Files
scratchbox/internal/storage/filesystem.go
T
2026-05-31 20:17:49 -05:00

366 lines
8.5 KiB
Go

package storage
import (
"compress/gzip"
"context"
"crypto/rand"
"encoding/hex"
"encoding/json"
"errors"
"fmt"
"io"
"os"
"path/filepath"
"strings"
"sync"
"time"
)
const (
indexFilename = "metadata.json"
filesDirName = "scratches"
idBytes = 12
maxCreateIDAttempts = 100
)
var ErrNotFound = errors.New("scratch not found")
type Metadata struct {
ID string `json:"id"`
FilePath string `json:"file_path"`
CreatedAt time.Time `json:"created_at"`
ExpiresAt time.Time `json:"expires_at"`
ContentType string `json:"content_type"`
OriginalName string `json:"original_name,omitempty"`
Size int64 `json:"size"`
}
type FilesystemStore struct {
dataDir string
filesDir string
index string
idGenerator func() (string, error)
mu sync.RWMutex
metadata map[string]Metadata
reservedIDs map[string]struct{}
}
func NewFilesystemStore(dataDir string) (*FilesystemStore, error) {
filesDir := filepath.Join(dataDir, filesDirName)
if err := os.MkdirAll(filesDir, 0o755); err != nil {
return nil, fmt.Errorf("create files dir: %w", err)
}
store := &FilesystemStore{
dataDir: dataDir,
filesDir: filesDir,
index: filepath.Join(dataDir, indexFilename),
idGenerator: generateID,
metadata: map[string]Metadata{},
reservedIDs: map[string]struct{}{},
}
if err := store.loadIndex(); err != nil {
return nil, err
}
if err := store.reconcileOnStartup(time.Now()); err != nil {
return nil, err
}
return store, nil
}
func (s *FilesystemStore) Create(ctx context.Context, body io.Reader, contentType string, ttl time.Duration) (Metadata, error) {
return s.CreateWithOriginalName(ctx, body, contentType, "", ttl)
}
func (s *FilesystemStore) CreateWithOriginalName(ctx context.Context, body io.Reader, contentType string, originalName string, ttl time.Duration) (Metadata, error) {
for attempt := 0; attempt < maxCreateIDAttempts; attempt++ {
id, err := s.idGenerator()
if err != nil {
return Metadata{}, fmt.Errorf("generate id: %w", err)
}
finalPath := filepath.Join(s.filesDir, id)
s.mu.RLock()
_, exists := s.metadata[id]
_, reserved := s.reservedIDs[id]
s.mu.RUnlock()
if exists || reserved {
continue
}
if _, err := os.Stat(finalPath); err == nil {
continue
} else if !errors.Is(err, os.ErrNotExist) {
return Metadata{}, fmt.Errorf("stat candidate scratch path: %w", err)
}
s.mu.Lock()
if _, exists := s.metadata[id]; exists {
s.mu.Unlock()
continue
}
if _, reserved := s.reservedIDs[id]; reserved {
s.mu.Unlock()
continue
}
// Reserve the ID before consuming the request body to avoid
// concurrent writers choosing the same ID.
s.reservedIDs[id] = struct{}{}
s.mu.Unlock()
tempFile, err := os.CreateTemp(s.filesDir, id+".tmp-*")
if err != nil {
s.mu.Lock()
delete(s.reservedIDs, id)
s.mu.Unlock()
return Metadata{}, fmt.Errorf("create temp file: %w", err)
}
gzipWriter := gzip.NewWriter(tempFile)
size, copyErr := io.Copy(gzipWriter, body)
gzipCloseErr := gzipWriter.Close()
closeErr := tempFile.Close()
if copyErr != nil {
_ = os.Remove(tempFile.Name())
s.mu.Lock()
delete(s.reservedIDs, id)
s.mu.Unlock()
return Metadata{}, fmt.Errorf("write scratch content: %w", copyErr)
}
if gzipCloseErr != nil {
_ = os.Remove(tempFile.Name())
s.mu.Lock()
delete(s.reservedIDs, id)
s.mu.Unlock()
return Metadata{}, fmt.Errorf("close gzip writer: %w", gzipCloseErr)
}
if closeErr != nil {
_ = os.Remove(tempFile.Name())
s.mu.Lock()
delete(s.reservedIDs, id)
s.mu.Unlock()
return Metadata{}, fmt.Errorf("close temp file: %w", closeErr)
}
if err := os.Rename(tempFile.Name(), finalPath); err != nil {
_ = os.Remove(tempFile.Name())
s.mu.Lock()
delete(s.reservedIDs, id)
s.mu.Unlock()
return Metadata{}, fmt.Errorf("atomically move scratch file: %w", err)
}
now := time.Now().UTC()
meta := Metadata{
ID: id,
FilePath: finalPath,
CreatedAt: now,
ExpiresAt: now.Add(ttl),
ContentType: contentType,
OriginalName: normalizeOriginalName(originalName),
Size: size,
}
s.mu.Lock()
delete(s.reservedIDs, id)
s.metadata[id] = meta
if err := s.persistLocked(); err != nil {
delete(s.metadata, id)
s.mu.Unlock()
_ = os.Remove(finalPath)
return Metadata{}, err
}
s.mu.Unlock()
return meta, nil
}
return Metadata{}, fmt.Errorf("failed to allocate unique scratch id after %d attempts", maxCreateIDAttempts)
}
func normalizeOriginalName(originalName string) string {
name := strings.TrimSpace(originalName)
if name == "" {
return ""
}
name = filepath.Base(name)
name = strings.TrimSpace(name)
if name == "." || name == string(filepath.Separator) {
return ""
}
return name
}
func (s *FilesystemStore) Get(id string) (Metadata, error) {
s.mu.RLock()
meta, ok := s.metadata[id]
s.mu.RUnlock()
if !ok {
return Metadata{}, ErrNotFound
}
return meta, nil
}
func (s *FilesystemStore) Open(id string) (io.ReadCloser, Metadata, error) {
meta, err := s.Get(id)
if err != nil {
return nil, Metadata{}, err
}
file, err := os.Open(meta.FilePath)
if err != nil {
if errors.Is(err, os.ErrNotExist) {
return nil, Metadata{}, ErrNotFound
}
return nil, Metadata{}, fmt.Errorf("open scratch file: %w", err)
}
gzipReader, err := gzip.NewReader(file)
if err != nil {
_ = file.Close()
return nil, Metadata{}, fmt.Errorf("create gzip reader: %w", err)
}
return &combinedReadCloser{
reader: gzipReader,
closers: []io.Closer{
gzipReader,
file,
},
}, meta, nil
}
func (s *FilesystemStore) Delete(id string) error {
s.mu.Lock()
meta, ok := s.metadata[id]
if !ok {
s.mu.Unlock()
return ErrNotFound
}
delete(s.metadata, id)
if err := s.persistLocked(); err != nil {
s.metadata[id] = meta
s.mu.Unlock()
return err
}
s.mu.Unlock()
if err := os.Remove(meta.FilePath); err != nil && !errors.Is(err, os.ErrNotExist) {
return fmt.Errorf("remove scratch file: %w", err)
}
return nil
}
func (s *FilesystemStore) DeleteExpired(now time.Time) (int, error) {
s.mu.Lock()
expired := make([]Metadata, 0)
for id, meta := range s.metadata {
if !meta.ExpiresAt.After(now) {
expired = append(expired, meta)
delete(s.metadata, id)
}
}
if len(expired) == 0 {
s.mu.Unlock()
return 0, nil
}
if err := s.persistLocked(); err != nil {
for _, meta := range expired {
s.metadata[meta.ID] = meta
}
s.mu.Unlock()
return 0, err
}
s.mu.Unlock()
for _, meta := range expired {
if err := os.Remove(meta.FilePath); err != nil && !errors.Is(err, os.ErrNotExist) {
return 0, fmt.Errorf("remove expired file %s: %w", meta.ID, err)
}
}
return len(expired), nil
}
func (s *FilesystemStore) loadIndex() error {
b, err := os.ReadFile(s.index)
if err != nil {
if errors.Is(err, os.ErrNotExist) {
return nil
}
return fmt.Errorf("read metadata index: %w", err)
}
loaded := map[string]Metadata{}
if err := json.Unmarshal(b, &loaded); err != nil {
return fmt.Errorf("decode metadata index: %w", err)
}
s.metadata = loaded
return nil
}
func (s *FilesystemStore) persistLocked() error {
tmp := s.index + ".tmp"
payload, err := json.MarshalIndent(s.metadata, "", " ")
if err != nil {
return fmt.Errorf("encode metadata index: %w", err)
}
if err := os.WriteFile(tmp, payload, 0o644); err != nil {
return fmt.Errorf("write metadata temp file: %w", err)
}
if err := os.Rename(tmp, s.index); err != nil {
return fmt.Errorf("replace metadata index: %w", err)
}
return nil
}
func (s *FilesystemStore) reconcileOnStartup(now time.Time) error {
s.mu.Lock()
defer s.mu.Unlock()
changed := false
for id, meta := range s.metadata {
if !meta.ExpiresAt.After(now) {
delete(s.metadata, id)
_ = os.Remove(meta.FilePath)
changed = true
continue
}
if _, err := os.Stat(meta.FilePath); err != nil {
if errors.Is(err, os.ErrNotExist) {
delete(s.metadata, id)
changed = true
continue
}
return fmt.Errorf("stat scratch %s: %w", id, err)
}
}
if !changed {
return nil
}
return s.persistLocked()
}
func generateID() (string, error) {
b := make([]byte, idBytes)
if _, err := rand.Read(b); err != nil {
return "", err
}
return hex.EncodeToString(b), nil
}
type combinedReadCloser struct {
reader io.Reader
closers []io.Closer
}
func (c *combinedReadCloser) Read(p []byte) (int, error) {
return c.reader.Read(p)
}
func (c *combinedReadCloser) Close() error {
var firstErr error
for _, closer := range c.closers {
if err := closer.Close(); err != nil && firstErr == nil {
firstErr = err
}
}
return firstErr
}