archivmail/internal/storage/storage.go

package storage

import (
	"crypto/sha256"
	"errors"
	"fmt"
	"io/fs"
	"os"
	"path/filepath"
	"time"
)

// Store is a file-based email storage using SHA256 for deduplication.
type Store struct {
	dir string
}

// StoreStats reports total mail count and size in bytes.
type StoreStats struct {
	TotalMails int64
	TotalBytes int64
}

// New initialises the storage directory, creating required subdirectories.
func New(dir string) (*Store, error) {
	for _, sub := range []string{"store", "attachments", "meta"} {
		if err := os.MkdirAll(filepath.Join(dir, sub), 0o755); err != nil {
			return nil, fmt.Errorf("storage: mkdir %s: %w", sub, err)
		}
	}
	return &Store{dir: dir}, nil
}

// Save writes raw email bytes to storage. The ID is the hex-encoded SHA256 of
// the content. If the file already exists, Save is a no-op (deduplication).
func (s *Store) Save(raw []byte, _ time.Time) (string, error) {
	sum := sha256.Sum256(raw)
	id := fmt.Sprintf("%x", sum[:]) // 64 hex chars

	path := s.filePath(id)
	if err := os.MkdirAll(filepath.Dir(path), 0o755); err != nil {
		return "", fmt.Errorf("storage: mkdir shard: %w", err)
	}

	// If file already exists, dedup: return same id without error.
	if _, err := os.Stat(path); err == nil {
		return id, nil
	}

	if err := os.WriteFile(path, raw, 0o644); err != nil {
		return "", fmt.Errorf("storage: write: %w", err)
	}

	return id, nil
}

// Load reads a stored email by its ID.
func (s *Store) Load(id string) ([]byte, error) {
	path := s.filePath(id)
	data, err := os.ReadFile(path)
	if err != nil {
		if errors.Is(err, os.ErrNotExist) {
			return nil, fmt.Errorf("storage: not found: %s", id)
		}
		return nil, fmt.Errorf("storage: read: %w", err)
	}
	return data, nil
}

// Delete removes a stored email by its ID.
func (s *Store) Delete(id string) error {
	path := s.filePath(id)
	if err := os.Remove(path); err != nil {
		if errors.Is(err, os.ErrNotExist) {
			return fmt.Errorf("storage: not found: %s", id)
		}
		return fmt.Errorf("storage: delete: %w", err)
	}
	return nil
}

// Stats walks the store directory and returns aggregate statistics.
func (s *Store) Stats() (*StoreStats, error) {
	var stats StoreStats
	err := filepath.WalkDir(filepath.Join(s.dir, "store"), func(_ string, d fs.DirEntry, err error) error {
		if err != nil {
			return err
		}
		if d.IsDir() {
			return nil
		}
		info, err := d.Info()
		if err != nil {
			return err
		}
		stats.TotalMails++
		stats.TotalBytes += info.Size()
		return nil
	})
	if err != nil {
		return nil, fmt.Errorf("storage: stats: %w", err)
	}
	return &stats, nil
}

// MailRef holds the ID and modification time of a stored mail.
type MailRef struct {
	ID      string
	ModTime time.Time
}

// FirstAndLastMail walks the store and returns the oldest and newest mail by
// file modification time. Returns nil for either if the store is empty.
func (s *Store) FirstAndLastMail() (first, last *MailRef, err error) {
	err = filepath.WalkDir(filepath.Join(s.dir, "store"), func(path string, d fs.DirEntry, werr error) error {
		if werr != nil {
			return werr
		}
		if d.IsDir() {
			return nil
		}
		info, err := d.Info()
		if err != nil {
			return err
		}
		ref := &MailRef{ID: d.Name(), ModTime: info.ModTime()}
		if first == nil || ref.ModTime.Before(first.ModTime) {
			first = ref
		}
		if last == nil || ref.ModTime.After(last.ModTime) {
			last = ref
		}
		return nil
	})
	if err != nil {
		return nil, nil, fmt.Errorf("storage: first/last: %w", err)
	}
	return first, last, nil
}

// filePath returns the on-disk path for a given mail ID.
// Uses 2-char prefix sharding: {dir}/store/{id[:2]}/{id}
func (s *Store) filePath(id string) string {
	return filepath.Join(s.dir, "store", id[:2], id)
}