<?php

// Load the TMX file into SimpleXML

$tmxFile = 'path/to/your/file.tmx';

$xml = simplexml_load_file($tmxFile);

// Check if the XML was loaded successfully

if ($xml === false) {

    die('Error loading TMX file.');

}

// Loop through translation units

foreach ($xml->body->tu as $tu) {

    // Extract source and target segments

    $sourceSegment = (string)$tu->tuv[0]->seg;

    $targetSegment = (string)$tu->tuv[1]->seg;

    // Extract tags

    $sourceTags = extractTags($sourceSegment);

    $targetTags = extractTags($targetSegment);

    // Output tags for each translation unit

    echo "Source Tags: " . implode(', ', $sourceTags) . PHP_EOL;

    echo "Target Tags: " . implode(', ', $targetTags) . PHP_EOL;

}

/**

 * Function to extract tags from a segment.

 *

 * @param string $segment

 * @return array

 */

function extractTags($segment) {

    $tags = [];

    preg_match_all('/<([^>]+)>/', $segment, $matches);

    if (!empty($matches[1])) {

        $tags = $matches[1];

    }

    return $tags;

}