sm-sitemap-plugin/classes/writer.go

package classes

import (
	"bufio"
	"encoding/xml"
	"io"
	"regexp"
	"strconv"
	"strings"
	"time"
)

// The XML namespaces of the sitemap files.
const (
	NSSitemap = "http://www.sitemaps.org/schemas/sitemap/0.9"
	NSXHTML   = "http://www.w3.org/1999/xhtml"
	NSImage   = "http://www.google.com/schemas/sitemap-image/1.1"
)

// IndexFile is one <sitemap> of the index: the file's absolute URL and the
// newest lastmod of its addresses (nil when none has one).
type IndexFile struct {
	Loc     string
	LastMod *time.Time
}

// FileAlternate is one xhtml:link alternate of a <url>, with an absolute
// href.
type FileAlternate struct {
	Hreflang string
	Href     string
}

// FileEntry is one <url> of a file: the absolute address, its lastmod and
// its hreflang alternates.
type FileEntry struct {
	Loc        string
	LastMod    *time.Time
	Alternates []FileAlternate
}

// fileNamePattern is the name of a per-group file.
var fileNamePattern = regexp.MustCompile(`^sitemap-([a-z0-9][a-z0-9-]*)-([1-9][0-9]{0,8})\.xml$`)

// FileName is the name of file n of group grp: sitemap-<grp>-<n>.xml.
func FileName(grp string, n int) string {
	return "sitemap-" + grp + "-" + strconv.Itoa(n) + ".xml"
}

// ParseFileName splits a file name made by FileName; ok is false for any
// other name.
func ParseFileName(name string) (grp string, n int, ok bool) {
	m := fileNamePattern.FindStringSubmatch(name)
	if m == nil {
		return "", 0, false
	}
	n, err := strconv.Atoi(m[2])
	if err != nil || n < 1 {
		return "", 0, false
	}
	return m[1], n, true
}

// lastmod formats a lastmod in RFC 3339, UTC.
func lastmod(t time.Time) string { return t.UTC().Format(time.RFC3339) }

// escape writes s with encoding/xml's escaping, which covers &, <, >, both
// quotes and control whitespace, so it is safe in text and in attribute
// values.
func escape(w *bufio.Writer, s string) {
	_ = xml.EscapeText(w, []byte(s))
}

// WriteIndex writes the sitemap index: one <sitemap> per file with its
// <loc> and, when known, its <lastmod>.
func WriteIndex(w io.Writer, files []IndexFile) error {
	bw := bufio.NewWriter(w)
	bw.WriteString(xml.Header)
	bw.WriteString(`<sitemapindex xmlns="` + NSSitemap + `">`)
	for _, f := range files {
		bw.WriteString("\n<sitemap><loc>")
		escape(bw, f.Loc)
		bw.WriteString("</loc>")
		if f.LastMod != nil {
			bw.WriteString("<lastmod>" + lastmod(*f.LastMod) + "</lastmod>")
		}
		bw.WriteString("</sitemap>")
	}
	bw.WriteString("\n</sitemapindex>\n")
	return bw.Flush()
}

// WriteURLSet writes one sitemap file: a <url> per entry with its <loc>, its
// <lastmod> when known and its xhtml:link alternates. The xhtml namespace
// is declared only when an entry has alternates. No <priority> and no
// <changefreq> are written (D-30).
func WriteURLSet(w io.Writer, entries []FileEntry) error {
	bw := bufio.NewWriter(w)
	bw.WriteString(xml.Header)
	bw.WriteString(`<urlset xmlns="` + NSSitemap + `"`)
	for _, e := range entries {
		if len(e.Alternates) > 0 {
			bw.WriteString(` xmlns:xhtml="` + NSXHTML + `"`)
			break
		}
	}
	bw.WriteString(">")
	for _, e := range entries {
		bw.WriteString("\n<url><loc>")
		escape(bw, e.Loc)
		bw.WriteString("</loc>")
		if e.LastMod != nil {
			bw.WriteString("<lastmod>" + lastmod(*e.LastMod) + "</lastmod>")
		}
		for _, a := range e.Alternates {
			bw.WriteString(`<xhtml:link rel="alternate" hreflang="`)
			escape(bw, a.Hreflang)
			bw.WriteString(`" href="`)
			escape(bw, a.Href)
			bw.WriteString(`"/>`)
		}
		bw.WriteString("</url>")
	}
	bw.WriteString("\n</urlset>\n")
	return bw.Flush()
}

// ImageEntry is one <url> of an image sitemap: the page's absolute address
// and its images' absolute addresses (at most MaxImagesPerURL).
type ImageEntry struct {
	Loc    string
	Images []string
}

// WriteImageURLSet writes one image sitemap file: a <url> per page with its
// <loc> and one <image:image><image:loc> per image. Only image:image and
// image:loc are written, the tags Google still reads (D-31).
func WriteImageURLSet(w io.Writer, entries []ImageEntry) error {
	bw := bufio.NewWriter(w)
	bw.WriteString(xml.Header)
	bw.WriteString(`<urlset xmlns="` + NSSitemap + `" xmlns:image="` + NSImage + `">`)
	for _, e := range entries {
		bw.WriteString("\n<url><loc>")
		escape(bw, e.Loc)
		bw.WriteString("</loc>")
		for i, img := range e.Images {
			if i >= MaxImagesPerURL {
				break
			}
			bw.WriteString("<image:image><image:loc>")
			escape(bw, img)
			bw.WriteString("</image:loc></image:image>")
		}
		bw.WriteString("</url>")
	}
	bw.WriteString("\n</urlset>\n")
	return bw.Flush()
}

// validSlug reports whether s may name a file group.
func validSlug(s string) bool {
	return s != "" && len(s) <= 64 && fileNamePattern.MatchString(FileName(s, 1)) && !strings.HasSuffix(s, "-")
}

// ValidSlug reports whether s may name a file group: lowercase letters,
// digits and hyphens, starting with a letter or digit, at most 64
// characters.
func ValidSlug(s string) bool { return validSlug(s) }