Tkakar/cat-1250 add zip zarr support (#436)

* Added tests

* Version update

* Version update

* Linting

* Used is_zip as param

* Specified is_zip as true

* Fixed is_zip check

Author: tkakar

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "vitessce"
-version = "3.5.10"
+version = "3.6.0"
 authors = [
   { name="Mark Keller", email="[email protected]" },
 ]

src/vitessce/constants.py

@@ -164,6 +164,7 @@ class FileType(DocEnum):
     ANNDATA_ZARR = "anndata.zarr", "Joint file type for AnnData objects"
     SPATIALDATA_ZARR = "spatialdata.zarr", "Joint file type for SpatialData objects"
     ANNDATA_H5AD = "anndata.h5ad", "Joint file type for AnnData objects"
+    ANNDATA_ZARR_ZIP = "anndata.zarr.zip", "Joint file type for AnnData object in a Zarr directory store that has been zipped"
     OBS_EMBEDDING_CSV = 'obsEmbedding.csv', "File type for obsEmbedding values stored in a CSV file"
     OBS_LOCATIONS_CSV = 'obsLocations.csv', "File type for obsLocations values stored in a CSV file"
     OBS_LABELS_CSV = 'obsLabels.csv', "File type for obsLabels values stored in a CSV file"

src/vitessce/wrappers.py

@@ -1192,7 +1192,7 @@ def raise_error_if_more_than_one(inputs):
 
 
 class AnnDataWrapper(AbstractWrapper):
-    def __init__(self, adata_path=None, adata_url=None, adata_store=None, adata_artifact=None, ref_path=None, ref_url=None, ref_artifact=None, obs_feature_matrix_path=None, feature_filter_path=None, initial_feature_filter_path=None, obs_set_paths=None, obs_set_names=None, obs_locations_path=None, obs_segmentations_path=None, obs_embedding_paths=None, obs_embedding_names=None, obs_embedding_dims=None, obs_spots_path=None, obs_points_path=None, feature_labels_path=None, obs_labels_path=None, convert_to_dense=True, coordination_values=None, obs_labels_paths=None, obs_labels_names=None, **kwargs):
+    def __init__(self, adata_path=None, adata_url=None, adata_store=None, adata_artifact=None, ref_path=None, ref_url=None, ref_artifact=None, obs_feature_matrix_path=None, feature_filter_path=None, initial_feature_filter_path=None, obs_set_paths=None, obs_set_names=None, obs_locations_path=None, obs_segmentations_path=None, obs_embedding_paths=None, obs_embedding_names=None, obs_embedding_dims=None, obs_spots_path=None, obs_points_path=None, feature_labels_path=None, obs_labels_path=None, convert_to_dense=True, coordination_values=None, obs_labels_paths=None, obs_labels_names=None, is_zip=None, **kwargs):
         """
         Wrap an AnnData object by creating an instance of the ``AnnDataWrapper`` class.
 
@@ -1220,6 +1220,7 @@ def __init__(self, adata_path=None, adata_url=None, adata_store=None, adata_arti
         :param list[str] obs_labels_names: The optional display names of columns containing observation labels (e.g., alternate cell IDs), instead of the default index in `obs` of the AnnData store.
         :param bool convert_to_dense: Whether or not to convert `X` to dense the zarr store (dense is faster but takes more disk space).
         :param coordination_values: Coordination values for the file definition.
+        :param is_zip: Boolean indicating whether the Zarr store is in a zipped format.
         :type coordination_values: dict or None
         :param \\*\\*kwargs: Keyword arguments inherited from :class:`~vitessce.wrappers.AbstractWrapper`
         """
@@ -1229,6 +1230,7 @@ def __init__(self, adata_path=None, adata_url=None, adata_store=None, adata_arti
         self._adata_url = adata_url
         self._adata_store = adata_store
         self._adata_artifact = adata_artifact
+        self.is_zip = is_zip
 
         # For reference spec JSON with .h5ad files
         self._ref_path = ref_path
@@ -1250,11 +1252,14 @@ def __init__(self, adata_path=None, adata_url=None, adata_store=None, adata_arti
             self.is_remote = False
             self.is_store = False
             self.zarr_folder = 'anndata.zarr'
+            if is_zip is None and '.zip' in str(adata_path):
+                self.is_zip = True
         elif adata_url is not None or adata_artifact is not None:
             self.is_remote = True
             self.is_store = False
             self.zarr_folder = None
-
+            if is_zip is None and '.zip' in str(adata_url):
+                self.is_zip = True
             # Store artifacts on AbstractWrapper.artifacts for downstream access,
             # e.g. in lamindb.save_vitessce_config
             if adata_artifact is not None:
@@ -1356,9 +1361,8 @@ def get_anndata_zarr(base_url):
             if len(options.keys()) > 0:
                 if self.is_h5ad:
                     options["refSpecUrl"] = self.get_ref_url(base_url, dataset_uid, obj_i)
-
                 obj_file_def = {
-                    "fileType": ft.ANNDATA_ZARR.value if not self.is_h5ad else ft.ANNDATA_H5AD.value,
+                    "fileType": ft.ANNDATA_ZARR_ZIP.value if self.is_zip else ft.ANNDATA_H5AD.value if self.is_h5ad else ft.ANNDATA_ZARR.value,
                     "url": self.get_zarr_url(base_url, dataset_uid, obj_i) if not self.is_h5ad else self.get_h5ad_url(base_url, dataset_uid, obj_i),
                     "options": options
                 }

tests/test_config.py

@@ -76,6 +76,52 @@ def test_config_add_dataset():
     }
 
 
+def test_config_add_anndata_zip_url():
+    vc = VitessceConfig(schema_version="1.0.15")
+    vc.add_dataset(name='My Dataset').add_object(
+        AnnDataWrapper(
+            adata_url="http://example.com/adata.zarr.zip",
+            obs_set_paths=["obs/louvain"],
+            is_zip=True
+        )
+    )
+
+    vc_dict = vc.to_dict()
+
+    assert vc_dict == {
+        "version": "1.0.15",
+        "name": "",
+        "description": "",
+        "datasets": [
+            {
+                'uid': 'A',
+                'name': 'My Dataset',
+                'files': [
+                    {
+                        "fileType": "anndata.zarr.zip",
+                        "url": "http://example.com/adata.zarr.zip",
+                        "options": {
+                            "obsSets": [
+                                {
+                                    "name": "louvain",
+                                    "path": "obs/louvain",
+                                }
+                            ]
+                        }
+                    }
+                ]
+            }
+        ],
+        'coordinationSpace': {
+            'dataset': {
+                'A': 'A'
+            },
+        },
+        "layout": [],
+        "initStrategy": "auto"
+    }
+
+
 def test_config_add_anndata_url():
     vc = VitessceConfig(schema_version="1.0.15")
     vc.add_dataset(name='My Dataset').add_object(

tests/test_wrappers.py

@@ -208,6 +208,24 @@ def test_obs_segmentations_ome_zarr(self):
             'url': 'http://localhost:8000/A/0/test.ome.zarr'
         })
 
+    def test_anndata_with_zip(self):
+        adata_path = data_path / 'test.zarr.zip'
+        w = AnnDataWrapper(adata_path,
+                           is_zip=True,
+                           obs_set_paths=['obs/CellType'], obs_set_names=['Cell Type'],
+                           obs_labels_names=['Cell Label'], obs_labels_paths=['obs/CellLabel'],
+                           obs_embedding_paths=['obsm/X_umap'], obs_embedding_names=['UMAP'])
+        w.local_dir_uid = 'anndata.zarr.zip'
+
+        file_def_creator = w.make_file_def_creator('A', 0)
+        file_def = file_def_creator('http://localhost:8000')
+        self.assertEqual(file_def, {'fileType': 'anndata.zarr.zip', 'url': 'http://localhost:8000/A/0/anndata.zarr.zip',
+                                    'options': {
+                                        'obsEmbedding': [{'path': 'obsm/X_umap', 'embeddingType': 'UMAP', 'dims': [0, 1]}],
+                                        'obsSets': [{'path': 'obs/CellType', 'name': 'Cell Type'}],
+                                        'obsLabels': [{'path': 'obs/CellLabel', 'obsLabelsType': 'Cell Label'}]
+                                    }})
+
     def test_anndata(self):
         adata_path = data_path / 'test.h5ad.zarr'
         w = AnnDataWrapper(adata_path,