Feature: Add dump feature to export all raw messages to a local directory (#443)

cmd/dump.go

@@ -0,0 +1,36 @@
+package cmd
+
+import (
+	"github.com/axllent/mailpit/config"
+	"github.com/axllent/mailpit/internal/dump"
+	"github.com/axllent/mailpit/internal/logger"
+	"github.com/spf13/cobra"
+)
+
+// dumpCmd represents the dump command
+var dumpCmd = &cobra.Command{
+	Use:   "dump <database> <output-dir>",
+	Short: "Dump all messages from a database to a directory",
+	Long: `Dump all messages stored in Mailpit into a local directory as individual files.
+
+The database can either be the database file (eg: --database /var/lib/mailpit/mailpit.db) or a
+URL of a running Mailpit instance (eg: --http http://127.0.0.1/). If dumping over HTTP, the URL
+should be the base URL of your running Mailpit instance, not the link to the API itself.`,
+	Args: cobra.ExactArgs(1),
+	Run: func(cmd *cobra.Command, args []string) {
+		if err := dump.Sync(args[0]); err != nil {
+			logger.Log().Fatal(err)
+		}
+	},
+}
+
+func init() {
+	rootCmd.AddCommand(dumpCmd)
+
+	dumpCmd.Flags().SortFlags = false
+
+	dumpCmd.Flags().StringVar(&config.Database, "database", config.Database, "Dump messages directly from a database file")
+	dumpCmd.Flags().StringVar(&config.TenantID, "tenant-id", config.TenantID, "Database tenant ID to isolate data (optional)")
+	dumpCmd.Flags().StringVar(&dump.URL, "http", dump.URL, "Dump messages via HTTP API (base URL of running Mailpit instance)")
+	dumpCmd.Flags().BoolVarP(&logger.VerboseLogging, "verbose", "v", logger.VerboseLogging, "Verbose logging")
+}

internal/dump/dump.go

@@ -0,0 +1,163 @@
+// Package dump is used to export all messages from mailpit into a directory
+package dump
+
+import (
+	"encoding/json"
+	"errors"
+	"io"
+	"net/http"
+	"os"
+	"path"
+	"regexp"
+	"strings"
+
+	"github.com/axllent/mailpit/config"
+	"github.com/axllent/mailpit/internal/logger"
+	"github.com/axllent/mailpit/internal/storage"
+	"github.com/axllent/mailpit/internal/tools"
+	"github.com/axllent/mailpit/server/apiv1"
+)
+
+var (
+	linkRe = regexp.MustCompile(`(?i)^https?:\/\/`)
+
+	outDir string
+
+	// Base URL of mailpit instance
+	base string
+
+	// URL is the base URL of a remove Mailpit instance
+	URL string
+
+	summary = []storage.MessageSummary{}
+)
+
+// Sync will sync all messages from the specified database or API to the specified output directory
+func Sync(d string) error {
+
+	outDir = path.Clean(d)
+
+	if URL != "" {
+		if !linkRe.MatchString(URL) {
+			return errors.New("Invalid URL")
+		}
+
+		base = strings.TrimRight(URL, "/") + "/"
+	}
+
+	if base == "" && config.Database == "" {
+		return errors.New("No database or API URL specified")
+	}
+
+	if !tools.IsDir(outDir) {
+		if err := os.MkdirAll(outDir, 0755); err != nil {
+			return err
+		}
+	}
+
+	if err := loadIDs(); err != nil {
+		return err
+	}
+
+	if err := saveMessages(); err != nil {
+		return err
+	}
+
+	return nil
+}
+
+// LoadIDs will load all message IDs from the specified database or API
+func loadIDs() error {
+	if base != "" {
+		// remote
+		logger.Log().Debugf("Fetching messages summary from %s", base)
+		res, err := http.Get(base + "api/v1/messages?limit=0")
+
+		if err != nil {
+			return err
+		}
+
+		body, err := io.ReadAll(res.Body)
+
+		if err != nil {
+			return err
+		}
+
+		var data apiv1.MessagesSummary
+		if err := json.Unmarshal(body, &data); err != nil {
+			return err
+		}
+
+		summary = data.Messages
+
+	} else {
+		// make sure the database isn't pruned while open
+		config.MaxMessages = 0
+
+		var err error
+		// local database
+		if err = storage.InitDB(); err != nil {
+			return err
+		}
+
+		logger.Log().Debugf("Fetching messages summary from %s", config.Database)
+
+		summary, err = storage.List(0, 0, 0)
+		if err != nil {
+			return err
+		}
+	}
+
+	if len(summary) == 0 {
+		return errors.New("No messages found")
+	}
+
+	return nil
+}
+
+func saveMessages() error {
+	for _, m := range summary {
+		out := path.Join(outDir, m.ID+".eml")
+
+		// skip if message exists
+		if tools.IsFile(out) {
+			continue
+		}
+
+		var b []byte
+
+		if base != "" {
+			res, err := http.Get(base + "api/v1/message/" + m.ID + "/raw")
+
+			if err != nil {
+				logger.Log().Errorf("Error fetching message %s: %s", m.ID, err.Error())
+				continue
+			}
+
+			b, err = io.ReadAll(res.Body)
+
+			if err != nil {
+				logger.Log().Errorf("Error fetching message %s: %s", m.ID, err.Error())
+				continue
+			}
+		} else {
+			var err error
+			b, err = storage.GetMessageRaw(m.ID)
+			if err != nil {
+				logger.Log().Errorf("Error fetching message %s: %s", m.ID, err.Error())
+				continue
+			}
+		}
+
+		if err := os.WriteFile(out, b, 0644); err != nil {
+			logger.Log().Errorf("Error writing message %s: %s", m.ID, err.Error())
+			continue
+		}
+
+		_ = os.Chtimes(out, m.Created, m.Created)
+
+		logger.Log().Debugf("Saved message %s to %s", m.ID, out)
+	}
+
+	return nil
+}

internal/storage/messages.go

@@ -175,9 +175,11 @@ func List(start int, beforeTS int64, limit int) ([]MessageSummary, error) {
 
 	q := sqlf.From(tenant("mailbox") + " m").
 		Select(`m.Created, m.ID, m.MessageID, m.Subject, m.Metadata, m.Size, m.Attachments, m.Read, m.Snippet`).
-		OrderBy("m.Created DESC").
+		OrderBy("m.Created DESC")
-		Limit(limit).
+
-		Offset(start)
+	if limit > 0 {
+		q = q.Limit(limit).Offset(start)
+	}
 
 	if beforeTS > 0 {
 		q = q.Where("Created < ?", beforeTS)

internal/tools/fs.go

@@ -0,0 +1,23 @@
+package tools
+
+import (
+	"os"
+	"path/filepath"
+)
+
+// IsFile returns whether a file exists and is readable
+func IsFile(path string) bool {
+	f, err := os.Open(filepath.Clean(path))
+	defer f.Close()
+	return err == nil
+}
+
+// IsDir returns whether a path is a directory
+func IsDir(path string) bool {
+	info, err := os.Stat(path)
+	if err != nil || os.IsNotExist(err) || !info.IsDir() {
+		return false
+	}
+
+	return true
+}