sm-sitemap-plugin/classes/generator.go

// Package classes generates the sitemap: it walks the definition items,
// resolves registry items through the framework's menu item type registry
// and writes the PHP-identical XML.
package classes

import (
	"context"
	"time"

	"git.golem15.com/golem15/sm-sitemap-plugin/models"
	"git.golem15.com/golem15/summercms/modules/pact"
)

// MaxURLs is the sitemap protocol's URL limit (PHP Definition::MAX_URLS):
// the 50,000th URL is written, the 50,001st is not.
const MaxURLs = 50000

// lastmodLayout is PHP date('c') in UTC: "+00:00", never "Z".
const lastmodLayout = "2006-01-02T15:04:05-07:00"

// Resolver resolves a registry item; *signpost.Registry implements it.
type Resolver interface {
	Resolve(ctx context.Context, key, reference, itemPattern string, nesting bool, unprefixedLocale string) (*pact.ResolvedMenuItem, error)
}

// Generator writes the sitemap of a definition. Root is the backend root URL
// (scheme and host without a trailing slash) that Url::to builds absolute
// URLs from. Resolver resolves registry items; nil skips them.
// UnprefixedLocale is handed to every Resolve as the locale without a URL
// prefix ("" leaves it to the type).
type Generator struct {
	Root             string
	Resolver         Resolver
	UnprefixedLocale string

	entries []URLEntry
	count   int
}

// itemInfo is PHP's $itemInfo: the URL (set marks it present), the record's
// modification time and the hreflang alternates.
type itemInfo struct {
	url        string
	set        bool
	mtime      *time.Time
	alternates []Alternate
}

// Generate returns the sitemap XML of items, or "" when there are none (PHP
// generateSitemap). A url item is added with its URL; a registry item is
// resolved with its URL pattern and nesting, its URL added when it has one
// and its sub-items added recursively. A type the registry does not know,
// or a Resolve that finds nothing, adds nothing.
func (g *Generator) Generate(ctx context.Context, items []models.DefinitionItem) (string, error) {
	if len(items) == 0 {
		return "", nil
	}
	g.entries, g.count = nil, 0
	for i := range items {
		item := &items[i]
		if item.Type == models.ItemTypeURL {
			g.addItem(item, itemInfo{})
			continue
		}
		resolved, err := g.resolve(ctx, item)
		if err != nil {
			return "", err
		}
		if resolved == nil {
			continue
		}
		if resolved.URL != "" {
			g.addItem(item, infoOf(*resolved))
		}
		g.addChildren(item, resolved.Items)
	}
	return Serialize(g.Stylesheet(), g.entries), nil
}

// Stylesheet is the stylesheet href PHP built:
// Url::buildUrl(Url::current(), ['path' => 'sitemap.xsl']).
func (g *Generator) Stylesheet() string {
	return replacePath(URLTo(g.Root, "/sitemap.xml"), "sitemap.xsl")
}

func (g *Generator) resolve(ctx context.Context, item *models.DefinitionItem) (*pact.ResolvedMenuItem, error) {
	if g.Resolver == nil || item.Type == "" {
		return nil, nil
	}
	if reg, ok := g.Resolver.(interface {
		Type(string) (pact.MenuItemType, bool)
	}); ok {
		if _, known := reg.Type(item.Type); !known {
			return nil, nil
		}
	}
	return g.Resolver.Resolve(ctx, item.Type, models.Text(item.Reference), models.Text(item.URLPattern), item.Nesting, g.UnprefixedLocale)
}

// addChildren adds every child with a URL, then its own children.
func (g *Generator) addChildren(item *models.DefinitionItem, children []pact.ResolvedMenuItem) {
	for _, child := range children {
		if child.URL != "" {
			g.addItem(item, infoOf(child))
		}
		g.addChildren(item, child.Items)
	}
}

func infoOf(r pact.ResolvedMenuItem) itemInfo {
	info := itemInfo{url: r.URL, set: true, mtime: r.MTime}
	for _, a := range r.Alternates {
		info.alternates = append(info.alternates, Alternate{Hreflang: a.Locale, Href: a.URL})
	}
	return info
}

// addItem is PHP Definition::addItemToSet.
func (g *Generator) addItem(item *models.DefinitionItem, info itemInfo) {
	if g.count >= MaxURLs {
		return
	}
	if !info.set {
		info.url = URLTo(g.Root, models.Text(item.URL))
	}
	lastmod := ""
	if info.mtime != nil && info.mtime.Unix() != 0 {
		lastmod = info.mtime.UTC().Format(lastmodLayout)
	}
	if info.url == "" || info.url == "0" {
		return
	}
	loc := URLTo(g.Root, info.url)
	g.count++
	entry := URLEntry{
		Loc:        loc,
		Lastmod:    lastmod,
		Changefreq: models.Text(item.Changefreq),
		Priority:   models.Text(item.Priority),
	}
	if len(info.alternates) > 1 {
		for _, a := range info.alternates {
			entry.Alternates = append(entry.Alternates, Alternate{Hreflang: a.Hreflang, Href: URLTo(g.Root, a.Href)})
		}
	}
	g.entries = append(g.entries, entry)
}