ENH: adds "UnitsDataArray.open" and "UnitsDataArray.create" class methods

podpac/__init__.py

@@ -44,7 +44,7 @@ def makedirs(name, mode=511, exist_ok=False):
 from podpac.core.node import Node, NodeException
 import podpac.core.authentication as authentication
 from podpac.core.utils import NodeTrait
-from podpac.core.units import ureg as units
+from podpac.core.units import ureg as units, UnitsDataArray
 
 # Organized submodules
 # These files are simply wrappers to create a curated namespace of podpac modules

podpac/core/managers/aws.py

@@ -1604,7 +1604,7 @@ def create_function(
     elif function_source_dist_zip is not None:
         with open(function_source_dist_zip, "rb") as f:
             lambda_config["Code"] = {}  # reset the code dict to make sure S3Bucket and S3Key are overridden
-            lambda_config["Code"]["ZipFile"]: f.read()
+            lambda_config["Code"]["ZipFile"] = f.read()
 
     else:
         raise ValueError("Function source is not defined")

podpac/core/node.py

@@ -18,7 +18,7 @@
 import traitlets as tl
 
 from podpac.core.settings import settings
-from podpac.core.units import ureg, UnitsDataArray, create_data_array
+from podpac.core.units import ureg, UnitsDataArray
 from podpac.core.utils import common_doc
 from podpac.core.utils import JSONEncoder, is_json_serializable
 from podpac.core.utils import _get_query_params_from_url, _get_from_url, _get_param
@@ -252,7 +252,7 @@ def create_output_array(self, coords, data=np.nan, **kwargs):
         if self.units is not None:
             attrs["units"] = ureg.Unit(self.units)
 
-        return create_data_array(coords, data=data, dtype=self.dtype, attrs=attrs, **kwargs)
+        return UnitsDataArray.create(coords, data=data, dtype=self.dtype, attrs=attrs, **kwargs)
 
     # -----------------------------------------------------------------------------------------------------------------
     # Serialization

podpac/core/test/test_units.py

@@ -12,8 +12,10 @@
 
 from podpac.core.units import ureg
 from podpac.core.units import UnitsDataArray
-from podpac.core.units import create_data_array
 from podpac.core.units import get_image
+from podpac.core.units import create_dataarray  # DEPRECATED
+
+from podpac.data import Array
 
 
 class TestUnitDataArray(object):
@@ -331,77 +333,129 @@ def setup_class(cls):
         cls.coords = Coordinates([[0, 1, 2], [0, 1, 2, 3]], dims=["lat", "lon"])
 
     def test_default(self):
-        a = create_data_array(self.coords)
+        a = UnitsDataArray.create(self.coords)
         assert isinstance(a, UnitsDataArray)
         assert a.shape == self.coords.shape
         assert np.all(np.isnan(a))
 
     def test_empty(self):
-        a = create_data_array(self.coords, data=None)
+        a = UnitsDataArray.create(self.coords, data=None)
         assert isinstance(a, UnitsDataArray)
         assert a.shape == self.coords.shape
         assert a.dtype == float
 
-        a = create_data_array(self.coords, data=None, dtype=bool)
+        a = UnitsDataArray.create(self.coords, data=None, dtype=bool)
         assert isinstance(a, UnitsDataArray)
         assert a.shape == self.coords.shape
         assert a.dtype == bool
 
     def test_zeros(self):
-        a = create_data_array(self.coords, data=0)
+        a = UnitsDataArray.create(self.coords, data=0)
         assert isinstance(a, UnitsDataArray)
         assert a.shape == self.coords.shape
         assert a.dtype == float
         assert np.all(a == 0.0)
 
-        a = create_data_array(self.coords, data=0, dtype=bool)
+        a = UnitsDataArray.create(self.coords, data=0, dtype=bool)
         assert isinstance(a, UnitsDataArray)
         assert a.shape == self.coords.shape
         assert a.dtype == bool
         assert np.all(~a)
 
     def test_ones(self):
-        a = create_data_array(self.coords, data=1)
+        a = UnitsDataArray.create(self.coords, data=1)
         assert isinstance(a, UnitsDataArray)
         assert a.shape == self.coords.shape
         assert a.dtype == float
         assert np.all(a == 1.0)
 
-        a = create_data_array(self.coords, data=1, dtype=bool)
+        a = UnitsDataArray.create(self.coords, data=1, dtype=bool)
         assert isinstance(a, UnitsDataArray)
         assert a.shape == self.coords.shape
         assert a.dtype == bool
         assert np.all(a)
 
     def test_full(self):
-        a = create_data_array(self.coords, data=10)
+        a = UnitsDataArray.create(self.coords, data=10)
         assert isinstance(a, UnitsDataArray)
         assert a.shape == self.coords.shape
         assert a.dtype == float
         assert np.all(a == 10)
 
-        a = create_data_array(self.coords, data=10, dtype=int)
+        a = UnitsDataArray.create(self.coords, data=10, dtype=int)
         assert isinstance(a, UnitsDataArray)
         assert a.shape == self.coords.shape
         assert a.dtype == int
         assert np.all(a == 10)
 
     def test_array(self):
         data = np.random.random(self.coords.shape)
-        a = create_data_array(self.coords, data=data)
+        a = UnitsDataArray.create(self.coords, data=data)
         assert isinstance(a, UnitsDataArray)
         assert a.dtype == float
         np.testing.assert_equal(a.data, data)
 
         data = np.round(10 * np.random.random(self.coords.shape))
-        a = create_data_array(self.coords, data=data, dtype=int)
+        a = UnitsDataArray.create(self.coords, data=data, dtype=int)
         assert isinstance(a, UnitsDataArray)
         assert a.dtype == int
         np.testing.assert_equal(a.data, data.astype(int))
 
     def test_invalid_coords(self):
         with pytest.raises(TypeError):
-            create_data_array((3, 4))
+            UnitsDataArray.create((3, 4))
+
+    def test_deprecate_create_dataarray(self):
+        with pytest.deprecated_call():
+            create_dataarray(self.coords, data=10)
+
+
+class TestOpenDataArray(object):
+    def test_open_after_create(self):
+        coords = Coordinates([[0, 1, 2], [0, 1, 2, 3]], dims=["lat", "lon"])
+        uda_1 = UnitsDataArray.create(coords, data=np.random.rand(3, 4))
+        ncdf = uda_1.to_netcdf()
+        uda_2 = UnitsDataArray.open(ncdf)
+
+        assert isinstance(uda_2, UnitsDataArray)
+        assert np.all(uda_2.data == uda_1.data)
+
+    def test_open_after_create_with_attrs(self):
+        coords = Coordinates([[0, 1, 2], [0, 1, 2, 3]], dims=["lat", "lon"], crs="EPSG:4193")
+        uda_1 = UnitsDataArray.create(coords, data=np.random.rand(3, 4), attrs={"some_attr": 5})
+        ncdf = uda_1.to_netcdf()
+        uda_2 = UnitsDataArray.open(ncdf)
+
+        assert isinstance(uda_2, UnitsDataArray)
+        assert np.all(uda_2.data == uda_1.data)
+
+        assert "some_attr" in uda_2.attrs
+        assert uda_2.attrs.get("some_attr") == uda_1.attrs.get("some_attr")
+
+        assert "crs" in uda_2.attrs
+        assert uda_2.attrs.get("crs") == uda_1.attrs.get("crs")
+
+    def test_open_after_eval(self):
+
+        # mock node
+        data = np.random.rand(5, 5)
+        lat = np.linspace(-10, 10, 5)
+        lon = np.linspace(-10, 10, 5)
+        native_coords = Coordinates([lat, lon], ["lat", "lon"])
+        node = Array(source=data, native_coordinates=native_coords)
+        uda = node.eval(node.native_coordinates)
+
+        ncdf = uda.to_netcdf()
+        uda_2 = UnitsDataArray.open(ncdf)
+
+        assert isinstance(uda_2, UnitsDataArray)
+        assert np.all(uda_2.data == uda.data)
+
+        assert "layer_style" in uda_2.attrs
+        assert uda_2.attrs.get("layer_style").json == uda.attrs.get("layer_style").json
+
+        assert "crs" in uda_2.attrs
+        assert uda_2.attrs.get("crs") == uda.attrs.get("crs")
 
 
 class TestGetImage(object):

podpac/core/units.py

@@ -11,6 +11,7 @@
 import operator
 from six import string_types
 import json
+import warnings
 
 from io import BytesIO
 import base64
@@ -29,6 +30,7 @@
 ureg = UnitRegistry()
 
 import podpac
+from podpac import Coordinates
 from podpac.core.settings import settings
 from podpac.core.utils import JSONEncoder
 from podpac.core.style import Style
@@ -283,6 +285,71 @@ def set(self, value, mask):
             # set self to have the same dims (and same order) as when first started
             self = self.transpose(*orig_dims)
 
+    @classmethod
+    def open(cls, *args, **kwargs):
+        """
+        Open an :class:`podpac.UnitsDataArray` from a file or file-like object containing a single data variable.
+
+        This is a wrapper around :func:`xarray.open_datarray`.
+        The inputs to this function are passed directly to :func:`xarray.open_datarray`.
+        See http://xarray.pydata.org/en/stable/generated/xarray.open_dataarray.html#xarray.open_dataarray.
+
+        The DataArray passed back from :func:`xarray.open_datarray` is used to create a units data array using :func:`creare_dataarray`.
+
+        Returns
+        -------
+        :class:`podpac.UnitsDataArray`
+        """
+        da = xr.open_dataarray(*args, **kwargs)
+        coords = Coordinates.from_xarray(da.coords, crs=da.attrs.get("crs"))
+
+        # pass in kwargs to constructor
+        uda_kwargs = {"attrs": da.attrs}
+        return cls.create(coords, data=da.data, **uda_kwargs)
+
+    @classmethod
+    def create(cls, coords, data=np.nan, dtype=float, **kwargs):
+        """Shortcut to create :class:`podpac.UnitsDataArray`
+
+        Parameters
+        ----------
+        coords : :class:`podpac.Coordinates`
+            PODPAC Coordinates
+        data : np.ndarray, optional
+            Data to fill in. Defaults to np.nan.
+        dtype : type, optional
+            Data type. Defaults to float.
+        **kwargs
+            keyword arguments to pass to :class:`podpac.UnitsDataArray` constructor
+
+        Returns
+        -------
+        :class:`podpac.UnitsDataArray`
+        """
+        if not isinstance(coords, podpac.Coordinates):
+            raise TypeError("`UnitsDataArray.create` expected Coordinates object, not '%s'" % type(coords))
+
+        if data is None:
+            data = np.empty(coords.shape, dtype=dtype)
+        elif np.shape(data) == ():
+            if data == 0:
+                data = np.zeros(coords.shape, dtype=dtype)
+            elif data == 1:
+                data = np.ones(coords.shape, dtype=dtype)
+            else:
+                data = np.full(coords.shape, data, dtype=dtype)
+        else:
+            data = data.astype(dtype)
+
+        # add crs attr
+        if "attrs" in kwargs:
+            if "crs" not in kwargs["attrs"]:
+                kwargs["attrs"]["crs"] = coords.crs
+        else:
+            kwargs["attrs"] = {"crs": coords.crs}
+
+        return cls(data, coords=coords.coords, dims=coords.idims, **kwargs)
+
 
 for tp in ("mul", "matmul", "truediv", "div"):
     meth = "__{:s}__".format(tp)
@@ -346,35 +413,14 @@ def func(self, *args, **kwargs):
 del func
 
 
-# ---------------------------------------------------------------------------------------------------------------------
-# Utility functions
-# ---------------------------------------------------------------------------------------------------------------------
-
-
-def create_data_array(coords, data=np.nan, dtype=float, **kwargs):
-    if not isinstance(coords, podpac.Coordinates):
-        raise TypeError("create_data_array expected Coordinates object, not '%s'" % type(coords))
-
-    if data is None:
-        data = np.empty(coords.shape, dtype=dtype)
-    elif np.shape(data) == ():
-        if data == 0:
-            data = np.zeros(coords.shape, dtype=dtype)
-        elif data == 1:
-            data = np.ones(coords.shape, dtype=dtype)
-        else:
-            data = np.full(coords.shape, data, dtype=dtype)
-    else:
-        data = data.astype(dtype)
-
-    # add crs attr
-    if "attrs" in kwargs:
-        if "crs" not in kwargs["attrs"]:
-            kwargs["attrs"]["crs"] = coords.crs
-    else:
-        kwargs["attrs"] = {"crs": coords.crs}
-
-    return UnitsDataArray(data, coords=coords.coords, dims=coords.idims, **kwargs)
+def create_dataarray(coords, data=np.nan, dtype=float, **kwargs):
+    """Deprecated. Use `UnitsDataArray.create()` in place.
+    """
+    warnings.warn(
+        "The `create_dataarray` function is deprecated and will be removed in podpac 2.0. Use the classmethod `UnitsDataArray.create()` instead.",
+        DeprecationWarning,
+    )
+    return UnitsDataArray.create(coords, data, dtype, **kwargs)
 
 
 def get_image(data, format="png", vmin=None, vmax=None, return_base64=False):

podpac/datalib/egi.py

@@ -29,7 +29,7 @@
 from podpac.data import DataSource
 from podpac import authentication
 from podpac import settings
-from podpac.core.units import UnitsDataArray, create_data_array
+from podpac.core.units import UnitsDataArray
 from podpac.core.node import node_eval
 
 # Set up logging
@@ -225,7 +225,7 @@ def eval(self, coordinates, output=None):
             )
             raise e
         # Force update on native_coordinates (in case of multiple evals)
-        self.set_trait('native_coordinates', self.get_native_coordinates())
+        self.set_trait("native_coordinates", self.get_native_coordinates())
 
         # run normal eval once self.data is prepared
         return super(EGI, self).eval(coordinates, output)

podpac/datalib/smap_egi.py

@@ -37,7 +37,7 @@ def isnat(a):
 import podpac.datalib
 from podpac.core.coordinates import Coordinates
 from podpac.datalib import EGI
-from podpac.core.units import create_data_array
+from podpac.core.units import UnitsDataArray
 
 SMAP_PRODUCT_DICT = {
     #'shortname':    ['lat_key', 'lon_key', 'data_key', 'quality_flag', 'default_verison']
@@ -143,7 +143,6 @@ def coverage(self):
         else:
             return (self.data_key, self.lat_key, self.lon_key)
 
-
     @tl.default("version")
     def _version_default(self):
         return SMAP_PRODUCT_DICT[self.product][4]
@@ -219,7 +218,7 @@ def read_file(self, filelike):
             c = Coordinates([time, lon, lat], dims=["time", "lon", "lat"], crs="epsg:6933")
 
         # make units data array with coordinates and data
-        return create_data_array(c, data=data)
+        return UnitsDataArray.create(c, data=data)
 
     def append_file(self, all_data, data):
         """Append data