verstak/internal/core/files/file.go

683 lines
19 KiB
Go

package files
import (
"crypto/sha256"
"database/sql"
"encoding/base64"
"fmt"
"io"
"os"
"os/exec"
"path/filepath"
"runtime"
"strings"
"time"
"verstak/internal/core/nodes"
"verstak/internal/core/storage"
"verstak/internal/core/util"
)
// Record represents a file entry linked to a node.
type Record struct {
ID string `json:"id"`
NodeID string `json:"node_id"`
Filename string `json:"filename"`
Path string `json:"path"` // relative to vault root
StorageMode string `json:"storage_mode"` // "vault" | "external"
Size int64 `json:"size"`
SHA256 string `json:"sha256,omitempty"`
MIME string `json:"mime,omitempty"`
CreatedAt time.Time `json:"created_at"`
UpdatedAt time.Time `json:"updated_at"`
LastSeenAt *time.Time `json:"last_seen_at,omitempty"`
Missing bool `json:"missing"`
}
// ImportSummary describes a scanned directory before import.
type ImportSummary struct {
Files int `json:"files"`
Folders int `json:"folders"`
TotalBytes int64 `json:"totalBytes"`
IsDangerous bool `json:"isDangerous"`
DangerReason string `json:"dangerReason,omitempty"`
}
// Service provides file operations inside a vault.
type Service struct {
db *storage.DB
vaultRoot string
nodes *nodes.Repository
}
// NewService creates a file service bound to a vault.
func NewService(db *storage.DB, vaultRoot string, nodeRepo *nodes.Repository) *Service {
return &Service{db: db, vaultRoot: vaultRoot, nodes: nodeRepo}
}
// DB returns the underlying storage.
func (s *Service) DB() *storage.DB {
return s.db
}
// --- security helpers ---
// ValidateName is an exported wrapper for validateName.
func ValidateName(name string) error {
return validateName(name)
}
// validateName rejects filenames with path separators, relative components,
// and other dangerous patterns.
func validateName(name string) error {
if name == "" {
return fmt.Errorf("name is required")
}
if strings.Contains(name, "/") || strings.Contains(name, "\\") {
return fmt.Errorf("name must not contain path separators")
}
if strings.Contains(name, "..") {
return fmt.Errorf("name must not contain '..'")
}
if strings.Contains(name, "\x00") {
return fmt.Errorf("name must not contain null bytes")
}
if len(name) > 255 {
return fmt.Errorf("name too long (max 255)")
}
return nil
}
// vaultPath resolves a relative vault path and checks it stays within jail.
func (s *Service) vaultPath(rel string) (string, error) {
abs := filepath.Join(s.vaultRoot, rel)
cleaned := filepath.Clean(abs)
if !strings.HasPrefix(cleaned, filepath.Clean(s.vaultRoot)) {
return "", fmt.Errorf("path escapes vault root")
}
return cleaned, nil
}
// absPathSafe resolves an absolute path and checks jail if it's under vault.
// For "external" mode records the path stored may be an absolute external path.
// This function only checks path safety — it does not enforce that external
// files must be inside the vault.
func (s *Service) absPathSafe(rec *Record) (string, error) {
if rec.StorageMode == "vault" {
return s.vaultPath(rec.Path)
}
abs, err := filepath.Abs(rec.Path)
if err != nil {
return "", fmt.Errorf("abs: %w", err)
}
return filepath.Clean(abs), nil
}
// --- public operations ---
// AddExternal registers an external file (absolute path) without copying.
func (s *Service) AddExternal(nodeID, absPath string) (*Record, error) {
info, err := os.Stat(absPath)
if err != nil {
return nil, fmt.Errorf("stat: %w", err)
}
absPath, _ = filepath.Abs(absPath)
return s.insertRecord(nodeID, filepath.Base(absPath), absPath, "external", info.Size(), "")
}
// CopyIntoVault copies an external file into the vault.
// The file lands at <vaultRoot>/spaces/<nodeSlug>/<filename>.
func (s *Service) CopyIntoVault(nodeID, absPath, nodeSlug string) (*Record, error) {
info, err := os.Stat(absPath)
if err != nil {
return nil, fmt.Errorf("stat: %w", err)
}
if nodeSlug == "" {
nodeSlug = nodeID[:8]
}
destDir := filepath.Join(s.vaultRoot, "spaces", nodeSlug)
if _, err := s.vaultPath(filepath.Join("spaces", nodeSlug)); err != nil {
return nil, fmt.Errorf("path safety: %w", err)
}
if err := os.MkdirAll(destDir, 0o750); err != nil {
return nil, fmt.Errorf("mkdir: %w", err)
}
filename := filepath.Base(absPath)
dest := filepath.Join(destDir, filename)
// If destination exists, add a numeric suffix.
if _, err := os.Stat(dest); err == nil {
ext := filepath.Ext(filename)
name := strings.TrimSuffix(filename, ext)
dest = filepath.Join(destDir, fmt.Sprintf("%s_%d%s", name, time.Now().Unix(), ext))
filename = filepath.Base(dest)
}
hash, err := copyAndHash(absPath, dest)
if err != nil {
return nil, fmt.Errorf("copy: %w", err)
}
relPath, _ := filepath.Rel(s.vaultRoot, dest)
if _, err := s.vaultPath(relPath); err != nil {
return nil, fmt.Errorf("path safety: %w", err)
}
return s.insertRecord(nodeID, filename, relPath, "vault", info.Size(), hash)
}
// Get returns a file record by ID.
func (s *Service) Get(id string) (*Record, error) {
row := s.db.QueryRow(
`SELECT id,node_id,filename,path,storage_mode,size,sha256,mime,
created_at,updated_at,last_seen_at,missing
FROM files WHERE id = ?`, id)
return scanRecord(row)
}
// ListByNode returns all files linked to a node.
func (s *Service) ListByNode(nodeID string) ([]Record, error) {
rows, err := s.db.Query(
`SELECT id,node_id,filename,path,storage_mode,size,sha256,mime,
created_at,updated_at,last_seen_at,missing
FROM files WHERE node_id = ? ORDER BY created_at`, nodeID)
if err != nil {
return nil, err
}
defer rows.Close()
return scanRecords(rows)
}
// MarkMissing flags a file as missing.
func (s *Service) MarkMissing(id string, missing bool) error {
m := 0
if missing {
m = 1
}
_, err := s.db.Exec(
`UPDATE files SET missing=?, updated_at=? WHERE id=?`,
m, time.Now().UTC().Format(time.RFC3339), id)
return err
}
// DeleteToTrash moves a vault file to .verstak/trash/ and removes the record.
func (s *Service) DeleteToTrash(id string) error {
rec, err := s.Get(id)
if err != nil {
return err
}
if rec.StorageMode == "vault" {
src, err := s.vaultPath(rec.Path)
if err != nil {
return err
}
trashDir := filepath.Join(s.vaultRoot, ".verstak", "trash")
if err := os.MkdirAll(trashDir, 0o750); err != nil {
return err
}
dest := filepath.Join(trashDir, rec.ID+"_"+rec.Filename)
// verify trash is inside vault
if _, err := s.vaultPath(filepath.Join(".verstak", "trash", rec.ID+"_"+rec.Filename)); err != nil {
return err
}
if err := os.Rename(src, dest); err != nil {
return fmt.Errorf("move to trash: %w", err)
}
}
_, err = s.db.Exec("DELETE FROM files WHERE id=?", id)
return err
}
// Open launches the file with the system default application.
func (s *Service) Open(id string) error {
rec, err := s.Get(id)
if err != nil {
return err
}
abs, err := s.absPathSafe(rec)
if err != nil {
return err
}
return openWithSystem(abs)
}
// maxPreviewSize is the maximum file size (5 MB) for inline preview.
const maxPreviewSize = 5 * 1024 * 1024
// ReadText reads a file's content as text, up to maxPreviewSize.
func (s *Service) ReadText(id string) (string, error) {
rec, err := s.Get(id)
if err != nil {
return "", err
}
if rec.Size > maxPreviewSize {
return "", fmt.Errorf("file too large for preview (%d bytes)", rec.Size)
}
abs, err := s.absPathSafe(rec)
if err != nil {
return "", err
}
b, err := os.ReadFile(abs)
if err != nil {
return "", fmt.Errorf("read: %w", err)
}
return string(b), nil
}
// ReadBase64 reads a file and returns a data URI (base64-encoded).
func (s *Service) ReadBase64(id string) (string, error) {
rec, err := s.Get(id)
if err != nil {
return "", err
}
if rec.Size > maxPreviewSize {
return "", fmt.Errorf("file too large for preview (%d bytes)", rec.Size)
}
abs, err := s.absPathSafe(rec)
if err != nil {
return "", err
}
b, err := os.ReadFile(abs)
if err != nil {
return "", fmt.Errorf("read: %w", err)
}
mime := rec.MIME
if mime == "" {
mime = "application/octet-stream"
}
return fmt.Sprintf("data:%s;base64,%s", mime, base64.StdEncoding.EncodeToString(b)), nil
}
// CreateEmptyFile creates a file node and an empty vault file.
func (s *Service) CreateEmptyFile(parentID, filename string) (*nodes.Node, error) {
if err := validateName(filename); err != nil {
return nil, fmt.Errorf("invalid filename: %w", err)
}
filename = s.uniqueTitle(parentID, filename)
node, err := s.nodes.Create(parentID, nodes.TypeFile, filename, "")
if err != nil {
return nil, err
}
dir := filepath.Join(s.vaultRoot, "spaces", node.Slug)
if err := os.MkdirAll(dir, 0o750); err != nil {
return nil, fmt.Errorf("mkdir: %w", err)
}
dest := filepath.Join(dir, filename)
f, err := os.Create(dest)
if err != nil {
return nil, fmt.Errorf("create file: %w", err)
}
f.Close()
relPath, _ := filepath.Rel(s.vaultRoot, dest)
// Verify dest is inside vault
if _, err := s.vaultPath(relPath); err != nil {
return nil, fmt.Errorf("path safety: %w", err)
}
_, err = s.insertRecord(node.ID, filename, relPath, "vault", 0, "")
return node, err
}
// Duplicate creates a copy of a node and its file record under the same parent.
func (s *Service) Duplicate(nodeID string) (*nodes.Node, error) {
original, err := s.nodes.GetActive(nodeID)
if err != nil {
return nil, err
}
parentID := ""
if original.ParentID != nil {
parentID = *original.ParentID
}
newName := s.copyTitle(parentID, original.Title)
node, err := s.nodes.Create(parentID, original.Type, newName, original.Section)
if err != nil {
return nil, err
}
if original.Type == nodes.TypeFile {
records, _ := s.ListByNode(original.ID)
if len(records) > 0 {
src := &records[0]
if src.StorageMode == "vault" {
srcPath, err := s.vaultPath(src.Path)
if err != nil {
return nil, err
}
dir := filepath.Join(s.vaultRoot, "spaces", node.Slug)
os.MkdirAll(dir, 0o750)
dst := filepath.Join(dir, newName)
hash, err := copyAndHash(srcPath, dst)
if err != nil {
return nil, fmt.Errorf("copy file: %w", err)
}
relPath, _ := filepath.Rel(s.vaultRoot, dst)
if _, err := s.vaultPath(relPath); err != nil {
return nil, fmt.Errorf("path safety: %w", err)
}
_, err = s.insertRecord(node.ID, newName, relPath, "vault", src.Size, hash)
if err != nil {
return nil, err
}
} else {
// External file: create a new record pointing to the same absolute path.
_, err = s.insertRecord(node.ID, newName, src.Path, "external", src.Size, src.SHA256)
if err != nil {
return nil, err
}
}
}
}
return node, nil
}
// AddPathCopy copies sourcePath (file or directory) into the vault under nodeID.
func (s *Service) AddPathCopy(nodeID, sourcePath string) ([]nodes.Node, error) {
return s.importPath(nodeID, sourcePath, true)
}
// AddPathLink links sourcePath (file or directory) without copying into vault.
func (s *Service) AddPathLink(nodeID, sourcePath string) ([]nodes.Node, error) {
return s.importPath(nodeID, sourcePath, false)
}
// PreviewImport scans sourcePath and returns a summary without importing.
func (s *Service) PreviewImport(sourcePath string) (*ImportSummary, error) {
info, err := os.Stat(sourcePath)
if err != nil {
return nil, fmt.Errorf("stat: %w", err)
}
if !info.IsDir() {
return &ImportSummary{Files: 1, TotalBytes: info.Size()}, nil
}
var sum ImportSummary
err = filepath.Walk(sourcePath, func(path string, fi os.FileInfo, err error) error {
if err != nil {
return filepath.SkipDir
}
if fi.IsDir() {
sum.Folders++
name := strings.ToLower(fi.Name())
if name == ".git" || name == "node_modules" || name == ".cache" {
sum.IsDangerous = true
sum.DangerReason = fmt.Sprintf("содержит %s", fi.Name())
}
return nil
}
sum.Files++
sum.TotalBytes += fi.Size()
return nil
})
if sum.Files > 1000 && !sum.IsDangerous {
sum.IsDangerous = true
sum.DangerReason = "более 1000 файлов"
}
if sum.TotalBytes > 1<<30 && !sum.IsDangerous {
sum.IsDangerous = true
sum.DangerReason = "более 1 GB"
}
return &sum, err
}
// DeleteNodeAndChildren soft-deletes a node and all descendants,
// moving vault files to trash.
func (s *Service) DeleteNodeAndChildren(nodeID string) error {
children, _ := s.nodes.ListChildren(nodeID, false)
for i := range children {
if err := s.DeleteNodeAndChildren(children[i].ID); err != nil {
return err
}
}
_ = s.deleteFileRecords(nodeID)
return s.nodes.SoftDelete(nodeID)
}
func (s *Service) deleteFileRecords(nodeID string) error {
records, err := s.ListByNode(nodeID)
if err != nil {
return err
}
for _, r := range records {
_ = s.DeleteToTrash(r.ID)
}
return nil
}
func (s *Service) importPath(parentID, sourcePath string, copyMode bool) ([]nodes.Node, error) {
info, err := os.Stat(sourcePath)
if err != nil {
return nil, fmt.Errorf("stat: %w", err)
}
if !info.IsDir() {
title := s.uniqueTitle(parentID, filepath.Base(sourcePath))
node, err := s.nodes.Create(parentID, nodes.TypeFile, title, "")
if err != nil {
return nil, err
}
if copyMode {
_, err = s.CopyIntoVault(node.ID, sourcePath, node.Slug)
} else {
_, err = s.AddExternal(node.ID, sourcePath)
}
if err != nil {
return nil, err
}
return []nodes.Node{*node}, nil
}
return s.importDir(parentID, sourcePath, info.Name(), copyMode)
}
func (s *Service) importDir(parentID, sourcePath, dirName string, copyMode bool) ([]nodes.Node, error) {
dirName = s.uniqueTitle(parentID, dirName)
folderNode, err := s.nodes.Create(parentID, nodes.TypeFolder, dirName, "")
if err != nil {
return nil, err
}
entries, err := os.ReadDir(sourcePath)
if err != nil {
return nil, err
}
var all []nodes.Node
all = append(all, *folderNode)
for _, entry := range entries {
childPath := filepath.Join(sourcePath, entry.Name())
if entry.IsDir() {
children, err := s.importDir(folderNode.ID, childPath, entry.Name(), copyMode)
if err != nil {
return nil, err
}
all = append(all, children...)
} else {
childNode, err := s.nodes.Create(folderNode.ID, nodes.TypeFile, entry.Name(), "")
if err != nil {
return nil, err
}
if copyMode {
_, err = s.CopyIntoVault(childNode.ID, childPath, childNode.Slug)
} else {
_, err = s.AddExternal(childNode.ID, childPath)
}
if err != nil {
return nil, err
}
all = append(all, *childNode)
}
}
return all, nil
}
func (s *Service) uniqueTitle(parentID, desired string) string {
children, _ := s.nodes.ListChildren(parentID, false)
used := make(map[string]bool, len(children))
for i := range children {
used[children[i].Title] = true
}
if !used[desired] {
return desired
}
for n := 2; ; n++ {
c := fmt.Sprintf("%s (%d)", desired, n)
if !used[c] {
return c
}
}
}
// copyTitle generates a unique "Name (copy).ext" style name for duplicates.
// For files with extensions: "photo.jpg" → "photo (copy).jpg", "photo (copy 2).jpg"
// For folders: "Docs" → "Docs (copy)", "Docs (copy 2)"
func (s *Service) copyTitle(parentID, desired string) string {
children, _ := s.nodes.ListChildren(parentID, false)
used := make(map[string]bool, len(children))
for i := range children {
used[children[i].Title] = true
}
ext := filepath.Ext(desired)
base := strings.TrimSuffix(desired, ext)
copyName := base + " (copy)" + ext
if !used[copyName] {
return copyName
}
for n := 2; ; n++ {
candidate := fmt.Sprintf("%s (copy %d)%s", base, n, ext)
if !used[candidate] {
return candidate
}
}
}
// UniqueTitleCopy returns a copy-style unique name for use in conflict resolution.
func (s *Service) UniqueTitleCopy(parentID, desired string) string {
return s.copyTitle(parentID, desired)
}
// --- implementation details ---
func (s *Service) insertRecord(nodeID, filename, path, mode string, size int64, sha string) (*Record, error) {
rec := &Record{
ID: util.UUID7(),
NodeID: nodeID,
Filename: filename,
Path: path,
StorageMode: mode,
Size: size,
SHA256: sha,
MIME: guessMIME(filename),
CreatedAt: time.Now().UTC(),
UpdatedAt: time.Now().UTC(),
}
_, err := s.db.Exec(
`INSERT INTO files (id,node_id,filename,path,storage_mode,size,sha256,mime,
created_at,updated_at,missing)
VALUES (?,?,?,?,?,?,?,?,?,?,0)`,
rec.ID, rec.NodeID, rec.Filename, rec.Path, rec.StorageMode,
rec.Size, rec.SHA256, rec.MIME,
rec.CreatedAt.Format(time.RFC3339), rec.UpdatedAt.Format(time.RFC3339))
if err != nil {
return nil, err
}
return rec, nil
}
func copyAndHash(src, dest string) (string, error) {
in, err := os.Open(src)
if err != nil {
return "", err
}
defer in.Close()
out, err := os.Create(dest)
if err != nil {
return "", err
}
defer out.Close()
h := sha256.New()
if _, err := io.Copy(io.MultiWriter(out, h), in); err != nil {
return "", err
}
return fmt.Sprintf("%x", h.Sum(nil)), nil
}
func guessMIME(filename string) string {
ext := strings.ToLower(filepath.Ext(filename))
switch ext {
case ".md", ".txt", ".go", ".py", ".js", ".ts", ".sh", ".sql", ".yml", ".yaml", ".json", ".toml", ".xml", ".html", ".css", ".csv", ".rst":
return "text/plain"
case ".png":
return "image/png"
case ".jpg", ".jpeg":
return "image/jpeg"
case ".gif":
return "image/gif"
case ".pdf":
return "application/pdf"
case ".docx":
return "application/vnd.openxmlformats-officedocument.wordprocessingml.document"
case ".xlsx":
return "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
case ".odt":
return "application/vnd.oasis.opendocument.text"
case ".zip":
return "application/zip"
}
return "application/octet-stream"
}
func openWithSystem(path string) error {
var cmd *exec.Cmd
switch runtime.GOOS {
case "linux":
cmd = exec.Command("xdg-open", path)
case "darwin":
cmd = exec.Command("open", path)
case "windows":
cmd = exec.Command("cmd", "/c", "start", "", path)
default:
return fmt.Errorf("unsupported platform")
}
return cmd.Start()
}
// --- scanning helpers ---
type scanFace interface {
Scan(dest ...interface{}) error
}
func scanRecord(s scanFace) (*Record, error) {
var r Record
var lastSeen sql.NullString
var createdStr, updatedStr string
err := s.Scan(
&r.ID, &r.NodeID, &r.Filename, &r.Path, &r.StorageMode,
&r.Size, &r.SHA256, &r.MIME,
&createdStr, &updatedStr, &lastSeen, &r.Missing)
if err == sql.ErrNoRows {
return nil, fmt.Errorf("file not found")
}
if err != nil {
return nil, err
}
r.CreatedAt, _ = time.Parse(time.RFC3339, createdStr)
r.UpdatedAt, _ = time.Parse(time.RFC3339, updatedStr)
if lastSeen.Valid {
t, _ := time.Parse(time.RFC3339, lastSeen.String)
r.LastSeenAt = &t
}
return &r, nil
}
func scanRecords(rows *sql.Rows) ([]Record, error) {
var out []Record
for rows.Next() {
r, err := scanRecord(rows)
if err != nil {
return nil, err
}
out = append(out, *r)
}
return out, rows.Err()
}