Skip to content
phpdftk

phpdftk/text

UAX #14 line breaking, UAX #9 bidi, OpenType GSUB/GPOS shaping. Pure PHP, no dependencies beyond ext-mbstring and ext-intl (for Unicode property lookups). The substrate phpdftk/html-to-pdf and phpdftk/svg-to-pdf use for line-fitting and glyph layout.

Terminal window
composer require phpdftk/text

Quick example

Section titled “Quick example”

Line breaking

Section titled “Line breaking”
use Phpdftk\Text\LineBreaker;
use Phpdftk\Text\LineBreakKind;


$breaker = new LineBreaker();
foreach ($breaker->breaks('The quick brown fox jumps') as $opp) {
    $opp->position;          // codepoint offset
    $opp->kind;              // LineBreakKind::Mandatory | Allowed | Prohibited
}

Bidi reordering

Section titled “Bidi reordering”
use Phpdftk\Text\Bidi;


$bidi = new Bidi();
$result = $bidi->analyze("Hello שלום world");
foreach ($result->runs as $run) {
    $run->text;              // segment
    $run->level;             // resolved embedding level (0 = LTR base, 1 = RTL)
    $run->start;             // codepoint offset in original
}

Shaping (with a font)

Section titled “Shaping (with a font)”
use Phpdftk\Text\LineBreaker;
// Shaping requires a font + glyph IDs — see html-to-pdf's text pipeline
// for how the shaper is wired in. The ShapedGlyph / ShapedRun structs
// are the per-line output.

What’s in the box

Section titled “What’s in the box”

UAX #14 line breaking — Full Unicode line-break property table baked in. Class-pair lookup yields one of three opportunities at each position:

  • Mandatory (BK, LF, CR+LF, NL) — forced break.
  • Allowed — break opportunity, the line-fitter decides.
  • Prohibited — never break here.

Tailoring rules: Phpdftk\Text\LineBreaker::strictness() selects between Strict, Normal, and Loose (CSS Text 3 line-break property).

UAX #9 bidi — Resolution of paragraph embedding level, explicit isolates (LRI / RLI / FSI), implicit run reordering. Returns segmented runs at one resolved level each — input to the line-fitter for visual ordering.

OpenType shaping — GSUB lookups: ligatures (liga, clig, dlig), small caps (smcp, c2sc), alternates (salt), localised forms (locl), case mappings (case). GPOS lookups: kerning (kern), mark positioning (mark, mkmk). Font-feature tags can be turned on/off per-call via the shaper’s feature mask.

Word + cluster boundaries — Word-segmentation per UAX #29 for word-spacing distribution. Grapheme-cluster iteration for letter-spacing application (so spacing lands between clusters, not between code points within a cluster).

Why phpdftk/text

Section titled “Why phpdftk/text”
phpdftk/textsymfony/stringext-intl alone
UAX #14 line-break opportunities✓ With CSS strictness tailoring✗✗
UAX #9 bidi resolution + isolates✓✗~ Via IntlBreakIterator
OpenType GSUB/GPOS shaping✓✗✗
Grapheme-cluster iteration✓✓✓
Pure PHP (no native shaper extension)✓✓✓
Font-feature mask per call✓n/an/a

Status

Section titled “Status”

The text package is the substrate shared by phpdftk/html-to-pdf and phpdftk/svg-to-pdf for line-fitting and glyph layout. Its surface tracks the CSS Text 3, CSS Text 4, and OpenType specs as those features get wired into the renderers above.

Currently shipped:

  • UAX #14 line breaking with line-break: strict|normal|loose tailoring
  • UAX #9 bidi with isolates and overrides
  • OpenType GSUB ligature classes (liga / clig / dlig)
  • OpenType GPOS kerning and mark positioning
  • Word + grapheme-cluster segmentation
  • Word-spacing / letter-spacing distribution
  • Hyphenation hooks (consumer supplies the dictionary)

Phase 4D substrate (deferred):

  • Arabic Universal Shaping Engine (joining types, contextual forms)
  • Indic shaping (reordering classes, syllable detection)
  • Complex GSUB lookups beyond the ligature classes
  • Embedded WOFF font loading (currently the consumer provides parsed font data via phpdftk/font-parser)

Roadmap

Section titled “Roadmap”

Text-shaping substrate work lives in docs/plans/html-and-svg.md Phase 4D and the full-spec compliance plan.

☕ Support phpdftk