Storefront sitemaps for GoCommerce

What it does

Configuration

One setting is required: the storefront’s URL. With it in Config the module starts switched on; otherwise switch it on and fill it in on the Plugins screen, which wins where both say something.

The module reads its Config struct, not the environment. The variable names are the ones the package’s own example or the reference binary uses; in your own main() you choose where each value comes from. Where a setting has a panel label, the admin’s settings drawer can hold it too, and a value typed there wins over Config.

Setting it up

1. Install the moduleImport it and pass sitemaps.New to gocommerce.New, as below — or run the reference binary with -sitemaps, which reads STOREFRONT_URL.
2. Set the storefront and pathsThe storefront URL, and the product, collection and page paths if your storefront’s differ from the defaults, on the Plugins screen.
3. Point /sitemap.xml at itHave the storefront proxy or redirect /sitemap.xml to /x/sitemaps/sitemap.xml — and /x/sitemaps/ too, since the index links to the files there.
4. Check it, then submit itOpen the index in a browser: the stylesheet shows what is listed. Then submit the storefront’s /sitemap.xml to the search engines.

import (
	"os"

	"github.com/itswadesh/gocommerce/core"
	"github.com/itswadesh/gocommerce/ext/sitemaps"
)

app, err := gocommerce.New(cfg,
	sitemaps.New(sitemaps.Config{StorefrontURL: os.Getenv("STOREFRONT_URL")}),
)

How it works

• An index and three files

  sitemap.xml lists products.xml and collections.xml, and pages.xml when the cms module is installed. All five routes are public and documented in the module’s OpenAPI file.

• Live, not stored

  Each file is built on request from active products, active collections and published pages, read 500 at a time, and sent with Cache-Control: public, max-age=3600.

• Last changed, by the day

  Every URL carries lastmod — the day it last changed, in UTC — and a change frequency: weekly for products and collections, monthly for content pages.

• Readable in a browser

  Each file points at a stylesheet the module serves beside it, so the first click after switching it on answers whether it worked.

• Links from the request’s host

  The index’s links use the host the request came to, honouring X-Forwarded-Host and X-Forwarded-Proto, so behind a proxy they carry the storefront’s address.

• Off means 404

  Switched off, every file is a 404. Switched on without a storefront URL, a 409 that names the setting to fill in.

What it does not do

Read these before an order depends on it. No store is known to run GoCommerce in production yet, so these come from the code, not from anyone’s experience.

• 50,000 URLs per fileThe package doc names it: a sitemap file holds 50,000 URLs, and a catalogue past that needs the products file split, which this does not do yet.
• Paths in Config are shadowedThe panel’s path fields have defaults, read before Config, so ProductPath, CollectionPath and PagePath in Config never take effect. Set the paths on the Plugins screen.
• No categoriesProducts, collections and content pages only. Category pages are not listed.
• No images or languagesEach entry is a location, a date and a frequency: no image entries, and no alternates for other languages.
• The proxy is yoursThe files are served by GoCommerce under /x/sitemaps/. The storefront has to proxy or redirect its /sitemap.xml; the module cannot answer at the storefront’s address.
• No change pingsIt tells no search engine that something changed. That is the indexnow module’s job.