** Each file's unicode-range (a function: the kit runs after the pen's code). */
function kitRanges() {
  return {
    latin: 'U+0000-00FF,U+0131,U+0152-0153,U+02BB-02BC,U+02C6,U+02DA,U+02DC,U+0304,U+0308,U+0329,'
      + 'U+2000-206F,U+20AC,U+2122,U+2191,U+2193,U+2212,U+2215,U+FEFF,U+FFFD',
    'latin-ext': 'U+0100-02BA,U+02BD-02C5,U+02C7-02CC,U+02CE-02D7,U+02DD-02FF,U+0304,U+0308,U+0329,'
      + 'U+1D00-1DBF,U+1E00-1E9F,U+1EF2-1EFF,U+2020,U+20A0-20AB,U+20AD-20C0,U+2113,U+2C60-2C7F,U+A720-A7FF',
    greek: 'U+0370-03FF',
  };
}

/** For the cjk, arabic and comics blocks: loads `faces` (as FONTS) for the
 *  sample `text`; `optional` skips a face Fontsource lacks. Files are added in
 *  the order asked, the last winning a shared character. Resolves to the count. */
async function loadFonts(faces, text = '', { optional = false } = {}) {
  kitStatus('Loading fonts…');
  const ranges = kitRanges();
  const jobs = [];
  for (const [family, specs] of Object.entries(faces)) {
    const id = fontsourceId(family);
    const todo = [...new Set(specs)].map((spec) => [parseInt(spec, 10), spec.endsWith('i') ? 'italic' : 'normal'])
      .filter(([weight, style]) => !hasFace(family, weight, style)); // before any await
    const at = fontsourceAt(id);
    const meta = optional || /[^\0-ÿ]/u.test(text) ? await fontsourceMeta(family) : null;
    const subsets = ['latin', ...kitSubsetsFor(text, meta)];
    for (const [weight, style] of todo) {
      if (optional && !(meta && (await fontsourceWeights(family, style)).includes(weight))) continue;
      for (const subset of subsets) {
        const url = `${await at}files/${id}-${subset}-${weight}-${style}.woff2`;
        const face = new FontFace(family, `url(${url}) format('woff2')`,
          { weight: String(weight), style, unicodeRange: ranges[subset] });
        jobs.push(face.load().catch(() => {
          if (subset === 'latin' && !optional) throw new Error(`Fontsource has no ${family} ${weight} ${style}`);
        }));
      }
    }
  }
  const ready = await Promise.all(jobs).catch((error) => { kitFail(error); throw error; });
  return ready.filter((face) => face && document.fonts.add(face)).length;
}

/** A loaded FontFace covers this face (fonts.check() says yes to any family). */
function hasFace(family, weight, style) {
  for (const face of document.fonts) {
    if (face.status !== 'loaded' || face.style !== style) continue;
    if (face.family.replace(/^["']|["']$/g, '') !== family) continue;
    const [low, high = low] = face.weight.split(' ').map(Number);
    if (weight >= low && weight <= high) return true;
  }
  return false;
}

/** The files beyond latin `text` needs that `meta`'s family ships. */
function kitSubsetsFor(text, meta) {
  return [[/[Ā-˿ᴀ-ᶿḀ-ỿ†ℓⱠ-Ɀ꜠-ꟿ]/u, 'latin-ext'], [/[Ͱ-Ͽ]/u, 'greek']]
    .filter(([re, x]) => re.test(text) && meta?.subsets?.includes(x)).map(([, x]) => x);
}

/** Fontsource's id for a family: 'Source Serif 4' → 'source-serif-4'. */
function fontsourceId(family) { return family.toLowerCase().replace(/\s+/g, '-'); }

/** A package's folder at the one version @5 names now (#766). */
function fontsourceAt(id, u = `https://cdn.jsdelivr.net/npm/@fontsource/${id}@`) {
  return (fontsourceAt.c ??= {})[id] ??= fetch(u + '5/package.json').then((r) => r.json())
    .then((p) => p.version || 5, () => 5).then((v) => `${u}${v}/`);
}

/** The family's Fontsource metadata (weights, styles, subsets), or null. */
function fontsourceMeta(family, id = fontsourceId(family)) {
  return (fontsourceMeta.c ??= {})[id] ??= fetch(`https://api.fontsource.org/v1/fonts/${id}`)
    .then((res) => (res.ok ? res.json() : null), () => null);
}

/** The weights the package ships in `style`: the API pairs none (no Libre
 *  Caslon 700 italic file), so italics are read from latin-italic.css. */
function fontsourceWeights(family, style, id = fontsourceId(family)) {
  return (fontsourceWeights.c ??= {})[id + style] ??= fontsourceMeta(family).then(async (m) => {
    if (!m?.styles.includes(style)) return m ? [] : [400, 700];
    const css = style === 'italic' ? await fetch(`${await fontsourceAt(id)}latin-italic.css`).then((r) => (r.ok ? r.text() : ''), () => '') : '';
    const w = [...css.matchAll(/weight: (\d+)/g)].map((x) => +x[1]);
    return w.length ? w : m.weights;
  });
}

// ─── Kit · viewer v2 ── the same in every recipe · postext.dev/cookbook
/** The pages as spreads on a dark desk, page 1 alone, then verso | recto,
 *  each painted when it scrolls near. Other options go to renderPageToCanvas. */
function showPages(docs, { title, width = 460, ...render } = {}) {
  const root = viewer(title);
  const pages = [docs].flat().flatMap((doc) =>
    doc.pages.map((page) => ({ doc, page, n: (doc.pageIndexOffset ?? 0) + page.index })));
  const spreads = [];
  let verso = null;
  for (const p of pages) {
    if (p.n % 2 === 1) { if (verso) spreads.push([verso, null]); verso = p; }
    else { spreads.push([verso, p]); verso = null; }
  }
  if (verso) spreads.push([verso, null]);
  const density = Math.min(window.devicePixelRatio || 1, 2);
  showPages.painter?.disconnect();
  const painter = new IntersectionObserver((entries) => {
    for (const { isIntersecting, target } of entries) {
      if (!isIntersecting) continue;
      painter.unobserve(target);
      const { doc, page } = target.postext;
      renderPageToCanvas(page, doc, target, { ...render, scale: (width * density) / page.width });
    }
  }, { rootMargin: '800px' });
  showPages.painter = painter;
  root.replaceChildren(...spreads.map((pair) => {
    const spread = document.createElement('div');
    spread.className = 'pt-spread';
    for (const p of pair) {
      const figure = document.createElement('figure');
      if (p) {
        const label = p.page.pageLabel || String(p.n + 1);
        const canvas = document.createElement('canvas');
        canvas.postext = p;
        canvas.style.aspectRatio = `${p.page.width} / ${p.page.height}`;
        canvas.setAttribute('role', 'img');
        canvas.setAttribute('aria-label', `Page ${label}`);
        const folio = document.createElement('figcaption');
        folio.textContent = label;
        figure.append(canvas, folio);
        painter.observe(canvas);
      } else figure.className = 'pt-blank';
      spread.append(figure);
    }
    return spread;
  }));
  kitStatus(`${pages.length} ${pages.length === 1 ? 'page' : 'pages'}`);
  document.documentElement.dataset.postext = 'ready';
  return pages.length;
}

/** The desk, the bar and the error reporting, created once. */
function viewer(title) {
  if (!document.getElementById('pt-kit')) {
    document.head.insertAdjacentHTML('beforeend', `<style id="pt-kit">
      :root { color-scheme: dark; }
      body { margin: 0; background: #0e1014; color: #b9bcc4; font: 13px/1.45 system-ui, sans-serif; }
      #pt-bar { position: sticky; top: 0; z-index: 1; display: flex; flex-wrap: wrap; align-items: center;
        gap: 6px 16px; padding: 10px 16px; background: rgb(14 16 20 / .92); backdrop-filter: blur(6px);
        border-bottom: 1px solid #23262d; }
      #pt-bar strong { color: #f4f1ea; font-weight: 600; }
      #pt-actions { display: flex; gap: 12px; margin-left: auto; }
      #pt-actions a, #pt-actions button { color: #d8a21a; font: inherit; background: none; border: 0; padding: 0; cursor: pointer; }
      #pages { display: grid; justify-items: center; gap: 48px; padding: 32px 16px 72px; }
      .pt-spread { display: flex; }
      .pt-spread figure { margin: 0; width: min(460px, 44vw); }
      .pt-spread canvas { display: block; width: 100%; background: #fff;
        box-shadow: 0 1px 2px rgb(0 0 0 / .5), 0 22px 44px -16px rgb(0 0 0 / .8); }
      .pt-spread figure:first-child canvas { box-shadow: inset -14px 0 14px -14px rgb(0 0 0 / .18), 0 1px 2px rgb(0 0 0 / .5), 0 22px 44px -16px rgb(0 0 0 / .8); }
      .pt-spread figcaption { margin-top: 10px; text-align: center; font: 600 10px/1 system-ui, sans-serif;
        letter-spacing: .18em; text-transform: uppercase; color: #6c7079; }
      .pt-blank { visibility: hidden; }
      @media (max-width: 760px) {
        .pt-spread { flex-direction: column; gap: 32px; }
        .pt-spread figure { width: min(460px, 92vw); }
        .pt-blank { display: none; }
      }
    </style>`);
    document.body.insertAdjacentHTML('afterbegin',
      '<header id="pt-bar"><strong id="pt-title"></strong><span id="pt-status" role="status"></span><span id="pt-actions"></span></header>');
    document.getElementById('pt-title').textContent = document.title || 'Postext';
    addEventListener('error', (event) => kitFail(event.error ?? event.message));
    addEventListener('unhandledrejection', (event) => kitFail(event.reason));
  }
  if (title) document.getElementById('pt-title').textContent = title;
  return document.getElementById('pages')
    ?? document.body.appendChild(Object.assign(document.createElement('main'), { id: 'pages' }));
}

function kitStatus(text) {
  viewer();
  document.getElementById('pt-status').textContent = text;
}

function kitFail(error) {
  document.documentElement.dataset.postext = 'error';
  kitStatus(`Error: ${error?.message ?? error}`);
}

// ─── /Kit ───────────────────────────────────────────────────────────────────────
```

## Variantes

### Signaler les coupures sans rien changer

Une revue qui vérifie ses épreuves à la main peut laisser la règle à `'warn'` : le Sandbox liste chaque coupure parmi ses avertissements, et les lignes restent telles que l’algorithme les a coupées.

```diff
-  citationBreaks: 'avoid', operatorBreaks: 'avoid',
+  citationBreaks: 'warn', operatorBreaks: 'warn',
```

### Garder la pénalité par défaut

À 10000, l’espacement reste un peu plus serré et il reste deux coupures par édition au lieu d’une.

```diff
-  citationBreakPenalty: 30000, operatorBreakPenalty: 30000,
```

## Erreurs fréquentes

- **Avertissement de mise en page: Citation en début de ligne** (`citationBreak`). Une ligne d’un paragraphe commence par une citation numérotée placée après un mot, ou par un morceau de celle-ci : « auteurs » en fin de ligne et « [12] » en tête de la suivante, ou « [14, » au-dessus de « 16] ». Un nombre entre crochets ou parenthèses après un nombre n’est pas une citation (« 2016 (3) »), ni un mot entre crochets. Le texte de droite à gauche et le chinois, le japonais ou le coréen sont laissés tels quels. Avec bodyText.citationBreaks sur 'warn', chaque coupure de ce genre est signalée ; sur 'avoid', celles que la coupure des lignes a gardées parce qu’aucune composition dans l’espacement des mots ne les évitait pour moins que bodyText.citationBreakPenalty. Le moteur note chacune dans doc.contentWarnings avec la citation, les deux lignes et leur page. Solution: Réglez bodyText.citationBreaks sur 'avoid' : la coupure des lignes donne un prix à une coupure avant une citation ou à l’intérieur et recompose le paragraphe là où l’espacement des mots le permet ; l’espace continue de s’étirer avec les autres. Une coupure qui reste coûte plus à éviter que bodyText.citationBreakPenalty, souvent parce que la citation est large pour la justification : augmentez la pénalité pour l’échanger contre des lignes plus lâches, ou écrivez une espace insécable (U+00A0) entre le mot et la citation pour lier cette paire quelle que soit la règle. Dans le portage d’un livre existant, gardez 'warn' : le portage reproduit les coupures de l’original. ([Documentation](https://postext.dev/fr/docs/configuration-text.md#texte-courant))
- **Avertissement de mise en page: Signe en bord de ligne** (`operatorBreak`). Une ligne d’un paragraphe finit ou commence par un signe isolé entre espaces, l’un de = < > ≤ ≥ ± × ≈ ≠ ÷ − + · → : « LR+ = » en fin de ligne avec « 18,29 » sur la suivante, ou « ± 0,03 » en tête de ligne. Un trait d’union entre deux mots n’est pas un signe. Le texte de droite à gauche et le chinois, le japonais ou le coréen sont laissés tels quels. Avec bodyText.operatorBreaks sur 'warn', chaque coupure de ce genre est signalée, une par coupure (une ligne qui commence et finit par un signe en donne deux) ; sur 'avoid', celles que la coupure des lignes a gardées parce qu’aucune composition dans l’espacement des mots ne les évitait pour moins que bodyText.operatorBreakPenalty. Le moteur note chacune dans doc.contentWarnings avec le signe, le bord (fin ou début), les deux lignes et leur page. Solution: Réglez bodyText.operatorBreaks sur 'avoid' : la coupure des lignes donne un prix à une coupure aux espaces de part et d’autre du signe et recompose le paragraphe là où l’espacement des mots le permet. Une coupure qui reste coûte plus à éviter que bodyText.operatorBreakPenalty, souvent dans une colonne étroite où le signe et ses deux termes sont larges : augmentez la pénalité pour l’échanger contre des lignes plus lâches, ou écrivez des espaces insécables (U+00A0, ou U+202F pour une fine) autour du signe pour le lier quelle que soit la règle. Dans le portage d’un livre existant, gardez 'warn' : le portage reproduit les coupures de l’original. ([Documentation](https://postext.dev/fr/docs/configuration-text.md#texte-courant))

- Les coupures se cherchent dans les lignes déjà coupées. Un changement de format, de corps ou de texte les déplace : lisez les avertissements de chaque composition au lieu de recopier les positions d’une composition antérieure.
- Pour garder un appel précis avec son mot quelle que soit la colonne, tapez une espace insécable devant lui dans le Markdown. La règle s’occupe du reste de l’article.

## Crédits

- Recette: Ignacio Ferro ([@drnachio](https://github.com/drnachio))
- Caractères: STIX Two Text (OFL-1.1), IBM Plex Sans (OFL-1.1), IBM Plex Sans Condensed (OFL-1.1)
- Code: MIT · Contenu d’exemple: CC-BY-4.0

## Recettes liées

- [Nº 184 · Une prose justifiée sans répétitions en bord de ligne](https://postext.dev/fr/cookbook/prose-without-stacks.md): Une page de Dickens composée deux fois dans une justification étroite : la première épreuve signale chaque répétition, la seconde la pèse dans la coupure. · Niveau 2 (Intermédiaire) · Fiction, théâtre et prose littéraire
- [Nº 095 · Un article médical au style Vancouver](https://postext.dev/fr/cookbook/medical-article-vancouver.md): Le compte rendu d’un essai clinique en A4 sur deux colonnes : appels numériques en exposant, bibliographie Vancouver numérotée et DOI cliquables dans le PDF. · Niveau 2 (Intermédiaire) · Articles et publications universitaires
- [Nº 014 · Des espaces réguliers en colonnes étroites justifiées](https://postext.dev/fr/cookbook/justification-lab.md): Coupure Knuth–Plass, espaces tenues entre 0,8 et 1,6 fois la normale, face à un témoin glouton aux lignes lâches, isolées et trop courtes repérées par le VDT. · Niveau 3 (Avancé) · Magazines et fanzines
