Attempt to fix indexing for Dask

This is a naive attempt to make `isel` work with Dask

Known limitation: it triggers the computation.

xarray/core/indexing.py

@@ -5,6 +5,7 @@
 from contextlib import suppress
 from datetime import timedelta
 from typing import Any, Callable, Iterable, List, Optional, Tuple, Union
+import dask.array as da
 
 import numpy as np
 import pandas as pd
@@ -307,7 +308,7 @@ def __init__(self, key):
         for k in key:
             if isinstance(k, slice):
                 k = as_integer_slice(k)
-            elif isinstance(k, np.ndarray):
+            elif isinstance(k, np.ndarray) or isinstance(k, da.Array):
                 if not np.issubdtype(k.dtype, np.integer):
                     raise TypeError(
                         f"invalid indexer array, does not have integer dtype: {k!r}"
@@ -320,7 +321,10 @@ def __init__(self, key):
                         "invalid indexer key: ndarray arguments "
                         f"have different numbers of dimensions: {ndims}"
                     )
-                k = np.asarray(k, dtype=np.int64)
+                if isinstance(k, da.Array):
+                    k = da.asarray(k, dtype=np.int64)
+                else:
+                    k = np.asarray(k, dtype=np.int64)
             else:
                 raise TypeError(
                     f"unexpected indexer type for {type(self).__name__}: {k!r}"
@@ -973,7 +977,6 @@ def _arrayize_vectorized_indexer(indexer, shape):
 
 def _dask_array_with_chunks_hint(array, chunks):
     """Create a dask array using the chunks hint for dimensions of size > 1."""
-    import dask.array as da
 
     if len(chunks) < array.ndim:
         raise ValueError("not enough chunks in hint")