Remove ad-hoc handling of NEP18 libraries in CI (#5740)

* CI versions refresh

* docs

* Update xarray/tests/test_ufuncs.py

Co-authored-by: Mathias Hauser <[email protected]>

* Update xarray/ufuncs.py

Co-authored-by: Mathias Hauser <[email protected]>

Co-authored-by: Mathias Hauser <[email protected]>

Author: crusaderky

.github/workflows/ci-additional.yaml

@@ -42,7 +42,6 @@ jobs:
           [
             "py37-bare-minimum",
             "py37-min-all-deps",
-            "py37-min-nep18",
             "py38-all-but-dask",
             "py38-flaky",
           ]

ci/requirements/py37-min-all-deps.yml

@@ -25,13 +25,14 @@ dependencies:
   - lxml=4.5  # Optional dep of pydap
   - matplotlib-base=3.2
   - nc-time-axis=1.2
-# netcdf follows a 1.major.minor[.patch] convention (see https://github.com/Unidata/netcdf4-python/issues/1090)
-# bumping the netCDF4 version is currently blocked by #4491
+  # netcdf follows a 1.major.minor[.patch] convention
+  # (see https://github.com/Unidata/netcdf4-python/issues/1090)
+  # bumping the netCDF4 version is currently blocked by #4491
   - netcdf4=1.5.3
   - numba=0.49
   - numpy=1.17
   - pandas=1.0
-  # - pint  # See py37-min-nep18.yml
+  - pint=0.15
   - pip
   - pseudonetcdf=3.1
   - pydap=3.2
@@ -44,7 +45,7 @@ dependencies:
   - scipy=1.4
   - seaborn=0.10
   - setuptools=40.4
-  # - sparse  # See py37-min-nep18.yml
+  - sparse=0.8
   - toolz=0.10
   - zarr=2.4
   - pip:

ci/requirements/py37-min-nep18.yml

@@ -95,12 +95,6 @@ dependencies:
 - **setuptools:** 42 months (but no older than 40.4)
 - **numpy:** 18 months
   (`NEP-29 <https://numpy.org/neps/nep-0029-deprecation_policy.html>`_)
-- **dask and dask.distributed:** 12 months
-- **sparse, pint** and other libraries that rely on
-  `NEP-18 <https://numpy.org/neps/nep-0018-array-function-protocol.html>`_
-  for integration: very latest available versions only, until the technology will have
-  matured. This extends to dask when used in conjunction with any of these libraries.
-  numpy >=1.17.
 - **all other libraries:** 12 months
 
 This means the latest minor (X.Y) version from N months prior. Patch versions (x.y.Z)
@@ -109,10 +103,7 @@ release is guaranteed to work.
 
 You can see the actual minimum tested versions:
 
-- `For NEP-18 libraries
-  <https://github.com/pydata/xarray/blob/main/ci/requirements/py37-min-nep18.yml>`_
-- `For everything else
-  <https://github.com/pydata/xarray/blob/main/ci/requirements/py37-min-all-deps.yml>`_
+`<https://github.com/pydata/xarray/blob/main/ci/requirements/py37-min-all-deps.yml>`_
 
 .. _installation-instructions:
 

doc/getting-started-guide/installing.rst

@@ -259,18 +259,16 @@ along a particular dimension, an exception is raised when you try to access
     arrays in a dataset share the same chunking alignment. Neither of these
     are currently done.
 
-NumPy ufuncs like ``np.sin`` currently only work on eagerly evaluated arrays
-(this will change with the next major NumPy release). We have provided
-replacements that also work on all xarray objects, including those that store
-lazy Dask arrays, in the :ref:`xarray.ufuncs <api.ufuncs>` module:
+NumPy ufuncs like ``np.sin`` transparently work on all xarray objects, including those
+that store lazy Dask arrays:
 
 .. ipython:: python
 
-    import xarray.ufuncs as xu
+    import numpy as np
 
-    xu.sin(rechunked)
+    np.sin(rechunked)
 
-To access Dask arrays directly, use the new
+To access Dask arrays directly, use the
 :py:attr:`DataArray.data <xarray.DataArray.data>` attribute. This attribute exposes
 array data either as a Dask array or as a NumPy array, depending on whether it has been
 loaded into Dask or not:
@@ -281,8 +279,8 @@ loaded into Dask or not:
 
 .. note::
 
-    In the future, we may extend ``.data`` to support other "computable" array
-    backends beyond Dask and NumPy (e.g., to support sparse arrays).
+    ``.data`` is also used to expose other "computable" array backends beyond Dask and
+    NumPy (e.g. sparse and pint arrays).
 
 .. _dask.automatic-parallelization:
 

doc/user-guide/dask.rst

@@ -2,6 +2,7 @@
 # it exists to let GitHub build the repository dependency graph
 # https://help.github.com/en/github/visualizing-repository-data-with-graphs/listing-the-packages-that-a-repository-depends-on
 
-numpy >= 1.15
-pandas >= 0.25
+numpy >= 1.17
+pandas >= 1.0
 setuptools >= 40.4
+typing-extensions >= 3.10

requirements.txt

@@ -161,12 +161,11 @@ def test_gufuncs():
 
 
 def test_xarray_ufuncs_deprecation():
-    with pytest.warns(PendingDeprecationWarning, match="xarray.ufuncs"):
+    with pytest.warns(FutureWarning, match="xarray.ufuncs"):
         xu.cos(xr.DataArray([0, 1]))
 
     with pytest.warns(None) as record:
         xu.angle(xr.DataArray([0, 1]))
-    record = [el.message for el in record if el.category == PendingDeprecationWarning]
     assert len(record) == 0
 
 

xarray/tests/test_ufuncs.py

@@ -46,10 +46,8 @@ def __init__(self, name):
     def __call__(self, *args, **kwargs):
         if self._name not in ["angle", "iscomplex"]:
             _warnings.warn(
-                "xarray.ufuncs will be deprecated when xarray no longer "
-                "supports versions of numpy older than v1.17. Instead, use "
-                "numpy ufuncs directly.",
-                PendingDeprecationWarning,
+                "xarray.ufuncs is deprecated. Instead, use numpy ufuncs directly.",
+                FutureWarning,
                 stacklevel=2,
             )