Avoid importing numpy

dandi/download.py

@@ -31,7 +31,6 @@
 from .dandiset import Dandiset
 from .exceptions import NotFoundError
 from .files import LocalAsset, find_dandi_files
-from .support.digests import get_digest, get_zarr_checksum
 from .support.iterators import IteratorWithAggregation
 from .support.pyout import naturalsize
 from .utils import (
@@ -552,6 +551,8 @@ def _download_file(
       possible checksums or other digests provided for the file. Only one
       will be used to verify download
     """
+    from .support.digests import get_digest
+
     if op.lexists(path):
         annex_path = op.join(toplevel_path, ".git", "annex")
         if existing is DownloadExisting.ERROR:
@@ -854,6 +855,8 @@ def _download_zarr(
     lock: Lock,
     jobs: int | None = None,
 ) -> Iterator[dict]:
+    from .support.digests import get_zarr_checksum
+
     download_gens = {}
     entries = list(asset.iterfiles())
     digests = {}

dandi/files/bases.py

@@ -27,7 +27,6 @@
 from dandi.dandiapi import RemoteAsset, RemoteDandiset, RESTFullAPIClient
 from dandi.metadata.core import get_default_metadata
 from dandi.misctypes import DUMMY_DANDI_ETAG, Digest, LocalReadableFile, P
-from dandi.support.digests import get_dandietag, get_digest
 from dandi.utils import yaml_load
 from dandi.validate_types import Scope, Severity, ValidationOrigin, ValidationResult
 
@@ -300,6 +299,8 @@ def get_metadata(
 
     def get_digest(self) -> Digest:
         """Calculate a dandi-etag digest for the asset"""
+        from dandi.support.digests import get_digest
+
         value = get_digest(self.filepath, digest="dandi-etag")
         return Digest.dandi_etag(value)
 
@@ -330,6 +331,8 @@ def iter_upload(
             ``"done"`` and an ``"asset"`` key containing the resulting
             `RemoteAsset`.
         """
+        from dandi.support.digests import get_dandietag
+
         asset_path = metadata.setdefault("path", self.path)
         client = dandiset.client
         yield {"status": "calculating etag"}

dandi/files/zarr.py

@@ -15,8 +15,6 @@
 from dandischema.digests.zarr import get_checksum
 from dandischema.models import BareAsset, DigestType
 import requests
-import zarr
-from zarr_checksum import ZarrChecksumTree
 
 from dandi import get_logger
 from dandi.consts import (
@@ -34,7 +32,6 @@
 )
 from dandi.metadata.core import get_default_metadata
 from dandi.misctypes import DUMMY_DANDI_ZARR_CHECKSUM, BasePath, Digest
-from dandi.support.digests import get_digest, get_zarr_checksum, md5file_nocache
 from dandi.utils import chunked, exclude_from_zarr, pluralize
 
 from .bases import LocalDirectoryAsset
@@ -95,6 +92,9 @@ def get_digest(self) -> Digest:
         directory, the algorithm will be the Dandi Zarr checksum algorithm; if
         it is a file, it will be MD5.
         """
+
+        from dandi.support.digests import get_digest, get_zarr_checksum
+
         if self.is_dir():
             return Digest.dandi_zarr(get_zarr_checksum(self.filepath))
         else:
@@ -151,6 +151,8 @@ def stat(self) -> ZarrStat:
         """Return various details about the Zarr asset"""
 
         def dirstat(dirpath: LocalZarrEntry) -> ZarrStat:
+            from dandi.support.digests import md5file_nocache
+
             size = 0
             dir_md5s = {}
             file_md5s = {}
@@ -175,6 +177,8 @@ def dirstat(dirpath: LocalZarrEntry) -> ZarrStat:
 
     def get_digest(self) -> Digest:
         """Calculate a dandi-zarr-checksum digest for the asset"""
+        from dandi.support.digests import get_zarr_checksum
+
         return Digest.dandi_zarr(get_zarr_checksum(self.filepath))
 
     def get_metadata(
@@ -192,6 +196,8 @@ def get_validation_errors(
         schema_version: str | None = None,
         devel_debug: bool = False,
     ) -> list[ValidationResult]:
+        import zarr
+
         errors: list[ValidationResult] = []
         try:
             data = zarr.open(str(self.filepath))
@@ -281,6 +287,10 @@ def iter_upload(
             ``"done"`` and an ``"asset"`` key containing the resulting
             `RemoteAsset`.
         """
+        # Importing zarr_checksum leads to importing numpy, which we want to
+        # avoid unless necessary
+        from zarr_checksum import ZarrChecksumTree
+
         # So that older clients don't get away with doing the wrong thing once
         # Zarr upload to embargoed Dandisets is implemented in the API:
         if dandiset.embargo_status is EmbargoStatus.EMBARGOED:
@@ -584,6 +594,8 @@ def register(self, e: LocalZarrEntry, digest: str | None = None) -> None:
 
     @staticmethod
     def _mkitem(e: LocalZarrEntry) -> UploadItem:
+        from dandi.support.digests import md5file_nocache
+
         digest = md5file_nocache(e.filepath)
         return UploadItem.from_entry(e, digest)
 
@@ -634,6 +646,8 @@ def upload_request(self) -> dict[str, str]:
 def _cmp_digests(
     asset_path: str, local_entry: LocalZarrEntry, remote_digest: str
 ) -> tuple[LocalZarrEntry, str, bool]:
+    from dandi.support.digests import md5file_nocache
+
     local_digest = md5file_nocache(local_entry.filepath)
     if local_digest != remote_digest:
         lgr.debug(

dandi/support/digests.py

@@ -9,6 +9,10 @@
 """Provides helper to compute digests (md5 etc) on files
 """
 
+# Importing this module imports fscacher, which imports joblib, which imports
+# numpy, which is a "heavy" import, so avoid importing this module at the top
+# level of a module.
+
 from __future__ import annotations
 
 from collections.abc import Callable
@@ -20,7 +24,6 @@
 
 from dandischema.digests.dandietag import DandiETag
 from fscacher import PersistentCache
-from zarr_checksum import ZarrChecksumTree
 
 from .threaded_walk import threaded_walk
 from ..utils import Hasher, exclude_from_zarr
@@ -101,6 +104,10 @@ def get_zarr_checksum(path: Path, known: dict[str, str] | None = None) -> str:
     passed in the ``known`` argument, which must be a `dict` mapping
     slash-separated paths relative to the root of the Zarr to hex digests.
     """
+    # Importing zarr_checksum leads to importing numpy, which we want to avoid
+    # unless necessary
+    from zarr_checksum import ZarrChecksumTree
+
     if path.is_file():
         s = get_digest(path, "md5")
         assert isinstance(s, str)