mygo/internal/service/file.go

package service

import (
	"bytes"
	"context"
	"crypto/sha256"
	"encoding/hex"
	"errors"
	"fmt"
	"io"
	"log/slog"
	"net/http"
	"strings"
	"time"

	"github.com/gabriel-vasile/mimetype"
	"github.com/google/uuid"

	"github.com/dhao2001/mygo/internal/model"
	"github.com/dhao2001/mygo/internal/repository"
	"github.com/dhao2001/mygo/internal/storage"
)

// FileInfo is the public representation of a file or directory.
type FileInfo struct {
	ID        string    `json:"id"`
	UserID    string    `json:"user_id"`
	ParentID  *string   `json:"parent_id"`
	Name      string    `json:"name"`
	Size      int64     `json:"size"`
	MimeType  string    `json:"mime_type"`
	IsDir     bool      `json:"is_dir"`
	Hash      string    `json:"hash,omitempty"`
	CreatedAt time.Time `json:"created_at"`
	UpdatedAt time.Time `json:"updated_at"`
}

// FileList is a paginated list of files.
type FileList struct {
	Files []FileInfo `json:"files"`
	Total int64      `json:"total"`
}

// FileService handles file management business logic.
type FileService struct {
	fileRepo      repository.FileRepository
	storage       storage.StorageBackend
	maxUploadSize int64
	logger        *slog.Logger
}

// NewFileService creates a FileService.
func NewFileService(
	fileRepo repository.FileRepository,
	storage storage.StorageBackend,
	maxUploadSize int64,
	logger *slog.Logger,
) *FileService {
	if logger == nil {
		logger = slog.Default()
	}
	return &FileService{
		fileRepo:      fileRepo,
		storage:       storage,
		maxUploadSize: maxUploadSize,
		logger:        logger,
	}
}

// Upload stores a file's content and creates its metadata record.
// The MIME type is always detected server-side from the file content.
func (s *FileService) Upload(ctx context.Context, userID string, parentID *string, fileName string, reader io.Reader) (*FileInfo, error) {
	if err := validateFileName(fileName); err != nil {
		return nil, err
	}

	if parentID != nil {
		if err := s.verifyParent(ctx, userID, *parentID); err != nil {
			return nil, err
		}
	}

	// Check for name conflicts in the target directory.
	if _, err := s.fileRepo.FindByNameAndParent(ctx, userID, parentID, fileName); err == nil {
		return nil, &model.AppError{Status: http.StatusConflict, Message: "a file with this name already exists in this directory"}
	} else if !errors.Is(err, model.ErrNotFound) {
		return nil, model.NewInternalError("check name conflict", err)
	}

	// Limit the reader if a max upload size is configured.
	if s.maxUploadSize > 0 {
		reader = io.LimitReader(reader, s.maxUploadSize+1)
	}

	// Detect MIME type from file content.
	head := make([]byte, 512)
	n, readErr := io.ReadFull(reader, head)
	if readErr != nil && readErr != io.ErrUnexpectedEOF && readErr != io.EOF {
		return nil, model.NewInternalError("read for mime detection", readErr)
	}
	head = head[:n]
	mimeType := mimetype.Detect(head).String()
	reader = io.MultiReader(bytes.NewReader(head), reader)

	fileID := uuid.NewString()
	storagePath := fmt.Sprintf("%s/%s", userID, fileID)

	// Compute SHA-256 hash while writing to storage.
	hasher := sha256.New()
	teeReader := io.TeeReader(reader, hasher)

	written, err := s.storage.Save(ctx, storagePath, teeReader)
	if err != nil {
		return nil, model.NewInternalError("save file", err)
	}

	if s.maxUploadSize > 0 && written > s.maxUploadSize {
		// Clean up the oversized file.
		_ = s.storage.Delete(ctx, storagePath)
		return nil, &model.AppError{Status: http.StatusBadRequest, Message: fmt.Sprintf("file size exceeds the maximum allowed size of %d bytes", s.maxUploadSize)}
	}

	file := &model.File{
		ID:          fileID,
		UserID:      userID,
		ParentID:    parentID,
		Name:        fileName,
		Size:        written,
		MimeType:    mimeType,
		StoragePath: storagePath,
		Hash:        hex.EncodeToString(hasher.Sum(nil)),
		Status:      model.StatusActive,
		IsDir:       false,
	}

	if err := s.fileRepo.Create(ctx, file); err != nil {
		// Compensate: remove the stored file.
		_ = s.storage.Delete(ctx, storagePath)
		if errors.Is(err, model.ErrDuplicate) {
			return nil, &model.AppError{Status: http.StatusConflict, Message: "a file with this name already exists in this directory"}
		}
		return nil, model.NewInternalError("create file record", err)
	}

	return modelToFileInfo(file), nil
}

// Download returns a reader for the file's content and its metadata.
// The caller must close the returned ReadCloser.
func (s *FileService) Download(ctx context.Context, userID, fileID string) (io.ReadCloser, *FileInfo, error) {
	file, err := s.getOwnedFile(ctx, userID, fileID)
	if err != nil {
		return nil, nil, err
	}

	if file.IsDir {
		return nil, nil, &model.AppError{Status: http.StatusBadRequest, Message: "cannot download a directory"}
	}

	reader, err := s.storage.Open(ctx, file.StoragePath)
	if err != nil {
		return nil, nil, model.NewInternalError("open file", err)
	}

	return reader, modelToFileInfo(file), nil
}

// Get returns metadata for a single file or directory.
func (s *FileService) Get(ctx context.Context, userID, fileID string) (*FileInfo, error) {
	file, err := s.getOwnedFile(ctx, userID, fileID)
	if err != nil {
		return nil, err
	}
	return modelToFileInfo(file), nil
}

// List returns the contents of a directory with pagination.
func (s *FileService) List(ctx context.Context, userID string, parentID *string, offset, limit int) (*FileList, error) {
	if parentID != nil {
		if err := s.verifyParent(ctx, userID, *parentID); err != nil {
			return nil, err
		}
	}

	files, total, err := s.fileRepo.FindByParentIDPaginated(ctx, userID, parentID, offset, limit)
	if err != nil {
		return nil, model.NewInternalError("list files", err)
	}

	infos := make([]FileInfo, 0, len(files))
	for i := range files {
		infos = append(infos, *modelToFileInfo(&files[i]))
	}

	return &FileList{Files: infos, Total: total}, nil
}

// Update renames or moves a file or directory.
func (s *FileService) Update(ctx context.Context, userID, fileID string, newName string, newParentID *string) (*FileInfo, error) {
	file, err := s.getOwnedFile(ctx, userID, fileID)
	if err != nil {
		return nil, err
	}

	// Validate new name if provided.
	if newName != "" {
		if err := validateFileName(newName); err != nil {
			return nil, err
		}
		file.Name = newName
	}

	// If parent is changing, verify the new parent.
	if newParentID != nil {
		if *newParentID == fileID {
			return nil, &model.AppError{Status: http.StatusBadRequest, Message: "cannot move a directory into itself"}
		}
		if err := s.verifyParent(ctx, userID, *newParentID); err != nil {
			return nil, err
		}
		file.ParentID = newParentID
	}

	// Check for name conflicts in the target directory.
	if newName != "" || newParentID != nil {
		conflict, err := s.fileRepo.FindByNameAndParent(ctx, userID, file.ParentID, file.Name)
		if err == nil && conflict.ID != fileID {
			return nil, &model.AppError{Status: http.StatusConflict, Message: "a file with this name already exists in the target directory"}
		} else if err != nil && !errors.Is(err, model.ErrNotFound) {
			return nil, model.NewInternalError("check name conflict", err)
		}
	}

	if err := s.fileRepo.Update(ctx, file); err != nil {
		if errors.Is(err, model.ErrDuplicate) {
			return nil, &model.AppError{Status: http.StatusConflict, Message: "a file with this name already exists in the target directory"}
		}
		return nil, model.NewInternalError("update file", err)
	}

	return modelToFileInfo(file), nil
}

// Delete soft-deletes a file or (empty) directory. Storage content is preserved.
func (s *FileService) Delete(ctx context.Context, userID, fileID string) error {
	file, err := s.getOwnedFile(ctx, userID, fileID)
	if err != nil {
		var ae *model.AppError
		if errors.As(err, &ae) && errors.Is(ae.Err, model.ErrNotFound) {
			return nil // idempotent: already deleted
		}
		return err
	}

	// Directories must be empty before deletion.
	if file.IsDir {
		_, total, err := s.fileRepo.FindByParentIDPaginated(ctx, userID, &fileID, 0, 1)
		if err != nil {
			return model.NewInternalError("check directory contents", err)
		}
		if total > 0 {
			return &model.AppError{Status: http.StatusConflict, Message: "directory is not empty"}
		}
	}

	if err := s.fileRepo.Delete(ctx, fileID); err != nil {
		return model.NewInternalError("delete file record", err)
	}

	return nil
}

// CreateDir creates a new directory.
func (s *FileService) CreateDir(ctx context.Context, userID string, parentID *string, dirName string) (*FileInfo, error) {
	if err := validateFileName(dirName); err != nil {
		return nil, err
	}

	if parentID != nil {
		if err := s.verifyParent(ctx, userID, *parentID); err != nil {
			return nil, err
		}
	}

	// Check for name conflicts in the target directory.
	if _, err := s.fileRepo.FindByNameAndParent(ctx, userID, parentID, dirName); err == nil {
		return nil, &model.AppError{Status: http.StatusConflict, Message: "a file with this name already exists in this directory"}
	} else if !errors.Is(err, model.ErrNotFound) {
		return nil, model.NewInternalError("check name conflict", err)
	}

	dir := &model.File{
		ID:       uuid.NewString(),
		UserID:   userID,
		ParentID: parentID,
		Name:     dirName,
		Status:   model.StatusActive,
		IsDir:    true,
	}

	if err := s.fileRepo.Create(ctx, dir); err != nil {
		if errors.Is(err, model.ErrDuplicate) {
			return nil, &model.AppError{Status: http.StatusConflict, Message: "a file with this name already exists in this directory"}
		}
		return nil, model.NewInternalError("create directory record", err)
	}

	return modelToFileInfo(dir), nil
}

// getOwnedFile fetches a file and verifies ownership.
func (s *FileService) getOwnedFile(ctx context.Context, userID, fileID string) (*model.File, error) {
	file, err := s.fileRepo.FindByID(ctx, fileID)
	if err != nil {
		if errors.Is(err, model.ErrNotFound) {
			return nil, &model.AppError{Status: http.StatusNotFound, Message: "file not found", Err: model.ErrNotFound}
		}
		return nil, model.NewInternalError("find file", err)
	}

	if file.UserID != userID {
		return nil, &model.AppError{Status: http.StatusForbidden, Message: "access denied", Err: model.ErrForbidden}
	}

	return file, nil
}

// verifyParent checks that a parent directory exists, belongs to the user, and is a directory.
func (s *FileService) verifyParent(ctx context.Context, userID, parentID string) error {
	parent, err := s.fileRepo.FindByID(ctx, parentID)
	if err != nil {
		if errors.Is(err, model.ErrNotFound) {
			return &model.AppError{Status: http.StatusNotFound, Message: "parent directory not found"}
		}
		return model.NewInternalError("find parent", err)
	}

	if parent.UserID != userID {
		return &model.AppError{Status: http.StatusForbidden, Message: "access denied", Err: model.ErrForbidden}
	}

	if !parent.IsDir {
		return &model.AppError{Status: http.StatusBadRequest, Message: "parent is not a directory"}
	}

	return nil
}

// validateFileName rejects names that are empty, contain path separators,
// null bytes, or are reserved names "." and "..".
func validateFileName(name string) error {
	if name == "" {
		return &model.AppError{Status: http.StatusBadRequest, Message: "file name must not be empty"}
	}
	if strings.ContainsAny(name, "/\\\x00") {
		return &model.AppError{Status: http.StatusBadRequest, Message: "file name contains invalid characters"}
	}
	if name == "." || name == ".." {
		return &model.AppError{Status: http.StatusBadRequest, Message: fmt.Sprintf("file name is reserved: %s", name)}
	}
	return nil
}

// modelToFileInfo converts a model.File to a FileInfo DTO.
func modelToFileInfo(f *model.File) *FileInfo {
	return &FileInfo{
		ID:        f.ID,
		UserID:    f.UserID,
		ParentID:  f.ParentID,
		Name:      f.Name,
		Size:      f.Size,
		MimeType:  f.MimeType,
		IsDir:     f.IsDir,
		Hash:      f.Hash,
		CreatedAt: f.CreatedAt,
		UpdatedAt: f.UpdatedAt,
	}
}