learn-platform-deploy/Backend/cmd/api/seo_public.go

package main

import (
	"encoding/xml"
	"net/http"
	"net/url"
	"strconv"
	"strings"
	"time"
)

type sitemapEntry struct {
	Loc     string
	LastMod string
}

type sitemapXML struct {
	XMLName xml.Name     `xml:"urlset"`
	XMLNS   string       `xml:"xmlns,attr"`
	URLs    []sitemapURL `xml:"url"`
}

type sitemapURL struct {
	Loc     string `xml:"loc"`
	LastMod string `xml:"lastmod,omitempty"`
}

func (a *app) sitemapHandler(w http.ResponseWriter, r *http.Request) {
	if a.db == nil {
		writeError(w, http.StatusServiceUnavailable, "database not configured")
		return
	}

	pages, err := a.list("pages", Doc{}, Sort{{Key: "updatedAt", Value: -1}})
	if err != nil {
		writeError(w, http.StatusInternalServerError, "failed to build sitemap")
		return
	}

	entries := collectSitemapEntries(pages, requestOrigin(r))
	body, err := renderSitemapXML(entries)
	if err != nil {
		writeError(w, http.StatusInternalServerError, "failed to render sitemap")
		return
	}

	w.Header().Set("Content-Type", "application/xml; charset=utf-8")
	w.WriteHeader(http.StatusOK)
	_, _ = w.Write([]byte(body))
}

func (a *app) robotsHandler(w http.ResponseWriter, r *http.Request) {
	body := renderRobotsTXT(requestOrigin(r))
	w.Header().Set("Content-Type", "text/plain; charset=utf-8")
	w.WriteHeader(http.StatusOK)
	_, _ = w.Write([]byte(body))
}

func collectSitemapEntries(pages []Doc, baseURL string) []sitemapEntry {
	seen := make(map[string]struct{}, len(pages))
	entries := make([]sitemapEntry, 0, len(pages))

	for _, page := range pages {
		if !isPublishedPublicPage(page) {
			continue
		}

		loc, ok := pagePublicURL(page, baseURL)
		if !ok {
			continue
		}
		if _, exists := seen[loc]; exists {
			continue
		}
		seen[loc] = struct{}{}

		entries = append(entries, sitemapEntry{
			Loc:     loc,
			LastMod: pageLastMod(page),
		})
	}

	return entries
}

func isPublishedPublicPage(page Doc) bool {
	published := false
	now := time.Now().UTC()

	if status, ok := asString(page["status"]); ok && strings.EqualFold(status, "published") {
		published = true
	}
	if value, ok := asBool(page["published"]); ok && value {
		published = true
	}
	if value, ok := asBool(page["isPublished"]); ok && value {
		published = true
	}
	if publishedAt, ok := asTime(page["publishedAt"]); ok {
		if publishedAt.UTC().After(now) {
			return false
		}
		published = true
	}

	if !published {
		return false
	}

	if value, ok := asBool(page["private"]); ok && value {
		return false
	}
	if value, ok := asBool(page["isPrivate"]); ok && value {
		return false
	}
	if visibility, ok := asString(page["visibility"]); ok && strings.EqualFold(visibility, "private") {
		return false
	}
	if access, ok := asString(page["access"]); ok && strings.EqualFold(access, "private") {
		return false
	}

	if !isIndexablePage(page) {
		return false
	}

	slug, ok := asString(page["slug"])
	return ok && strings.TrimSpace(slug) != ""
}

func isIndexablePage(page Doc) bool {
	seo, ok := asMap(page["seo"])
	if !ok {
		return true
	}

	robotsRaw, exists := seo["robots"]
	if !exists || robotsRaw == nil {
		return true
	}

	robots, ok := asMap(robotsRaw)
	if !ok {
		return true
	}

	if index, exists := asBool(robots["index"]); exists && !index {
		return false
	}

	return true
}

func pagePublicURL(page Doc, baseURL string) (string, bool) {
	seo, _ := asMap(page["seo"])
	if canonical, ok := asString(seo["canonicalUrl"]); ok && canonical != "" && isAbsoluteHTTPURL(canonical) {
		return canonical, true
	}

	slug, ok := asString(page["slug"])
	if !ok {
		return "", false
	}
	slug = strings.Trim(strings.TrimSpace(slug), "/")
	if slug == "" {
		return "", false
	}

	return strings.TrimRight(baseURL, "/") + "/pages/" + slug, true
}

func pageLastMod(page Doc) string {
	raw, ok := page["updatedAt"]
	if !ok {
		return ""
	}

	if t, ok := asTime(raw); ok {
		return t.UTC().Format(time.RFC3339)
	}

	return ""
}

func renderSitemapXML(entries []sitemapEntry) (string, error) {
	xmlDoc := sitemapXML{XMLNS: "http://www.sitemaps.org/schemas/sitemap/0.9"}
	xmlDoc.URLs = make([]sitemapURL, 0, len(entries))
	for _, entry := range entries {
		xmlDoc.URLs = append(xmlDoc.URLs, sitemapURL{Loc: entry.Loc, LastMod: entry.LastMod})
	}

	body, err := xml.MarshalIndent(xmlDoc, "", "  ")
	if err != nil {
		return "", err
	}

	return xml.Header + string(body), nil
}

func renderRobotsTXT(baseURL string) string {
	base := strings.TrimRight(baseURL, "/")
	return "User-agent: *\nAllow: /\nSitemap: " + base + "/sitemap.xml\n"
}

func requestOrigin(r *http.Request) string {
	scheme := r.Header.Get("X-Forwarded-Proto")
	if scheme == "" {
		if r.TLS != nil {
			scheme = "https"
		} else {
			scheme = "http"
		}
	}

	host := r.Header.Get("X-Forwarded-Host")
	if host == "" {
		host = r.Host
	}

	if host == "" {
		host = "localhost:8080"
	}

	return scheme + "://" + host
}

func isAbsoluteHTTPURL(value string) bool {
	u, err := url.Parse(strings.TrimSpace(value))
	if err != nil {
		return false
	}
	if u.Host == "" {
		return false
	}
	return u.Scheme == "http" || u.Scheme == "https"
}

func asMap(raw any) (map[string]any, bool) {
	switch value := raw.(type) {
	case map[string]any:
		return value, true
	case Doc:
		return map[string]any(value), true
	default:
		return nil, false
	}
}

func asString(raw any) (string, bool) {
	value, ok := raw.(string)
	if !ok {
		return "", false
	}
	value = strings.TrimSpace(value)
	if value == "" {
		return "", false
	}
	return value, true
}

func asBool(raw any) (bool, bool) {
	value, ok := raw.(bool)
	return value, ok
}

func asTime(raw any) (time.Time, bool) {
	switch value := raw.(type) {
	case time.Time:
		if value.IsZero() {
			return time.Time{}, false
		}
		return value, true
	case int:
		if value <= 0 {
			return time.Time{}, false
		}
		return time.UnixMilli(int64(value)), true
	case int32:
		if value <= 0 {
			return time.Time{}, false
		}
		return time.UnixMilli(int64(value)), true
	case int64:
		if value <= 0 {
			return time.Time{}, false
		}
		return time.UnixMilli(value), true
	case float64:
		if value <= 0 {
			return time.Time{}, false
		}
		return time.UnixMilli(int64(value)), true
	case string:
		trimmed := strings.TrimSpace(value)
		if trimmed == "" {
			return time.Time{}, false
		}
		if parsed, err := strconv.ParseInt(trimmed, 10, 64); err == nil && parsed > 0 {
			return time.UnixMilli(parsed), true
		}
		if parsed, err := time.Parse(time.RFC3339, trimmed); err == nil && !parsed.IsZero() {
			return parsed, true
		}
	}

	return time.Time{}, false
}