blurp/blurp.go

package main

import (
	"encoding/json"
	"fmt"
	"io"
	"io/ioutil"
	"log/slog"
	"net/http"
	"net/url"
	"os"
	"path/filepath"
	"time"

	auth "git.coopcloud.tech/decentral1se/gtslib-auth-keyring"
	"git.coopcloud.tech/decentral1se/gtslib/client/accounts"
	"git.coopcloud.tech/decentral1se/gtslib/client/statuses"
	"git.coopcloud.tech/decentral1se/gtslib/models"
	"github.com/peterhellberg/link"
	"github.com/pkg/errors"
	"github.com/spf13/cobra"
)

var (
	user           string // account username
	weeks          int    // number of weeks of statuses to retain
	rate           int    // requests per second
	dry            bool   // whether or not to run in dry mode
	skipFavourites bool   // whether or not to skip deleting favourites
	skipPinned     bool   // whether or not to skip deleting pinned statuses
)

func init() {
	loginCmd.Flags().StringVarP(&user, "user", "u", "", "username@domain of account")
	rootCmd.AddCommand(loginCmd)

	archiveCmd.Flags().IntVarP(&rate, "rate", "r", 1, "send a request every 'r' seconds")
	rootCmd.AddCommand(archiveCmd)

	deleteCmd.Flags().IntVarP(&weeks, "weeks", "w", 2, "keep statuses NEWER than no. of weeks")
	deleteCmd.Flags().IntVarP(&rate, "rate", "r", 1, "send a request every 'r' seconds")
	deleteCmd.Flags().BoolVarP(&dry, "dry", "d", false, "dry run mode (NO DELETION)")
	deleteCmd.Flags().BoolVarP(&skipFavourites, "skip-favourites", "f", false, "skip deletion of favourites")
	deleteCmd.Flags().BoolVarP(&skipPinned, "skip-pinnned", "p", false, "skip deletion of pinned statuses")
	rootCmd.AddCommand(deleteCmd)
}

// main is the command-line entrypoint.
func main() {
	if err := rootCmd.Execute(); err != nil {
		os.Exit(1) // NOTE(d1): Execute takes care of error output
	}
}

// rootCmd is the root command.
var rootCmd = &cobra.Command{
	Use:   "blurp",
	Short: "A GoToSocial status management tool",
}

// loginCmd is the login command.
var loginCmd = &cobra.Command{
	Use:   "login",
	Short: "Log in",
	RunE: func(cmd *cobra.Command, args []string) error {
		return auth.Login(user, auth.WithName("blurp"))
	},
}

// archiveCmd is the archive command.
var archiveCmd = &cobra.Command{
	Use:   "archive",
	Short: "Archive all account statuses (favourites, boosted, etc.)",
	RunE: func(cmd *cobra.Command, args []string) error {
		authClient, err := auth.NewAuthClient(user)
		if err != nil {
			slog.Error("unable to create auth client", "error", err)
			os.Exit(1)
		}

		acc, err := getAccount(authClient)
		if err != nil {
			slog.Error("unable to retrieve account", "error", err)
			os.Exit(1)
		}

		statuses, err := readAllPaged(authClient, acc.ID)
		if err != nil {
			slog.Error("unable to download paged response", "error", err)
			os.Exit(1)
		}

		basePath := filepath.Join(".", "archive")
		if err := os.MkdirAll(basePath, 0755); err != nil {
			slog.Error("unable to create status directory", "error", err)
			os.Exit(1)
		}

		for _, status := range statuses {
			basePath = filepath.Join(".", "archive")

			if len(status.MediaAttachments) > 0 {
				basePath = filepath.Join(basePath, status.ID)

				if err := os.MkdirAll(basePath, 0755); err != nil {
					slog.Error("unable to create status directory", "error", err)
					os.Exit(1)
				}

				for _, media := range status.MediaAttachments {
					parsed, err := url.Parse(media.URL)
					if err != nil {
						slog.Error("unable to parse media URL", "error", err)
						os.Exit(1)
					}

					imagePath := filepath.Join(basePath, filepath.Base(parsed.Path))
					if _, err := os.Stat(imagePath); errors.Is(err, os.ErrNotExist) {
						if err := httpGetFile(imagePath, media.URL); err != nil {
							slog.Error("unable to download file", "error", err)
							os.Exit(1)
						}
						slog.Info(fmt.Sprintf("archived %s", imagePath))
					}
				}
			}

			payload, err := json.MarshalIndent(status, "", " ")
			if err != nil {
				slog.Error("unable to marshal", "error", err)
				os.Exit(1)
			}

			jsonPath := filepath.Join(basePath, fmt.Sprintf("%s.json", status.ID))
			if _, err := os.Stat(jsonPath); errors.Is(err, os.ErrNotExist) {
				if err = ioutil.WriteFile(jsonPath, payload, 0644); err != nil {
					slog.Error("unable to write JSON file", "error", err)
					os.Exit(1)
				}
				slog.Info(fmt.Sprintf("archived %s", jsonPath))
			}
		}

		return nil
	},
}

var deleteCmd = &cobra.Command{
	Use:   "delete",
	Short: "Delete statuses like tears in the rain",
	Example: `
You can use "--weeks/-w 0" to signal "all posts". In combination with
"--dry/-d", you can get a clear idea of what blurp is going to do with each
status.

  blurp delete --dry --weeks 0`,
	RunE: func(cmd *cobra.Command, args []string) error {
		authClient, err := auth.NewAuthClient(user)
		if err != nil {
			slog.Error("unable to create auth client", "error", err)
			os.Exit(1)
		}

		slog.Info(fmt.Sprintf("keeping statuses NEWER than %d weeks", weeks))

		if dry {
			slog.Info("DRY RUN MODE ENABLED - STATUS DELETION DISABLED")
		}

		acc, err := getAccount(authClient)
		if err != nil {
			slog.Error("unable to retrieve account", "error", err)
			os.Exit(1)
		}

		slog.Info("retrieving statuses... this may take a moment...")

		allStatuses, err := readAllPaged(authClient, acc.ID)
		if err != nil {
			slog.Error("unable to download paged response", "error", err)
			os.Exit(1)
		}

		ISO8601 := "2006-01-02T15:04:05.000Z"
		for _, status := range allStatuses {
			t, err := time.Parse(ISO8601, status.CreatedAt)
			if err != nil {
				slog.Error("unable to parse status 'CreatedAt' value", "error", err)
				os.Exit(1)
			}

			if status.Favourited && skipFavourites {
				slog.Info(fmt.Sprintf("skipping %s (skip favourites: %v)", status.ID, skipFavourites))
				continue
			}

			if status.Pinned && skipPinned {
				slog.Info(fmt.Sprintf("skipping %s (skip pinned: %v)", status.ID, skipPinned))
				continue
			}

			numHours := time.Duration(168 * weeks)
			if t.Before(time.Now().Add(-time.Hour * numHours)) {
				if !dry {
					_, err := authClient.Client.Statuses.StatusDelete(&statuses.StatusDeleteParams{
						ID: status.ID,
					}, authClient.Auth)
					if err != nil {
						slog.Error("unable to delete status", "error", err)
						os.Exit(1)
					}
				}

				msg := fmt.Sprintf("deleted %s (created: %s)", status.ID, t.Format(time.DateOnly))
				if dry {
					msg = fmt.Sprintf("DRY RUN: %s", msg)
				}
				slog.Info(msg)

				time.Sleep(time.Duration(rate) * time.Second)
			} else {
				slog.Info(fmt.Sprintf("keeping %s (created: %s)", status.ID, t.Format(time.DateOnly)))
			}
		}

		return nil
	},
}

// getAccount returns the currently authenticated account.
func getAccount(authClient *auth.Client) (*models.Account, error) {
	err := authClient.Wait()
	if err != nil {
		return nil, err
	}

	resp, err := authClient.Client.Accounts.AccountVerify(nil, authClient.Auth)
	if err != nil {
		return nil, errors.WithStack(err)
	}

	return resp.GetPayload(), nil
}

// httpGetFile downloads a file from the internet.
func httpGetFile(filepath, url string) error {
	out, err := os.Create(filepath)
	if err != nil {
		return fmt.Errorf("httpGetFile: unable to create '%s': %s", filepath, err)
	}
	defer out.Close()

	resp, err := http.Get(url)
	if err != nil {
		return fmt.Errorf("httpGetFile: unable to HTTP GET '%s'", url)
	}
	defer resp.Body.Close()

	if resp.StatusCode != http.StatusOK {
		return fmt.Errorf("httpGetFile: HTTP GET response code %v for '%s'", resp.StatusCode, url)
	}

	_, err = io.Copy(out, resp.Body)
	if err != nil {
		return fmt.Errorf("httpGetFile: unable to copy HTTP GET response to disk: %s", err)
	}

	return nil
}

// parseLinkMaxID extracts the `max_id` from the `next` link for paging to older items.
func parseLinkMaxID(linkHeader string) (*string, error) {
	next := link.Parse(linkHeader)["next"]
	if next == nil {
		// No link header in that direction means end of results.
		return nil, nil
	}
	nextUrl, err := url.Parse(next.URI)
	if err != nil {
		return nil, errors.Wrap(err, "couldn't parse next page URL")
	}
	nextMaxID := nextUrl.Query().Get("max_id")
	if nextMaxID == "" {
		return nil, errors.New("couldn't find next page max ID")
	}
	return &nextMaxID, err
}

func readAllPaged(authClient *auth.Client, accID string) ([]*models.Status, error) {
	var all []*models.Status
	var maxID *string

	for {
		err := authClient.Wait()
		if err != nil {
			return all, errors.WithStack(err)
		}

		params := &accounts.AccountStatusesParams{ID: accID, MaxID: maxID}
		resp, err := authClient.Client.Accounts.AccountStatuses(params, authClient.Auth)
		if err != nil {
			slog.Error("error fetching page", "error", err)
			return all, errors.WithStack(err)
		}

		maxID, err = parseLinkMaxID(resp.Link)
		if err != nil {
			slog.Error("error parsing Link header", "error", err)
			return all, errors.WithStack(err)
		}
		if maxID == nil {
			// End of pages.
			break
		}

		time.Sleep(time.Duration(rate) * time.Second)

		all = append(all, resp.GetPayload()...)
	}

	return all, nil
}