Load collection-level assets into xarray

docs/source/tutorial.rst

@@ -135,3 +135,50 @@ Intake-stac can turn this object into an Intake catalog:
 
     catalog = intake.open_stac_item_collection('single-file-stac.json')
     list(catalog)
+
+Using xarray-assets
+-------------------
+
+Intake-stac uses the `xarray-assets`_ STAC extension to automatically use the appropriate keywords to load a STAC asset into a data container.
+
+Intake-stac will automatically use the keywords from the `xarray-assets`_ STAC extension, if present, when loading data into a container.
+For example, the STAC collection at <https://planetarycomputer.microsoft.com/api/stac/v1/collections/daymet-annual-hi> defines an
+asset ``zarr-https`` with the metadata ``"xarray:open_kwargs": {"consolidated": true}"`` to indicate that this dataset should be
+opened with the ``consolidated=True`` keyword argument. This will be used automatically by ``.to_dask()``
+
+
+.. code-block:: python
+
+   >>> collection = intake.open_stac_collection(
+   ...     "https://planetarycomputer.microsoft.com/api/stac/v1/collections/daymet-annual-hi"
+   ... )
+
+   >>> source = collection.get_asset("zarr-https")
+   >>> source.to_dask()
+   <xarray.Dataset>
+   Dimensions:                  (nv: 2, time: 41, x: 284, y: 584)
+   Coordinates:
+       lat                      (y, x) float32 dask.array<chunksize=(584, 284), meta=np.ndarray>
+       lon                      (y, x) float32 dask.array<chunksize=(584, 284), meta=np.ndarray>
+     * time                     (time) datetime64[ns] 1980-07-01T12:00:00 ... 20...
+     * x                        (x) float32 -5.802e+06 -5.801e+06 ... -5.519e+06
+     * y                        (y) float32 -3.9e+04 -4e+04 ... -6.21e+05 -6.22e+05
+   Dimensions without coordinates: nv
+   Data variables:
+       lambert_conformal_conic  int16 ...
+       prcp                     (time, y, x) float32 dask.array<chunksize=(1, 584, 284), meta=np.ndarray>
+       swe                      (time, y, x) float32 dask.array<chunksize=(1, 584, 284), meta=np.ndarray>
+       time_bnds                (time, nv) datetime64[ns] dask.array<chunksize=(1, 2), meta=np.ndarray>
+       tmax                     (time, y, x) float32 dask.array<chunksize=(1, 584, 284), meta=np.ndarray>
+       tmin                     (time, y, x) float32 dask.array<chunksize=(1, 584, 284), meta=np.ndarray>
+       vp                       (time, y, x) float32 dask.array<chunksize=(1, 584, 284), meta=np.ndarray>
+   Attributes:
+       Conventions:       CF-1.6
+       Version_data:      Daymet Data Version 4.0
+       Version_software:  Daymet Software Version 4.0
+       citation:          Please see http://daymet.ornl.gov/ for current Daymet ...
+       references:        Please see http://daymet.ornl.gov/ for current informa...
+       source:            Daymet Software Version 4.0
+       start_year:        1980
+
+.. _xarray-assets: https://github.com/stac-extensions/xarray-assets

intake_stac/catalog.py

@@ -4,6 +4,7 @@
 import pystac
 from intake.catalog import Catalog
 from intake.catalog.local import LocalCatalogEntry
+from intake.source import DataSource
 from pkg_resources import get_distribution
 from pystac.extensions.eo import EOExtension
 
@@ -38,6 +39,7 @@
     'application/json': 'textfiles',
     'application/geo+json': 'geopandas',
     'application/geopackage+sqlite3': 'geopandas',
+    'application/vnd+zarr': 'zarr',
     'application/xml': 'textfiles',
 }
 
@@ -165,6 +167,60 @@ class StacCollection(StacCatalog):
     name = 'stac_catalog'
     _stac_cls = pystac.Collection
 
+    def get_asset(
+        self,
+        key,
+        storage_options=None,
+        merge_asset_storage_options=True,
+        merge_asset_open_kwargs=True,
+        **kwargs,
+    ):
+        r"""
+        Get a datasource for a collection-level asset.
+
+        Parameters
+        ----------
+        key : str, optional
+            The asset key to use if multiple Zarr assets are provided.
+        storage_options : dict, optional
+            Additional arguments for the backend fsspec filesystem.
+        merge_asset_storage_option : bool, default True
+            Whether to merge the storage options provided by the asset under the
+            ``xarray:storage_options`` key with `storage_options`.
+        merge_asset_open_kwargs : bool, default True
+            Whether to merge the keywords provided by the asset under the
+            ``xarray:open_kwargs`` key with ``**kwargs``.
+        **kwargs
+            Additional keyword options are provided to the loader, for example ``consolidated=True``
+            to pass to :meth:`xarray.open_zarr`.
+
+        Notes
+        -----
+        The Media Type of the asset will be used to determine how to load the data.
+
+        Returns
+        -------
+        DataSource
+            The dataset described by the asset loaded into a dask-backed object.
+        """
+        try:
+            asset = self._stac_obj.assets[key]
+        except KeyError:
+            raise KeyError(
+                f'No asset named {key}. Should be one of {list(self._stac_obj.assets)}'
+            ) from None
+
+        storage_options = storage_options or {}
+        if merge_asset_storage_options:
+            asset_storage_options = asset.extra_fields.get('xarray:storage_options', {})
+            storage_options.update(asset_storage_options)
+
+        if merge_asset_open_kwargs:
+            asset_open_kwargs = asset.extra_fields.get('xarray:open_kwargs', {})
+            kwargs.update(asset_open_kwargs)
+
+        return StacAsset(key, asset)(storage_options=storage_options, **kwargs)
+
 
 class StacItemCollection(AbstractStacCatalog):
     """
@@ -230,6 +286,20 @@ class StacItem(AbstractStacCatalog):
     name = 'stac_item'
     _stac_cls = pystac.Item
 
+    def __getitem__(self, key):
+        result = super().__getitem__(key)
+        # TODO: handle non-string assets?
+        asset = self._entries[key]
+        storage_options = asset._stac_obj.extra_fields.get('xarray:storage_options', {})
+        open_kwargs = asset._stac_obj.extra_fields.get('xarray:open_kwargs', {})
+
+        if isinstance(result, DataSource):
+            kwargs = result._captured_init_kwargs
+            kwargs = {**kwargs, **dict(storage_options=storage_options), **open_kwargs}
+            result = result(*result._captured_init_args, **kwargs)
+
+        return result
+
     def _load(self):
         """
         Load the STAC Item.
@@ -357,6 +427,7 @@ def __init__(self, key, asset):
         Construct an Intake catalog 'Source' from a STAC Item Asset.
         asset = pystac.item.Asset
         """
+        self._stac_obj = asset
         driver = self._get_driver(asset)
 
         super().__init__(