sm-sitemap-plugin/console/import.go

package console

import (
	"context"
	"encoding/json"
	"errors"
	"fmt"
	"os"
	"strings"

	"git.golem15.com/golem15/sm-sitemap-plugin/models"
	"git.golem15.com/golem15/summercms/modules/backpack"
	"git.golem15.com/golem15/summercms/modules/bonfire"
	"gorm.io/gorm"
)

// Import is sitemap:import FILE [--append]: it reads a JSON array of URL
// entries ({"url": "/path", "lastmod": "2026-10-28"}; lastmod optional) into
// the settings' additional addresses, in one transaction. Without --append
// the imported addresses replace the additional addresses; with --append
// they are added after them, an address already listed is skipped. An
// entry whose url is not a path or an http(s) URL (models.AllowedAddress),
// or whose lastmod is not a date, stops the import before anything is
// written. The Phase 17 keys changefreq and priority are ignored, and
// --theme is accepted for the PHP command's sake. The files follow at the
// next rebuild (sitemap:rebuild).
func Import(app *backpack.App) bonfire.Command {
	return bonfire.Command{
		Name:        "sitemap:import",
		Description: "Import additional sitemap addresses from a JSON manifest file",
		Long: `Reads a JSON array of URL entries ({"url": "/path", "lastmod":
"2026-10-28"}; lastmod is optional) and puts them into the sitemap's
additional addresses (Settings > Sitemaps > Additional addresses), in one
transaction. Every entry needs a url that is a path or an http(s) URL; one bad
entry stops the import before anything is written. The old changefreq and
priority keys are ignored: the sitemaps no longer write them.

By default the import replaces the additional addresses. --append adds the
addresses after them and skips those already listed. --theme is accepted for
the PHP command's sake and does nothing. The files change at the next rebuild
(sitemap:rebuild, a settings save or the nightly run). It ends with
"Imported N addresses into the sitemap's additional addresses."`,
		Args: []bonfire.Arg{
			{Name: "file", Description: "Path to the JSON manifest (relative to the working directory or absolute)", Required: true},
		},
		Flags: []bonfire.Flag{
			{Name: "theme", Description: "Accepted for the PHP command; ignored"},
			{Name: "append", Description: "Add the addresses instead of replacing the additional addresses", Bare: true},
		},
		Examples: []bonfire.Example{
			{Command: "sitemap:import urls.json", Description: "Replace the additional addresses with the manifest's addresses"},
			{Command: "sitemap:import urls.json --append", Description: "Add the addresses and keep the existing ones"},
		},
		Run: func(ctx context.Context, in bonfire.Input, out bonfire.Output) error {
			file, _ := in.Argument("file")
			db, done, err := lookupDB(ctx, app)
			if err != nil {
				return err
			}
			defer done()
			return ImportFile(ctx, db, file, flagValue(in, "theme"), flagOn(in, "append"), out)
		},
	}
}

// ImportFile runs sitemap:import against db.
func ImportFile(ctx context.Context, db *gorm.DB, file, theme string, appendMode bool, out bonfire.Output) error {
	raw, err := os.ReadFile(file)
	if err != nil {
		if errors.Is(err, os.ErrNotExist) {
			return fmt.Errorf("File not found: %s", file)
		}
		return fmt.Errorf("Cannot read %s: %w", file, err)
	}
	entries, labels, ok := models.LegacyEntries(raw)
	if !ok {
		return errors.New("Invalid JSON: expected an array of URL entries")
	}
	rows := make([]models.ExtraURL, 0, len(entries))
	for i, entry := range entries {
		var fields map[string]json.RawMessage
		_ = json.Unmarshal(entry, &fields)
		url, err := models.DecodeText(fields["url"])
		if fields["url"] == nil || err != nil || strings.TrimSpace(models.Text(url)) == "" {
			return fmt.Errorf("Entry %s is missing a 'url' field", labels[i])
		}
		u := strings.TrimSpace(models.Text(url))
		if !models.AllowedAddress(u) {
			return fmt.Errorf("Entry %s has a url that is not a path or an http(s) URL", labels[i])
		}
		row := models.ExtraURL{URL: u}
		if raw, ok := fields["lastmod"]; ok {
			mod, err := models.DecodeText(raw)
			if err != nil || (strings.TrimSpace(models.Text(mod)) != "" && !models.ValidDate(models.Text(mod))) {
				return fmt.Errorf("Entry %s has a lastmod that is not a date (YYYY-MM-DD)", labels[i])
			}
			row.LastMod = strings.TrimSpace(models.Text(mod))
		}
		rows = append(rows, row)
	}
	var kept, added int
	err = db.WithContext(ctx).Transaction(func(tx *gorm.DB) error {
		s, err := loadSettingsForUpdate(tx)
		if err != nil {
			return err
		}
		next := []models.ExtraURL{}
		seen := map[string]bool{}
		if appendMode {
			for _, e := range s.Extra() {
				next = append(next, e)
				seen[e.URL] = true
			}
			kept = len(next)
		}
		for _, r := range rows {
			if seen[r.URL] {
				continue
			}
			seen[r.URL] = true
			next = append(next, r)
			added++
		}
		return saveExtra(tx, next)
	})
	if err != nil {
		return err
	}
	out.Success(fmt.Sprintf("Imported %d addresses into the sitemap's additional addresses.", added))
	if kept > 0 {
		out.Info(fmt.Sprintf("Kept %d existing addresses.", kept))
	}
	if theme != "" {
		out.Warning(fmt.Sprintf("Theme '%s' was given; the sitemap has one set of additional addresses, so the addresses were imported there.", theme))
	}
	out.Info("Run sitemap:rebuild (or save the sitemap settings) to update the files now.")
	return nil
}