From d603a3f47cdb857c0fd808a28c79596d7f6d162e Mon Sep 17 00:00:00 2001
From: Jacob Nesbitt <jjnesbitt2@gmail.com>
Date: Wed, 4 Dec 2024 12:18:10 -0500
Subject: [PATCH] Include zarr asset sizes when validating version metadata

Co-authored-by: Aaron Kanzer <aaronkanzer@gmail.com>
---
 dandiapi/api/services/metadata/__init__.py |  5 +-
 dandiapi/api/tests/test_tasks.py           | 57 ++++++++++++++++++++++
 2 files changed, 61 insertions(+), 1 deletion(-)

diff --git a/dandiapi/api/services/metadata/__init__.py b/dandiapi/api/services/metadata/__init__.py
index 0656916a1..b7c7792ba 100644
--- a/dandiapi/api/services/metadata/__init__.py
+++ b/dandiapi/api/services/metadata/__init__.py
@@ -7,6 +7,7 @@
 from dandischema.metadata import aggregate_assets_summary, validate
 from django.conf import settings
 from django.db import transaction
+from django.db.models.query_utils import Q
 from django.utils import timezone
 
 from dandiapi.api.models import Asset, Version
@@ -124,7 +125,9 @@ def _build_validatable_version_metadata(version: Version) -> dict:
         metadata_for_validation['doi'] = '10.80507/dandi.123456/0.123456.1234'
         metadata_for_validation['assetsSummary'] = {
             'schemaKey': 'AssetsSummary',
-            'numberOfBytes': 1 if version.assets.filter(blob__size__gt=0).exists() else 0,
+            'numberOfBytes': 1
+            if version.assets.filter(Q(blob__size__gt=0) | Q(zarr__size__gt=0)).exists()
+            else 0,
             'numberOfFiles': 1 if version.assets.exists() else 0,
         }
         return metadata_for_validation
diff --git a/dandiapi/api/tests/test_tasks.py b/dandiapi/api/tests/test_tasks.py
index 3cb632bb9..85020359b 100644
--- a/dandiapi/api/tests/test_tasks.py
+++ b/dandiapi/api/tests/test_tasks.py
@@ -2,6 +2,7 @@
 
 import datetime
 import hashlib
+from pathlib import Path
 from typing import TYPE_CHECKING
 
 from django.conf import settings
@@ -10,9 +11,13 @@
 from guardian.shortcuts import assign_perm
 import pytest
 from rest_framework.test import APIClient
+from zarr_checksum import compute_zarr_checksum
+from zarr_checksum.checksum import EMPTY_CHECKSUM
+from zarr_checksum.generators import ZarrArchiveFile
 
 from dandiapi.api import tasks
 from dandiapi.api.models import Asset, AssetBlob, Version
+from dandiapi.zarr.models import ZarrArchiveStatus
 
 from .fuzzy import URN_RE, UTC_ISO_TIMESTAMP_RE
 
@@ -321,6 +326,58 @@ def test_validate_version_metadata_no_assets(
     ]
 
 
+@pytest.mark.django_db
+def test_validate_version_metadata_empty_zarr_asset(
+    draft_version: Version, zarr_archive_factory, draft_asset_factory
+):
+    asset = draft_asset_factory(
+        blob=None,
+        zarr=zarr_archive_factory(
+            status=ZarrArchiveStatus.COMPLETE, checksum=EMPTY_CHECKSUM, size=0, file_count=0
+        ),
+        status=Asset.Status.VALID,
+    )
+    assert asset.size == 0
+    draft_version.assets.add(asset)
+
+    # Since the zarr asset has zero size, a validation error should be produced
+    tasks.validate_version_metadata_task(draft_version.id)
+    draft_version.refresh_from_db()
+    assert draft_version.status == Version.Status.INVALID
+    assert draft_version.validation_errors == [
+        {
+            'field': 'assetsSummary',
+            'message': 'Value error, '
+            'A Dandiset containing no files or zero bytes is not publishable',
+        }
+    ]
+
+
+@pytest.mark.django_db
+def test_validate_version_metadata_only_zarr_assets(
+    draft_version: Version, zarr_archive_factory, draft_asset_factory
+):
+    asset = draft_asset_factory(
+        blob=None,
+        zarr=zarr_archive_factory(
+            status=ZarrArchiveStatus.COMPLETE,
+            checksum=compute_zarr_checksum(
+                [ZarrArchiveFile(path=Path('foo/bar'), size=100, digest=hashlib.md5().hexdigest())]
+            ),
+            size=100,
+            file_count=1,
+        ),
+        status=Asset.Status.VALID,
+    )
+    assert asset.size > 0
+    draft_version.assets.add(asset)
+
+    tasks.validate_version_metadata_task(draft_version.id)
+    draft_version.refresh_from_db()
+    assert draft_version.status == Version.Status.VALID
+    assert not draft_version.validation_errors
+
+
 @pytest.mark.django_db
 def test_publish_task(
     api_client: APIClient,