CoCoOne commited on
Commit
0561e04
·
1 Parent(s): ec086ef

Ignore hidden placeholder files in validator

Browse files
Files changed (1) hide show
  1. validator.py +9 -2
validator.py CHANGED
@@ -138,6 +138,13 @@ def _iter_data_refs(text: str) -> list[str]:
138
  return refs
139
 
140
 
 
 
 
 
 
 
 
141
  def cleanup_work_dir(work_dir: str | Path | None) -> None:
142
  if not work_dir:
143
  return
@@ -360,7 +367,7 @@ def validate_task_dir(
360
  if any(token in description for token in STALE_TOKENS):
361
  errors.append(f'{prefix}.description still contains stale source paths or legacy directories')
362
 
363
- actual_data_files = {p for p in data_dir.rglob('*') if p.is_file()} if data_dir.exists() else set()
364
  uncovered = sorted(actual_data_files - covered_files)
365
  if uncovered:
366
  errors.append('data/ contains undeclared files: ' + ', '.join(rel(p, task_dir) for p in uncovered[:20]))
@@ -429,7 +436,7 @@ def validate_task_dir(
429
  errors.append(f'{prefix}.path does not exist: {path_value}')
430
  referenced_images.add(path_value)
431
 
432
- actual_image_files = {str(p.relative_to(target_dir)) for p in images_dir.rglob('*') if p.is_file()} if images_dir.exists() else set()
433
  extra_images = sorted(actual_image_files - referenced_images)
434
  missing_images = sorted(referenced_images - actual_image_files)
435
  if extra_images:
 
138
  return refs
139
 
140
 
141
+ def _iter_visible_files(root: Path) -> set[Path]:
142
+ return {
143
+ path for path in root.rglob('*')
144
+ if path.is_file() and not any(part.startswith('.') for part in path.relative_to(root).parts)
145
+ }
146
+
147
+
148
  def cleanup_work_dir(work_dir: str | Path | None) -> None:
149
  if not work_dir:
150
  return
 
367
  if any(token in description for token in STALE_TOKENS):
368
  errors.append(f'{prefix}.description still contains stale source paths or legacy directories')
369
 
370
+ actual_data_files = _iter_visible_files(data_dir) if data_dir.exists() else set()
371
  uncovered = sorted(actual_data_files - covered_files)
372
  if uncovered:
373
  errors.append('data/ contains undeclared files: ' + ', '.join(rel(p, task_dir) for p in uncovered[:20]))
 
436
  errors.append(f'{prefix}.path does not exist: {path_value}')
437
  referenced_images.add(path_value)
438
 
439
+ actual_image_files = {str(p.relative_to(target_dir)) for p in _iter_visible_files(images_dir)} if images_dir.exists() else set()
440
  extra_images = sorted(actual_image_files - referenced_images)
441
  missing_images = sorted(referenced_images - actual_image_files)
442
  if extra_images: