README

HTML file filter for the CAT Framework.

Installation

composer require catframework/filter-html

Requires ext-dom and ext-libxml.

Usage

use CatFramework\FilterHtml\HtmlFilter;

$filter = new HtmlFilter();

// Extract translatable segments
$document = $filter->extract('page.html', 'en', 'fr');

foreach ($document->getSegmentPairs() as $pair) {
    $pair->target = new Segment('seg-t', [$translatedText]);
}

// Write the translated file
$filter->rebuild($document, 'page.fr.html');

What gets extracted

The filter uses a block element taxonomy to decide segmentation boundaries:

Block elements (each becomes at most one segment): <p>, <div>, <h1>–<h6>, <li>, <td>, <th>, <dt>, <dd>, <blockquote>, <figcaption>, <caption>

A block element with only text / inline children → extracted as one segment.
A block element that itself contains other block elements → recursed into (not extracted as a whole).

Inline elements inside a segment become InlineCode pairs so translators see placeholders rather than raw HTML tags: <b>, <strong>, <i>, <em>, <a>, <span>, <sub>, <sup>, <code>, <abbr>, <u>, <small>, <mark>

Void elements (<br>, <img>, <input>, etc.) become standalone InlineCode placeholders.

Whitespace-only blocks are silently skipped.

Skeleton format

[
    'html'    => string,   // serialized DOMDocument with {{SEG:NNN}} tokens in place of block content
    'seg_map' => [         // segId => token string
        'seg-1' => '{{SEG:001}}',
        // …
    ],
]

Limitations

Full HTML documents only: the filter expects a <body> element. Fragment-only strings (no wrapping body) will produce no segments.
Structural elements outside the body (<head>, <title>, <meta>) are not extracted.
Unknown non-block elements are treated as inline and wrapped as InlineCode pairs.
Invalid nesting (block element inside an inline context) is silently ignored.

opencat / filter-html

Maintainers

Package info

Statistics

Security