Merge remote-tracking branch 'upstream/main' into add-pint-registry

* upstream/main:
  Update README.rst
  Add earthcube 2021 notebook link to readme
  v0.5.2
  Some CMIP6 support
  Add pooch to binder environment (#223)
  `add_bounds` uses `keys` rather than `dims` (#221)

Author: dcherian

.binder/environment.yml

@@ -7,5 +7,6 @@ dependencies:
   - netcdf4
   - pip
   - xarray
+  - pooch
   - pip:
     - git+https://github.com/xarray-contrib/cf-xarray

README.rst

@@ -34,4 +34,6 @@ A lightweight convenience wrapper for using CF attributes on xarray objects.
 
 For example you can use ``.cf.mean("latitude")`` instead of ``.mean("lat")`` if appropriate attributes are set! This allows you to write code that does not require knowledge of specific dimension or coordinate names particular to a dataset.
 
-See more in the introductory notebook `here <https://cf-xarray.readthedocs.io/en/latest/examples/introduction.html>`_.
+See more in the `introductory notebook <https://cf-xarray.readthedocs.io/en/latest/examples/introduction.html>`_.
+
+Try out our Earthcube 2021 Annual Meeting notebook `submission <https://binder.pangeo.io/v2/gh/malmans2/cf-xarray-earthcube/main?filepath=DC_01_cf-xarray.ipynb>`_.

cf_xarray/accessor.py

@@ -1586,15 +1586,15 @@ def get_bounds_dim_name(self, key: str) -> str:
         assert self._obj.sizes[bounds_dim] in [2, 4]
         return bounds_dim
 
-    def add_bounds(self, dims: Union[Hashable, Iterable[Hashable]]):
+    def add_bounds(self, keys: Union[str, Iterable[str]]):
         """
         Returns a new object with bounds variables. The bounds values are guessed assuming
         equal spacing on either side of a coordinate label.
 
         Parameters
         ----------
-        dims : Hashable or Iterable[Hashable]
-            Either a single dimension name or a list of dimension names.
+        keys : str or Iterable[str]
+            Either a single key or a list of keys corresponding to dimensions.
 
         Returns
         -------
@@ -1609,12 +1609,16 @@ def add_bounds(self, dims: Union[Hashable, Iterable[Hashable]]):
         The bounds variables are automatically named f"{dim}_bounds" where ``dim``
         is a dimension name.
         """
-        if isinstance(dims, Hashable):
-            dimensions = (dims,)
-        else:
-            dimensions = dims
+        if isinstance(keys, str):
+            keys = [keys]
+
+        dimensions = set()
+        for key in keys:
+            dimensions.update(
+                apply_mapper(_get_dims, self._obj, key, error=False, default=[key])
+            )
 
-        bad_dims: Set[Hashable] = set(dimensions) - set(self._obj.dims)
+        bad_dims: Set[str] = dimensions - set(self._obj.dims)
         if bad_dims:
             raise ValueError(
                 f"{bad_dims!r} are not dimensions in the underlying object."

cf_xarray/criteria.py

@@ -5,6 +5,7 @@
 """
 
 
+import copy
 import re
 from typing import MutableMapping, Tuple
 
@@ -74,19 +75,20 @@
 }
 
 # "long_name" and "standard_name" criteria are the same. For convenience.
-coordinate_criteria["long_name"] = coordinate_criteria["standard_name"]
+coordinate_criteria["long_name"] = copy.deepcopy(coordinate_criteria["standard_name"])
+coordinate_criteria["long_name"]["X"] += ("cell index along first dimension",)
+coordinate_criteria["long_name"]["Y"] += ("cell index along second dimension",)
 
 #: regular expressions for guess_coord_axis
 regex = {
     "time": re.compile("\\bt\\b|(time|min|hour|day|week|month|year)[0-9]*"),
-    "vertical": re.compile(
-        "(z|nav_lev|gdep|lv_|bottom_top|sigma|h(ei)?ght|altitude|depth|"
+    "Z": re.compile(
+        "(z|nav_lev|gdep|lv_|[o]*lev|bottom_top|sigma|h(ei)?ght|altitude|depth|"
         "isobaric|pres|isotherm)[a-z_]*[0-9]*"
     ),
-    "Y": re.compile("y"),
+    "Y": re.compile("y|j|nlat|nj"),
     "latitude": re.compile("y?(nav_lat|lat|gphi)[a-z0-9]*"),
-    "X": re.compile("x"),
+    "X": re.compile("x|i|nlon|ni"),
     "longitude": re.compile("x?(nav_lon|lon|glam)[a-z0-9]*"),
 }
-regex["Z"] = regex["vertical"]
 regex["T"] = regex["time"]

cf_xarray/tests/test_accessor.py

@@ -634,6 +634,10 @@ def test_add_bounds(obj, dims):
         assert added[dim].attrs["bounds"] == name
         assert_allclose(added[name].reset_coords(drop=True), expected[dim])
 
+    # Test multiple dimensions
+    assert not {"x1_bounds", "x2_bounds"} <= set(multiple.variables)
+    assert {"x1_bounds", "x2_bounds"} <= set(multiple.cf.add_bounds("X").variables)
+
 
 def test_bounds():
     ds = airds.copy(deep=True).cf.add_bounds("lat")
@@ -776,9 +780,9 @@ def _make_names(prefixes):
         "nav_lev",
     ]
 )
-_X_NAMES = _make_names(["x"])
-_Y_NAMES = _make_names(["y"])
-_Z_NAMES = _VERTICAL_NAMES
+_X_NAMES = _make_names(["x", "nlon", "i", "ni"])
+_Y_NAMES = _make_names(["y", "nlat", "j", "nj"])
+_Z_NAMES = _VERTICAL_NAMES + ["olevel", "level", "zlevel"]
 _LATITUDE_NAMES = _make_names(["lat", "latitude", "gphi", "nav_lat"])
 _LONGITUDE_NAMES = _make_names(["lon", "longitude", "glam", "nav_lon"])
 
@@ -1189,3 +1193,24 @@ def test_differentiate_positive_upward(obj):
     obj.z.attrs["positive"] = "zzz"
     with pytest.raises(ValueError):
         obj.cf.differentiate("z", positive_upward=True)
+
+
+def test_cmip6_attrs():
+    da = xr.DataArray(
+        np.ones((10, 10)),
+        dims=("nlon", "nlat"),
+        coords={
+            "nlon": (
+                "nlon",
+                np.arange(10),
+                {"long_name": "cell index along first dimension"},
+            ),
+            "nlat": (
+                "nlat",
+                np.arange(10),
+                {"long_name": "cell index along second dimension"},
+            ),
+        },
+    )
+    assert da.cf.axes["X"] == ["nlon"]
+    assert da.cf.axes["Y"] == ["nlat"]

doc/whats-new.rst

@@ -3,11 +3,16 @@
 What's New
 ----------
 
-v0.5.2 (unreleased)
+v0.5.3 (unreleased)
 ===================
-
 - Begin adding support for units with a unit registry for pint arrays. :pr:`197`.
   By `Jon Thielen`_ and `Justus Magin`_.
+
+v0.5.2 (May 11, 2021)
+=====================
+
+- Add some explicit support for CMIP6 output. By `Deepak Cherian`_.
+- Replace the ``dims`` argument of :py:meth:`Dataset.cf.add_bounds` with ``keys``, allowing to use CF keys. By `Mattia Almansi`_.
 - Added :py:attr:`DataArray.cf.formula_terms` and :py:attr:`Dataset.cf.formula_terms`.
   By `Deepak Cherian`_.
 - Added :py:attr:`Dataset.cf.bounds` to return a dictionary mapping valid keys to the variable names of their bounds. By `Mattia Almansi`_.