archivmail/internal/storage/attachments.go

package storage

import (
	"context"
	"crypto/sha256"
	"fmt"
	"os"
	"path/filepath"

	"github.com/archivmail/pkg/mailparser"
)

// saveAttachments deduplicates and stores attachments from a parsed email.
// Each unique attachment (by SHA-256 hash) is stored once on disk.
// email_attachments links attachments to their email record.
func (s *Store) saveAttachments(ctx context.Context, emailID string, pm *mailparser.ParsedMail) error {
	if s.db == nil || len(pm.Attachments) == 0 {
		return nil
	}

	for _, att := range pm.Attachments {
		if len(att.Data) == 0 {
			continue
		}

		sum := sha256.Sum256(att.Data)
		hash := fmt.Sprintf("%x", sum[:])

		// Check if this attachment is already stored
		var attID int64
		err := s.db.QueryRow(ctx, `SELECT id FROM attachments WHERE hash = $1`, hash).Scan(&attID)
		if err != nil {
			// Not found — compress and store
			toWrite := att.Data
			compression := "none"
			if s.compressEnabled {
				compressed, cerr := compressGzip(att.Data)
				if cerr == nil && len(compressed) < len(att.Data) {
					toWrite = compressed
					compression = "gzip"
				}
			}

			attPath := s.attachmentPath(hash)
			if err := os.MkdirAll(filepath.Dir(attPath), 0o755); err != nil {
				return fmt.Errorf("storage: attachment mkdir: %w", err)
			}
			if _, statErr := os.Stat(attPath); os.IsNotExist(statErr) {
				if err := os.WriteFile(attPath, toWrite, 0o644); err != nil {
					return fmt.Errorf("storage: attachment write: %w", err)
				}
			}

			// Register in storage_objects
			var soID int64
			soErr := s.db.QueryRow(ctx, `
				INSERT INTO storage_objects (storage_type, path, compression, size_original, size_compressed, checksum)
				VALUES ('filesystem', $1, $2, $3, $4, $5)
				RETURNING id
			`, attPath, compression, int64(len(att.Data)), int64(len(toWrite)), hash).Scan(&soID)

			// Insert attachment record
			var insertErr error
			if soErr == nil {
				insertErr = s.db.QueryRow(ctx, `
					INSERT INTO attachments (filename, mime_type, size_bytes, hash, storage_id)
					VALUES ($1, $2, $3, $4, $5)
					ON CONFLICT (hash) DO UPDATE SET hash = EXCLUDED.hash
					RETURNING id
				`, att.Filename, att.ContentType, int64(len(att.Data)), hash, soID).Scan(&attID)
			} else {
				insertErr = s.db.QueryRow(ctx, `
					INSERT INTO attachments (filename, mime_type, size_bytes, hash)
					VALUES ($1, $2, $3, $4)
					ON CONFLICT (hash) DO UPDATE SET hash = EXCLUDED.hash
					RETURNING id
				`, att.Filename, att.ContentType, int64(len(att.Data)), hash).Scan(&attID)
			}
			if insertErr != nil {
				continue // non-fatal: mail is saved, attachment linking is best-effort
			}
		}

		// Link attachment to email
		_, _ = s.db.Exec(ctx, `
			INSERT INTO email_attachments (email_id, attachment_id)
			VALUES ($1, $2)
			ON CONFLICT DO NOTHING
		`, emailID, attID)
	}

	return nil
}

// LoadAttachment reads and decompresses an attachment by its SHA-256 hash.
func (s *Store) LoadAttachment(hash string) ([]byte, error) {
	path := s.attachmentPath(hash)
	data, err := os.ReadFile(path)
	if err != nil {
		return nil, fmt.Errorf("storage: attachment not found: %s", hash)
	}
	return maybeDecompress(data)
}

// attachmentPath returns the on-disk path for a given attachment hash.
// Uses 2-level 2-char prefix sharding: {dir}/attachments/{ab}/{cd}/{hash}
func (s *Store) attachmentPath(hash string) string {
	return filepath.Join(s.dir, "attachments", hash[:2], hash[2:4], hash)
}