README

GNU gettext PO file filter for the CAT Framework.

Installation

composer require catframework/filter-po

Usage

use CatFramework\FilterPo\PoFilter;

$filter = new PoFilter();

// Extract translatable segments from a PO file
$document = $filter->extract('messages.po', 'en', 'fr');

foreach ($document->getSegmentPairs() as $pair) {
    echo $pair->source->getPlainText() . PHP_EOL;
    // … send to MT, TM lookup, or human translator …
    $pair->target = new Segment('seg-t', [$translatedText]);
}

// Write the translated PO file
$filter->rebuild($document, 'messages.fr.po');

What gets extracted

PO escape sequences (\n, \t, \", \\) are decoded on extraction and re-encoded on rebuild.

Skeleton format

The skeleton stored in BilingualDocument::$skeleton is:

[
    'lines'   => string[],  // full file split by "\n", msgstr blocks replaced by tokens
    'seg_map' => [          // segId => token string
        'seg-1' => '{{SEG:001}}',
        'seg-2' => '{{SEG:002}}',
        // …
    ],
]

Limitations

• Plural translations: only msgstr[0] is filled; msgstr[1..n] require manual post-editing.
• msgctxt (message context) is recognised and skipped over during parsing but not exposed as metadata on the segment.
• Binary PO files (.mo) are not supported.