Parsing product descriptions for 1C-Bitrix content

Parsing Product Descriptions to Populate 1C-Bitrix

Empty or duplicated descriptions mean zero SEO value and poor conversion. But writing copy manually for 5,000 products is not feasible. Parsing descriptions from manufacturer or distributor websites is a practical way to populate a catalog quickly — provided the process is well organized and content duplication is avoided.

Sources for descriptions

The choice of source determines both quality and legal risk:

Manufacturer's website — the most relevant content, but often protected against scraping and may contain copyright. Use as a basis for rewriting, not as the final text.

Aggregators (Yandex Market, OZON, Wildberries) — large volume of descriptions, standardized format. The same legal caveats apply.

Authorized distributors — more permissive about content reuse, often actively interested in its distribution.

Manufacturer databases (Icecat, Synccentric) — a legal option with an API, paid, but provides clean, licensed data.

Extracting and cleaning text

The description on the source site typically lives inside <div class="description"> or a similar container. Extract it via DomCrawler:

$description = $crawler->filter('.product-description')->text();

After extraction, mandatory cleaning steps:

• strip_tags() to remove HTML (or filter allowed tags)
• Remove promotional inserts like "Buy in our store"
• Normalize whitespace and line breaks
• Remove competitor brand mentions

If HTML formatting must be preserved (bold, lists), use Symfony\Component\DomCrawler\Crawler::html() and filter through HTMLPurifier.

Writing to 1C-Bitrix fields

1C-Bitrix separates descriptions into two fields:

• PREVIEW_TEXT — short description (for listings)
• DETAIL_TEXT — full description (for the product card)

When parsing a long description: first paragraph → PREVIEW_TEXT, full text → DETAIL_TEXT. The text type is set via PREVIEW_TEXT_TYPE and DETAIL_TEXT_TYPE (values: text or html).

Updating an element:

$el = new CIBlockElement();
$el->Update($elementId, [
    'PREVIEW_TEXT' => $shortDesc,
    'PREVIEW_TEXT_TYPE' => 'html',
    'DETAIL_TEXT' => $fullDesc,
    'DETAIL_TEXT_TYPE' => 'html',
]);

Handling already populated cards

Do not overwrite blindly — managers may have manually improved descriptions. Add the following logic:

1. If DETAIL_TEXT is empty — write the parsed text
2. If it is populated — set the DESCRIPTION_SOURCE = parsed property only on initial population
3. On subsequent parser runs — skip elements without the DESCRIPTION_SOURCE flag (meaning the text was edited manually)

Work timeline

Total: 4–6 working days. If AI-based rewriting via GPT API is needed after parsing, add 1–2 more days for integration.