# kompress e-shop scrape

Snapshot of the live catalogue at <http://e-shop.kompress.sk> (PrestaShop 1.5/1.6),
taken 2026-07-29. Only the product catalogue is captured — no CMS pages, no orders.

Re-run with `python3 scrape.py` (stdlib only, no dependencies). It is idempotent:
already-downloaded images are skipped.

## Files

| File | Contents |
| --- | --- |
| `categories.json` | 30 categories, flat list ordered parent-before-child |
| `products.json` | 118 products, ordered by PrestaShop id |
| `images/<product_id>/<image_id>.jpg` | 181 product images, original uploads |
| `images/categories/<category_id>.jpg` | 30 category thumbnails, original uploads |

## Category schema

```jsonc
{
  "id": 8,                     // PrestaShop id_category, kept as the stable key
  "name": "Gáza, role, štvorce, prírezy",
  "title": "…",                // link title attr; often a longer marketing blurb
  "url": "http://e-shop.kompress.sk/8-gaza-role-…",
  "parent_id": 7,              // null for top-level
  "path": ["Zdravotnícke", "Gáza, role, štvorce, prírezy"],
  "depth": 1,                  // 0 = top level
  "description_html": "<p>…</p>",
  "description": "…",          // plain-text rendering
  "image": { "url": "…/img/c/8.jpg", "file": "images/categories/8.jpg" },
  "product_ids": [59, 62, …],
  "product_count": 8
}
```

The tree is 2 levels deep: 3 roots (`Zdravotnícke`, `Somatické`, `Pre invalidov`),
27 children. The roots hold a few products directly; `Somatické` and `Pre invalidov`
hold none of their own.

## Product schema

```jsonc
{
  "id": 73,                    // PrestaShop id_product
  "name": "Návlek na obuv",
  "url": "http://e-shop.kompress.sk/-netkane-textie-…/73-navlek-na-obuv.html",
  "reference": "6811",         // the "Kód:" on the page; "" when the shop has none
  "price": 0.04,               // tax included, EUR — this is the displayed price
  "price_display": "0,04 €",
  "price_tax_excluded": 0.033333,
  "price_without_reduction": 0.04,
  "tax_rate": 20.0,
  "currency": "EUR",
  "quantity_available": 1000,
  "available_for_order": true,
  "has_attributes": false,     // true when the product has variant combinations
  "attribute_groups": [],      // see "Variants" below
  "variants": [],
  "short_description_html": "<p>…</p>",
  "short_description": "…",
  "description_html": "<p>…</p><ul><li>…</li></ul>",
  "description": "…",          // plain text, list items prefixed with •
  "breadcrumb": [{ "id": 7, "name": "Zdravotnícke", "url": "…" }, …],
  "category_path": ["Zdravotnícke", " Netkané textíe, Pervin, SMS - …"],
  "categories": [{ "id": 10, "name": "…", "path": [...] }],   // every category listing it
  "default_image_id": 279,
  "images": [{ "id": 279, "url": "…/img/p/2/7/9/279.jpg",
               "file": "images/73/279.jpg", "bytes": 53443 }]
}
```

`breadcrumb`/`category_path` is the product's primary category trail. `categories`
lists every category the product is actually listed in — 7 products appear in more
than one, so treat product↔category as many-to-many.

## Variants

20 of the 118 products sell in several sizes/packagings — 74 combinations in total,
across 4 attribute groups: `Rozmer`, `Veľkosť`, `Balenie`, `Počet útržkov v balení`.
Each is its own purchasable unit with its own code, price and stock, so these map to
`product_variants` rather than to plain products.

```jsonc
"attribute_groups": [
  { "id": 4, "name": "Rozmer", "slug": "rozmer",
    "attributes": [ { "id": 21, "label": "10cm x 10cm", "default": false }, … ] }
],
"variants": [
  {
    "id": 30,                          // id_product_attribute
    "attribute_ids": [23],
    "label": "5cm x 5cm",              // groups joined with " / " when there are several
    "attributes": [{ "group": "rozmer", "label": "5cm x 5cm" }],
    "reference": "1211",               // per-variant Kód; "" on some products
    "quantity_available": 500,
    "price_impact_tax_excluded": -2.35,
    "price_tax_excluded": 1.04,
    "price": 1.14,                     // tax included — what the shop shows when selected
    "minimal_quantity": 1,
    "image_id": 100,                   // null when the variant has no own image
    "is_default": true
  }
]
```

`price` is computed as `(product.price_tax_excluded + price_impact_tax_excluded) *
(1 + tax_rate/100)`, matching `themes/default/js/product.js`. Verified: every
product's displayed price equals its default variant's computed price, and every
product's `quantity_available` equals the sum of its variants' stock.

On variant products the top-level `price`/`reference` describe the **default**
variant, and `quantity_available` is the total across variants. 14 of the 74
variants have an empty `reference`.

## Images

`images[].url` points at PrestaShop's untouched original upload
(`/img/p/<digits split by slash>/<image_id>.jpg`), not a generated thumbnail. For
product 73 that is 53 KB versus 10 KB for the `-large_default` the page shows. Every
downloaded file was verified to be a real JPEG; sizes run 3 KB – 634 KB.

Product 270 (`Plachta na posteľ z PVC`) has no image in the shop at all — its
`images` array is empty and `default_image_id` is null.

## Coverage notes

- Product URLs came from crawling every category listing, unioned with `sitemap.xml`.
  All category listings fit on a single page, so nothing was lost to pagination.
- `sitemap.xml` is stale: it advertises 182 products, but 83 of those (ids 8–55,
  72, 86, 96, 103–105, 109, 118–197 among them) return HTTP 404 — they are
  disabled/deleted products, verified individually via
  `index.php?id_product=<id>&controller=product`. 118 live products remain.
- Some products are served under both an old and a new slug (e.g. product 59 as
  `…59-gaza-hydrofilna-skladana-zlozky-z-gazy.html` and `…59-gaza-hydrofilna-skladana.html`).
  These are deduplicated by id, keeping the URL the category listing links to.
- 29 products have an empty top-level `reference` — that is genuine, the shop leaves
  the Kód field blank for them (the block is rendered `display:none`). 20 of those 29
  12 of those 29 are variant products, where the code lives on each variant instead.