Update minimum dependencies for 0.20 (#5917)

* conda ci requirements

* update setup.cfg for 0.20

* update dependency list in whatsnew

* try removing setuptools

* Fix

* update

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

* fix test

* add notes

* fix pint

* fix accessor

* One more fix

* one last fix?

Co-authored-by: keewis <[email protected]>
Co-authored-by: dcherian <[email protected]>
Co-authored-by: Deepak Cherian <[email protected]>

Author: 3 people

ci/requirements/py37-bare-minimum.yml

@@ -10,6 +10,6 @@ dependencies:
   - pytest-cov
   - pytest-env
   - pytest-xdist
-  - numpy=1.17
-  - pandas=1.0
+  - numpy=1.18
+  - pandas=1.1
   - typing_extensions=3.7

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

@@ -10,29 +10,31 @@ dependencies:
   - python=3.7
   - boto3=1.13
   - bottleneck=1.3
+  # cartopy 0.18 conflicts with pynio
   - cartopy=0.17
   - cdms2=3.1
   - cfgrib=0.9
-  - cftime=1.1
+  - cftime=1.2
   - coveralls
-  - dask=2.24
-  - distributed=2.24
+  - dask=2.30
+  - distributed=2.30
   - h5netcdf=0.8
   - h5py=2.10
+  # hdf5 1.12 conflicts with h5py=2.10
   - hdf5=1.10
   - hypothesis
   - iris=2.4
-  - lxml=4.5  # Optional dep of pydap
-  - matplotlib-base=3.2
+  - lxml=4.6  # Optional dep of pydap
+  - matplotlib-base=3.3
   - 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
   - netcdf4=1.5.3
-  - numba=0.49
-  - numpy=1.17
-  - pandas=1.0
-  - pint=0.15
+  - numba=0.51
+  - numpy=1.18
+  - pandas=1.1
+  - pint=0.16
   - pip
   - pseudonetcdf=3.1
   - pydap=3.2
@@ -42,13 +44,11 @@ dependencies:
   - pytest-env
   - pytest-xdist
   - rasterio=1.1
-  - scipy=1.4
-  - seaborn=0.10
-  # don't need to pin setuptools, now that we don't depend on it
-  - setuptools
-  - sparse=0.8
-  - toolz=0.10
+  - scipy=1.5
+  - seaborn=0.11
+  - sparse=0.11
+  - toolz=0.11
   - typing_extensions=3.7
-  - zarr=2.4
+  - zarr=2.5
   - pip:
     - numbagg==0.1

doc/whats-new.rst

@@ -49,12 +49,24 @@ Breaking changes
 ~~~~~~~~~~~~~~~~
 - The minimum versions of some dependencies were changed:
 
-  ============ ====== ====
-  Package      Old    New
-  ============ ====== ====
-  dask         2.15   2.24
-  distributed  2.15   2.24
-  ============ ====== ====
+  =============== ====== ====
+  Package         Old    New
+  =============== ====== ====
+  cftime          1.1    1.2
+  dask            2.15   2.30
+  distributed     2.15   2.30
+  lxml            4.5    4.6
+  matplotlib-base 3.2    3.3
+  numba           0.49   0.51
+  numpy           1.17   1.18
+  pandas          1.0    1.1
+  pint            0.15   0.16
+  scipy           1.4    1.5
+  seaborn         0.10   0.11
+  sparse          0.8    0.11
+  toolz           0.10   0.11
+  zarr            2.4    2.5
+  =============== ====== ====
 
 - The ``__repr__`` of a :py:class:`xarray.Dataset`'s ``coords`` and ``data_vars``
   ignore ``xarray.set_option(display_max_rows=...)`` and show the full output

setup.cfg

@@ -76,8 +76,8 @@ zip_safe = False  # https://mypy.readthedocs.io/en/latest/installed_packages.htm
 include_package_data = True
 python_requires = >=3.7
 install_requires =
-    numpy >= 1.17
-    pandas >= 1.0
+    numpy >= 1.18
+    pandas >= 1.1
     importlib-metadata; python_version < '3.8'
     typing_extensions >= 3.7; python_version < '3.8'
 

xarray/backends/zarr.py

@@ -1,6 +1,5 @@
 import os
 import warnings
-from distutils.version import LooseVersion
 
 import numpy as np
 
@@ -353,10 +352,7 @@ def open_group(
             synchronizer=synchronizer,
             path=group,
         )
-        if LooseVersion(zarr.__version__) >= "2.5.0":
-            open_kwargs["storage_options"] = storage_options
-        elif storage_options:
-            raise ValueError("Storage options only compatible with zarr>=2.5.0")
+        open_kwargs["storage_options"] = storage_options
 
         if chunk_store:
             open_kwargs["chunk_store"] = chunk_store

xarray/coding/cftime_offsets.py

@@ -42,7 +42,6 @@
 
 import re
 from datetime import timedelta
-from distutils.version import LooseVersion
 from functools import partial
 from typing import ClassVar, Optional
 
@@ -243,14 +242,7 @@ def _shift_month(date, months, day_option="start"):
         day = _days_in_month(reference)
     else:
         raise ValueError(day_option)
-    if LooseVersion(cftime.__version__) < LooseVersion("1.0.4"):
-        # dayofwk=-1 is required to update the dayofwk and dayofyr attributes of
-        # the returned date object in versions of cftime between 1.0.2 and
-        # 1.0.3.4.  It can be removed for versions of cftime greater than
-        # 1.0.3.4.
-        return date.replace(year=year, month=month, day=day, dayofwk=-1)
-    else:
-        return date.replace(year=year, month=month, day=day)
+    return date.replace(year=year, month=month, day=day)
 
 
 def roll_qtrday(other, n, month, day_option, modby=3):

xarray/coding/cftimeindex.py

@@ -134,12 +134,6 @@ def _parse_iso8601_with_reso(date_type, timestr):
             # TODO: Consider adding support for sub-second resolution?
             replace[attr] = int(value)
             resolution = attr
-    if LooseVersion(cftime.__version__) < LooseVersion("1.0.4"):
-        # dayofwk=-1 is required to update the dayofwk and dayofyr attributes of
-        # the returned date object in versions of cftime between 1.0.2 and
-        # 1.0.3.4.  It can be removed for versions of cftime greater than
-        # 1.0.3.4.
-        replace["dayofwk"] = -1
     return default.replace(**replace), resolution
 
 

xarray/coding/times.py

@@ -1,7 +1,6 @@
 import re
 import warnings
 from datetime import datetime, timedelta
-from distutils.version import LooseVersion
 from functools import partial
 
 import numpy as np
@@ -269,19 +268,13 @@ def decode_cf_datetime(num_dates, units, calendar=None, use_cftime=None):
 
 
 def to_timedelta_unboxed(value, **kwargs):
-    if LooseVersion(pd.__version__) < "0.25.0":
-        result = pd.to_timedelta(value, **kwargs, box=False)
-    else:
-        result = pd.to_timedelta(value, **kwargs).to_numpy()
+    result = pd.to_timedelta(value, **kwargs).to_numpy()
     assert result.dtype == "timedelta64[ns]"
     return result
 
 
 def to_datetime_unboxed(value, **kwargs):
-    if LooseVersion(pd.__version__) < "0.25.0":
-        result = pd.to_datetime(value, **kwargs, box=False)
-    else:
-        result = pd.to_datetime(value, **kwargs).to_numpy()
+    result = pd.to_datetime(value, **kwargs).to_numpy()
     assert result.dtype == "datetime64[ns]"
     return result
 

xarray/core/accessor_dt.py

@@ -1,5 +1,4 @@
 import warnings
-from distutils.version import LooseVersion
 
 import numpy as np
 import pandas as pd
@@ -336,9 +335,6 @@ def isocalendar(self):
         if not is_np_datetime_like(self._obj.data.dtype):
             raise AttributeError("'CFTimeIndex' object has no attribute 'isocalendar'")
 
-        if LooseVersion(pd.__version__) < "1.1.0":
-            raise AttributeError("'isocalendar' not available in pandas < 1.1.0")
-
         values = _get_date_field(self._obj.data, "isocalendar", np.int64)
 
         obj_type = type(self._obj)
@@ -383,12 +379,7 @@ def weekofyear(self):
             FutureWarning,
         )
 
-        if LooseVersion(pd.__version__) < "1.1.0":
-            weekofyear = Properties._tslib_field_accessor(
-                "weekofyear", "The week ordinal of the year", np.int64
-            ).fget(self)
-        else:
-            weekofyear = self.isocalendar().week
+        weekofyear = self.isocalendar().week
 
         return weekofyear
 

xarray/core/missing.py

@@ -1,6 +1,5 @@
 import datetime as dt
 import warnings
-from distutils.version import LooseVersion
 from functools import partial
 from numbers import Number
 from typing import Any, Callable, Dict, Hashable, Sequence, Union
@@ -557,16 +556,8 @@ def _localize(var, indexes_coords):
     """
     indexes = {}
     for dim, [x, new_x] in indexes_coords.items():
-        if np.issubdtype(new_x.dtype, np.datetime64) and LooseVersion(
-            np.__version__
-        ) < LooseVersion("1.18"):
-            # np.nanmin/max changed behaviour for datetime types in numpy 1.18,
-            # see https://github.com/pydata/xarray/pull/3924/files
-            minval = np.min(new_x.values)
-            maxval = np.max(new_x.values)
-        else:
-            minval = np.nanmin(new_x.values)
-            maxval = np.nanmax(new_x.values)
+        minval = np.nanmin(new_x.values)
+        maxval = np.nanmax(new_x.values)
         index = x.to_index()
         imin = index.get_loc(minval, method="nearest")
         imax = index.get_loc(maxval, method="nearest")

xarray/tests/__init__.py

@@ -61,9 +61,6 @@ def LooseVersion(vstring):
 
 
 has_matplotlib, requires_matplotlib = _importorskip("matplotlib")
-has_matplotlib_3_3_0, requires_matplotlib_3_3_0 = _importorskip(
-    "matplotlib", minversion="3.3.0"
-)
 has_scipy, requires_scipy = _importorskip("scipy")
 has_pydap, requires_pydap = _importorskip("pydap.client")
 has_netCDF4, requires_netCDF4 = _importorskip("netCDF4")
@@ -77,7 +74,6 @@ def LooseVersion(vstring):
 has_nc_time_axis, requires_nc_time_axis = _importorskip("nc_time_axis")
 has_rasterio, requires_rasterio = _importorskip("rasterio")
 has_zarr, requires_zarr = _importorskip("zarr")
-has_zarr_2_5_0, requires_zarr_2_5_0 = _importorskip("zarr", minversion="2.5.0")
 has_fsspec, requires_fsspec = _importorskip("fsspec")
 has_iris, requires_iris = _importorskip("iris")
 has_cfgrib, requires_cfgrib = _importorskip("cfgrib")
@@ -86,8 +82,7 @@ def LooseVersion(vstring):
 has_sparse, requires_sparse = _importorskip("sparse")
 has_cupy, requires_cupy = _importorskip("cupy")
 has_cartopy, requires_cartopy = _importorskip("cartopy")
-# Need Pint 0.15 for __dask_tokenize__ tests for Quantity wrapped Dask Arrays
-has_pint_0_15, requires_pint_0_15 = _importorskip("pint", minversion="0.15")
+has_pint, requires_pint = _importorskip("pint")
 has_numexpr, requires_numexpr = _importorskip("numexpr")
 
 # some special cases

xarray/tests/test_accessor_dt.py

@@ -71,7 +71,7 @@ def setup(self):
     )
     def test_field_access(self, field) -> None:
 
-        if LooseVersion(pd.__version__) >= "1.1.0" and field in ["week", "weekofyear"]:
+        if field in ["week", "weekofyear"]:
             data = self.times.isocalendar()["week"]
         else:
             data = getattr(self.times, field)
@@ -98,13 +98,6 @@ def test_field_access(self, field) -> None:
     )
     def test_isocalendar(self, field, pandas_field) -> None:
 
-        if LooseVersion(pd.__version__) < "1.1.0":
-            with pytest.raises(
-                AttributeError, match=r"'isocalendar' not available in pandas < 1.1.0"
-            ):
-                self.data.time.dt.isocalendar()[field]
-            return
-
         # pandas isocalendar has dtypy UInt32Dtype, convert to Int64
         expected = pd.Int64Index(getattr(self.times.isocalendar(), pandas_field))
         expected = xr.DataArray(
@@ -185,13 +178,6 @@ def test_dask_field_access(self, field) -> None:
     def test_isocalendar_dask(self, field) -> None:
         import dask.array as da
 
-        if LooseVersion(pd.__version__) < "1.1.0":
-            with pytest.raises(
-                AttributeError, match=r"'isocalendar' not available in pandas < 1.1.0"
-            ):
-                self.data.time.dt.isocalendar()[field]
-            return
-
         expected = getattr(self.times_data.dt.isocalendar(), field)
 
         dask_times_arr = da.from_array(self.times_arr, chunks=(5, 5, 50))

xarray/tests/test_backends.py

@@ -71,7 +71,6 @@
     requires_scipy,
     requires_scipy_or_netCDF4,
     requires_zarr,
-    requires_zarr_2_5_0,
 )
 from .test_coding_times import (
     _ALL_CALENDARS,
@@ -2400,7 +2399,6 @@ def create_zarr_target(self):
 
 
 @requires_fsspec
-@requires_zarr_2_5_0
 def test_zarr_storage_options():
     pytest.importorskip("aiobotocore")
     ds = create_test_data()

xarray/tests/test_dask.py

@@ -24,7 +24,7 @@
     assert_frame_equal,
     assert_identical,
     raise_if_dask_computes,
-    requires_pint_0_15,
+    requires_pint,
     requires_scipy_or_netCDF4,
 )
 from .test_backends import create_tmp_file
@@ -297,7 +297,7 @@ def test_persist(self):
         self.assertLazyAndAllClose(u + 1, v)
         self.assertLazyAndAllClose(u + 1, v2)
 
-    @requires_pint_0_15(reason="Need __dask_tokenize__")
+    @requires_pint
     def test_tokenize_duck_dask_array(self):
         import pint
 
@@ -748,7 +748,7 @@ def test_from_dask_variable(self):
         a = DataArray(self.lazy_array.variable, coords={"x": range(4)}, name="foo")
         self.assertLazyAndIdentical(self.lazy_array, a)
 
-    @requires_pint_0_15(reason="Need __dask_tokenize__")
+    @requires_pint
     def test_tokenize_duck_dask_array(self):
         import pint