Shopify through HasData
Two tools, both keyed on a storefront url. hasdata_shopify_products_getProducts returns products, optionally narrowed to a collection, and hasdata_shopify_collections_getCollections returns the collections themselves. Both page with limit and page.
The url is a store's domain, so the tools work against any Shopify storefront rather than against one account. Nothing here is authenticated, so what comes back is what the store publishes.
The field names are Shopify's, not ours
These payloads are snake_case, unlike the rest of HasData's, because they mirror Shopify's own product shape. A product carries id, title, handle, body_html, vendor, product_type, tags, options, images, variants, created_at, published_at and updated_at. Code that reaches for productType or bodyHtml finds nothing.
body_html is HTML, not text. Strip the markup before putting a description in front of a reader.
Price lives on the variant, and it is a string
A product has no price of its own. Each entry in variants carries price, compare_at_price, sku, its option values and an available flag.
Both prices are strings, such as "16.50". Parse them to numbers before sorting, comparing or averaging, because a string sort puts "9.00" above "16.00" and the answer looks plausible. compare_at_price can be null, so test it before calling anything a markdown.
available is per variant, so a product is not sold out until every variant says so. The figure to quote is a range across variants or the price of a named one, and a product with a single variant is the only case with a single price.
A collection carries id, title, handle, description, products_count, published_at and updated_at. A products_count of zero is ordinary for a staged or seasonal collection rather than a sign the call failed.
A wrong storefront still answers 200
A URL that is not a classic Shopify storefront returns a successful call with an error string where products should be, and requestMetadata.status still reads ok. Test for the products array rather than the status. A headless store often fails on its custom domain and works on its <shop>.myshopify.com address, so that is the retry worth making.
Paging
page starts at 1 and limit caps the page at 250. collection takes the collection handle rather than its title or id.
There is no pagination block in the products response, unlike the other tools here, so the only stop signal is a page that comes back short or empty. Set limit to 250 and read products_count from the collection first, because a page costs the same whatever its size and a catalogue walked twenty at a time is twelve times the calls.
