PHP Requires Chilkat v11.7.0+

Expand Ligatures and Normalize Unicode in Extracted PDF Text

See more PDF Text Extraction Examples

Demonstrates the "expandLigatures" and "normalize" clean-up options. Typeset PDFs use single glyphs for "fi" and "fl"; by default they are expanded to their letters. Unicode normalization is NFC by default; NFKC also folds compatibility characters such as circled digits, the trademark sign and vulgar fractions into plain characters, and "none" leaves the text as the PDF has it. Chilkat2 PHP

<?php

use Chilkat2\JsonObject;
use Chilkat2\Pdf;

// This example requires the Chilkat API to have been previously unlocked.
// See Global Unlock Sample for sample code.

// The PDF used in this example is available at ligatures.pdf

$pdf = new Pdf();
try {
    $pdf->LoadFile('qa_data/pdf/ligatures.pdf');

    $json = new JsonObject();
    $jsonOut = new JsonObject();

    // The first line is set with the fi and fl ligature glyphs; the second has circled digits,
    // a trademark sign and a vulgar fraction.
    $pdf->ExtractText('text', $json, $jsonOut);
} catch (\Chilkat2\Exception $e) {
    echo $e->getMessage(), "\n";
    exit;
}

echo "---- default (expandLigatures: true, normalize: NFC) ----\n";
echo $jsonOut->StringOf('text'), "\n";

// Keep the ligature characters as they are in the PDF.
$json->UpdateBool('expandLigatures', false);
$json->UpdateString('normalize', 'none');
try {
    $pdf->ExtractText('text', $json, $jsonOut);
} catch (\Chilkat2\Exception $e) {
    echo $e->getMessage(), "\n";
    exit;
}

echo "---- expandLigatures: false, normalize: none ----\n";
echo $jsonOut->StringOf('text'), "\n";

// NFKC folds the compatibility characters too (and expands ligatures on its own).
$json->UpdateString('normalize', 'NFKC');
try {
    $pdf->ExtractText('text', $json, $jsonOut);
} catch (\Chilkat2\Exception $e) {
    echo $e->getMessage(), "\n";
    exit;
}

echo "---- normalize: NFKC ----\n";
echo $jsonOut->StringOf('text'), "\n";

// Sample output (shown here in ASCII: (fi) and (fl) stand for the single ligature characters, (1) and (2) for the circled digits,
// (TM) for the trademark sign and 1/2 for the vulgar fraction; the NFKC output has plain "1", "2", "TM" and "1", fraction slash, "2"):

// ---- default (expandLigatures: true, normalize: NFC) ----
// Typeset Text
//
// The final flight of the season was officially confirmed on the fifth. Rooms (1) and (2) are reserved; the Harbor Cafe(TM) serves 1/2 portions.
// ---- expandLigatures: false, normalize: none ----
// Typeset Text
//
// The (fi)nal (fl)ight of the season was of(fi)cially con(fi)rmed on the (fi)fth. Rooms (1) and (2) are reserved; the Harbor Cafe(TM) serves 1/2 portions.
// ---- normalize: NFKC ----
// Typeset Text
//
// The final flight of the season was officially confirmed on the fifth. Rooms 1 and 2 are reserved; the Harbor CafeTM serves 1/2 portions.