Arabic, Hebrew and Hindi

Arabic, Hebrew and Hindi with Latin words and numbers inside them, a table laid out right to left, and six scripts justified with kashidas, the stretched joins of Arabic-style writing. Inter falls back to Noto fonts for the scripts it lacks, and the text is shaped with HarfBuzz (Tagua.Shaping.HarfBuzz).

Standards: PDF/A-2b · PDF/UA-1, checked with veraPDF on every change. Download the PDF.

What it shows

• The Unicode Bidirectional Algorithm, with embedded numbers
• A right-to-left table
• Font fallback from Inter to Noto
• Kashida justification: Arabic, Syriac, N’Ko, Mandaic, Adlam, Mongolian
• Each block tagged with its language (PDF/UA-1)

The template

multilingual.report.yaml

# yaml-language-server: $schema=../schema/template.schema.json
# Right-to-left and complex scripts: Arabic, Hebrew and Hindi with embedded Latin words and numbers,
# a right-to-left table, Arabic, Syriac, N'Ko, Mandaic, Adlam and Mongolian justified with kashidas, and
# PDF/A-2b + PDF/UA-1. Inter falls back to Noto fonts for
# the scripts it lacks. Rendered with the HarfBuzz shaper (Tagua.Shaping.HarfBuzz).
version: 1
title: Multilingual text
culture: en-GB

document:
  author: Acme Ltd
  compliance: [pdf-a-2b, pdf-ua-1]

page:
  size: A4
  margins: 22mm

fonts:
  Inter:
    regular: fonts/Inter-Regular.ttf
    bold: fonts/Inter-Bold.ttf
    fallback: [NotoArabic, NotoHebrew, NotoDevanagari, NotoSyriac, NotoNKo, NotoMandaic, NotoAdlam, NotoMongolian]
  NotoArabic:     { regular: fonts/NotoSansArabic-Regular.ttf }
  NotoHebrew:     { regular: fonts/NotoSansHebrew-Regular.ttf }
  NotoDevanagari: { regular: fonts/NotoSansDevanagari-Regular.ttf }
  NotoSyriac:     { regular: fonts/NotoSansSyriac-Regular.ttf }
  NotoNKo:        { regular: fonts/NotoSansNKo-Regular.ttf }
  NotoMandaic:    { regular: fonts/NotoSansMandaic-Regular.ttf }
  NotoAdlam:      { regular: fonts/NotoSansAdlam-Regular.ttf }
  NotoMongolian:  { regular: fonts/NotoSansMongolian-Regular.ttf }

styles:
  default: { font: Inter, fontSize: 11pt, lineHeight: 1.6, color: "#1F1F1F" }
  label:   { fontSize: 8.5pt, color: "#6B6B6B", padding: 10pt 0 2pt 0 }
  th:      { fontWeight: bold, borderBottom: "0.75pt solid #1F1F1F", padding: 0 0 3pt 0 }
  total:   { fontWeight: bold, borderTop: "0.75pt solid #1F1F1F", padding: 3pt 0 0 0 }

data:
  notice:
    fields: { Arabic: string, Hebrew: string, Hindi: string, Syriac: string, Nko: string, Mandaic: string, Adlam: string, Mongolian: string, Invoice: string }
  items:
    fields: { Item: string, Quantity: int32, Price: decimal }

body:
  dataset: notice

  detail:
    - content:
        - { type: text, value: "Right-to-left and complex scripts", fontSize: 18pt, fontWeight: bold, tag: h1 }
        - type: text
          value: >-
            Each paragraph takes its direction from its first strong character. Arabic and Hindi are
            shaped with HarfBuzz; words and numbers in other scripts keep their own order.

    - content:
        - { type: text, value: "Arabic, justified with kashidas", style: label }
    - layout: row          # right to left: the narrow paragraph sits at the right
      textDirection: rtl
      content:
        - { type: text, value: "{Arabic}", language: ar, align: justify, width: 330pt }
    - content:
        - { type: text, value: "Hebrew", style: label }
        - { type: text, value: "{Hebrew}", language: he }
        - { type: text, value: "Hindi", style: label }
        - { type: text, value: "{Hindi}", language: hi }
        - { type: text, value: "Syriac and N'Ko, justified with kashidas (tatweel and lajanyalan)", style: label }
    - layout: row
      textDirection: rtl
      gap: 24pt
      content:
        - { type: text, value: "{Syriac}", language: aii, align: justify, width: 1fr }
        - { type: text, value: "{Nko}", language: nqo, align: justify, width: 1fr }
    - content:
        - { type: text, value: "Mandaic, Adlam and Mongolian (set horizontally), justified with kashidas", style: label }
    - layout: row
      textDirection: rtl
      gap: 24pt
      content:
        - { type: text, value: "{Mandaic}", language: myz, align: justify, width: 1fr }
        - { type: text, value: "{Adlam}", language: ff-Adlm, align: justify, width: 1fr }
        # Mongolian runs left to right: undo the row's inherited rtl for its words.
        - { type: text, value: "{Mongolian}", language: mn-Mong, align: justify, width: 1fr, textDirection: auto }

    - padding: 16pt 0 0 0
      content:
        - { type: text, value: "A right-to-left table (textDirection: rtl)", style: label }
        - type: table
          dataset: items
          textDirection: rtl
          language: ar
          headerStyle: th
          columns:
            - { header: الصنف, value: "{Item}", width: 1fr }
            - { header: الكمية, value: "{Quantity}", width: 60pt, align: end }
            - { header: السعر, value: "{Price:N2}", width: 80pt, align: end }
          footer:
            - style: total
              cells:
                - { value: المجموع, span: 2, isHeader: true }
                - { value: "{Sum(Quantity * Price):N2}", align: end }
        - { type: text, value: "Invoice {Invoice} · فاتورة {Invoice} · חשבונית {Invoice}", style: label }

  pageFooter:
    - content:
        - { type: text, value: "Page {PageNumber} of {TotalPages}", align: right, color: "#6B6B6B", fontSize: 8pt }

Render it yourself

Save the template, put the fonts it names next to it (Inter and the Noto fonts are free, under the SIL Open Font License), and run:

tagua render multilingual.report.yaml --open

Without a data file, tagua render makes up sample rows from the field types; see Preview as you edit.