Wauplin HF staff commited on
Commit
368ace1
1 Parent(s): 559fc40

Update user_history.py

Browse files
Files changed (1) hide show
  1. user_history.py +0 -70
user_history.py CHANGED
@@ -15,7 +15,6 @@ Useful links:
15
  - Source file: https://huggingface.co/spaces/Wauplin/gradio-user-history/blob/main/user_history.py
16
  - Discussions: https://huggingface.co/spaces/Wauplin/gradio-user-history/discussions
17
  """
18
- import hashlib
19
  import json
20
  import os
21
  import shutil
@@ -38,12 +37,6 @@ def setup(folder_path: str | Path | None = None) -> None:
38
  user_history.folder_path = _resolve_folder_path(folder_path)
39
  user_history.initialized = True
40
 
41
- # Clean duplicates
42
- try:
43
- _clean_duplicates()
44
- except Exception as e:
45
- print(f"Failed to clean duplicates: {e}")
46
-
47
 
48
  def render() -> None:
49
  user_history = _UserHistory()
@@ -428,66 +421,3 @@ def _fetch_admins() -> List[str]:
428
  if response.status_code == 200:
429
  return sorted((member["user"] for member in response.json()), key=lambda x: x.lower())
430
  return [namespace]
431
-
432
-
433
- #######
434
- #######
435
-
436
- # TODO: remove this once from IllusionDiffusion once cleaned
437
-
438
-
439
- def _clean_duplicates() -> None:
440
- user_history = _UserHistory()
441
- if not (user_history.initialized and user_history.folder_path.exists()):
442
- # Must be initialized correctly
443
- return
444
-
445
- _lock = user_history.folder_path / "_clean_duplicates.lock"
446
- _is_done_file = user_history.folder_path / "_clean_duplicates_is_done" # Only 1 replica will do it, once for all
447
-
448
- with FileLock(_lock):
449
- if _is_done_file.exists(): # if True, another replica already did it
450
- return
451
-
452
- for subpath in user_history.folder_path.iterdir():
453
- if subpath.is_file():
454
- continue
455
-
456
- history_file = subpath / "history.jsonl"
457
- if not history_file.exists():
458
- continue
459
-
460
- # Read history
461
- images = [json.loads(line) for line in history_file.read_text().splitlines()]
462
-
463
- # Select unique images
464
- curated_images = []
465
- seen_hashes = set()
466
- seen_paths = set()
467
- for image in images:
468
- image_hash = _file_hash(Path(image["path"]))
469
- if image_hash is None:
470
- continue
471
- if image_hash in seen_hashes:
472
- continue
473
- seen_hashes.add(image_hash)
474
- seen_paths.add(Path(image["path"]))
475
- curated_images.append(image)
476
-
477
- # Remove duplicates + save history
478
- for path in subpath.glob("images/*"):
479
- if path not in seen_paths:
480
- try:
481
- path.unlink()
482
- except OSError:
483
- pass
484
- history_file.write_text("\n".join(json.dumps(image) for image in curated_images))
485
-
486
- _is_done_file.touch()
487
-
488
-
489
- def _file_hash(path: Path) -> str | None:
490
- """Return the hash of a file. No need to read by chunks."""
491
- if path.is_file():
492
- return hashlib.md5(path.read_bytes()).hexdigest()
493
- return None
 
15
  - Source file: https://huggingface.co/spaces/Wauplin/gradio-user-history/blob/main/user_history.py
16
  - Discussions: https://huggingface.co/spaces/Wauplin/gradio-user-history/discussions
17
  """
 
18
  import json
19
  import os
20
  import shutil
 
37
  user_history.folder_path = _resolve_folder_path(folder_path)
38
  user_history.initialized = True
39
 
 
 
 
 
 
 
40
 
41
  def render() -> None:
42
  user_history = _UserHistory()
 
421
  if response.status_code == 200:
422
  return sorted((member["user"] for member in response.json()), key=lambda x: x.lower())
423
  return [namespace]