Exclude special dotfiles from dandi digest -d zarr-checksum

dandi/cli/tests/test_digest.py

@@ -1,5 +1,6 @@
 import os
 from pathlib import Path
+import subprocess
 
 from click.testing import CliRunner
 import numpy as np
@@ -62,3 +63,20 @@ def test_digest_empty_zarr(tmp_path: Path) -> None:
         r = runner.invoke(digest, ["--digest", "zarr-checksum", "empty.zarr"])
         assert r.exit_code == 0
         assert r.output == "empty.zarr: 481a2f77ab786a0f45aafd5db0971caa-0--0\n"
+
+
+def test_digest_zarr_with_excluded_dotfiles():
+    # This test assumes that the Zarr serialization format never changes
+    runner = CliRunner()
+    with runner.isolated_filesystem():
+        dt = np.dtype("<i8")
+        zarr.save(
+            "sample.zarr", np.arange(1000, dtype=dt), np.arange(1000, 0, -1, dtype=dt)
+        )
+        subprocess.run(["git", "init"], cwd="sample.zarr", check=True)
+        os.mkdir("sample.zarr/.dandi")
+        Path("sample.zarr", ".dandi", "somefile.txt").touch()
+        Path("sample.zarr", ".gitattributes").touch()
+        r = runner.invoke(digest, ["--digest", "zarr-checksum", "sample.zarr"])
+        assert r.exit_code == 0
+        assert r.output == "sample.zarr: 4313ab36412db2981c3ed391b38604d6-5--1516\n"

dandi/support/digests.py

@@ -23,7 +23,7 @@
 from fscacher import PersistentCache
 
 from .threaded_walk import threaded_walk
-from ..utils import auto_repr
+from ..utils import auto_repr, exclude_from_zarr
 
 lgr = logging.getLogger("dandi.support.digests")
 
@@ -124,7 +124,9 @@ def digest_file(f: Path) -> Tuple[Path, str, int]:
         return (f, dgst, os.path.getsize(f))
 
     zcc = ZCTree()
-    for p, digest, size in threaded_walk(path, digest_file):
+    for p, digest, size in threaded_walk(
+        path, digest_file, exclude=lambda p: p.parent == path and exclude_from_zarr(p)
+    ):
         zcc.add(p.relative_to(path), digest, size)
     return zcc.get_digest()
 

dandi/support/threaded_walk.py

@@ -28,6 +28,7 @@ def threaded_walk(
     dirpath: Union[str, Path],
     func: Optional[Callable[[Path], Any]] = None,
     threads: int = 60,
+    exclude: Optional[Callable[[Path], Any]] = None,
 ) -> Iterable[Any]:
     if not os.path.isdir(dirpath):
         return
@@ -54,7 +55,9 @@ def worker() -> None:
                     break
             try:
                 for p in path.iterdir():
-                    if p.is_dir():
+                    if exclude is not None and exclude(p):
+                        log.debug("Excluding %s from traversal", p)
+                    elif p.is_dir():
                         with lock:
                             tasks += 1
                             paths.append(p)

dandi/utils.py

@@ -788,3 +788,11 @@ def is_page2_url(page1: str, page2: str) -> bool:
     bits2 = urlparse(page2)
     params2 = parse_qs(bits2.query)
     return (bits1[:3], params1, bits1.fragment) == (bits2[:3], params2, bits2.fragment)
+
+
+def exclude_from_zarr(path: Path) -> bool:
+    """
+    Returns `True` if the ``path`` is a file or directory that should be
+    excluded from consideration when located at the root of a Zarr
+    """
+    return path.name in (".dandi", ".datalad", ".git", ".gitattributes", ".gitmodules")