Merge remote-tracking branch 'upstream/master' into propagate-attrs

* upstream/master: (34 commits)
  Fix bug in computing means of cftime.datetime arrays (pydata#4344)
  fix some str accessor inconsistencies (pydata#4339)
  pin matplotlib in ci/requirements/doc.yml (pydata#4340)
  Clarify drop_vars return value. (pydata#4244)
  Support explicitly setting a dimension order with to_dataframe() (pydata#4333)
  Increase support window of all dependencies (pydata#4296)
  Implement interp for interpolating between chunks of data (dask) (pydata#4155)
  Add @mathause to current core developers. (pydata#4335)
  install sphinx-autosummary-accessors from conda-forge (pydata#4332)
  Use sphinx-accessors-autosummary (pydata#4323)
  ndrolling fixes (pydata#4329)
  DOC: fix typo argmin -> argmax in DataArray.argmax docstring (pydata#4327)
  pin sphinx to 3.1(pydata#4326)
  nd-rolling (pydata#4219)
  Implicit dask import 4164 (pydata#4318)
  allow customizing the inline repr of a duck array (pydata#4248)
  silence the known docs CI issues (pydata#4316)
  enh: fixed pydata#4302 (pydata#4315)
  Remove all unused and warn-raising methods from AbstractDataStore (pydata#4310)
  Fix map_blocks example (pydata#4305)
  ...

HOW_TO_RELEASE.md

@@ -1,4 +1,4 @@
-# How to issue an xarray release in 17 easy steps
+# How to issue an xarray release in 20 easy steps
 
 Time required: about an hour.
 
@@ -23,9 +23,9 @@ upstream        https://github.com/pydata/xarray (push)
     git log v{0.X.Y-1}.. --format=%aN | sort -u | perl -pe 's/\n/$1, /'
     ```
     Add these into `whats-new.rst` somewhere :)
- 2. Write a release summary: ~50 words describing the high level features. This
+ 3. Write a release summary: ~50 words describing the high level features. This
     will be used in the release emails, tweets, GitHub release notes, etc. 
- 3. Look over whats-new.rst and the docs. Make sure "What's New" is complete
+ 4. Look over whats-new.rst and the docs. Make sure "What's New" is complete
     (check the date!) and add the release summary at the top.
     Things to watch out for:
     - Important new features should be highlighted towards the top.
@@ -34,46 +34,46 @@ upstream        https://github.com/pydata/xarray (push)
       due to a bad merge. Check for these before a release by using git diff,
       e.g., `git diff v{0.X.Y-1} whats-new.rst` where {0.X.Y-1} is the previous
       release.
- 4. If possible, open a PR with the release summary and whatsnew changes.
- 4. After merging, again ensure your master branch is synced to upstream:
+ 5. If possible, open a PR with the release summary and whatsnew changes.
+ 6. After merging, again ensure your master branch is synced to upstream:
      ```sh
      git pull upstream master
      ```
- 4. If you have any doubts, run the full test suite one final time!
+ 7. If you have any doubts, run the full test suite one final time!
       ```sh
       pytest
       ```
- 5. Check that the ReadTheDocs build is passing.
- 6. On the master branch, commit the release in git:
+ 8. Check that the ReadTheDocs build is passing.
+ 9. On the master branch, commit the release in git:
       ```s
       git commit -am 'Release v{0.X.Y}'
       ```
- 7. Tag the release:
+10. Tag the release:
       ```sh
       git tag -a v{0.X.Y} -m 'v{0.X.Y}'
       ```
- 8. Build source and binary wheels for PyPI:
+11. Build source and binary wheels for PyPI:
       ```sh
       git clean -xdf  # this deletes all uncommitted changes!
       python setup.py bdist_wheel sdist
       ```
- 9. Use twine to check the package build:
+12. Use twine to check the package build:
       ```sh
       twine check dist/xarray-{0.X.Y}*
       ```
-10. Use twine to register and upload the release on PyPI. Be careful, you can't
+13. Use twine to register and upload the release on PyPI. Be careful, you can't
     take this back!
       ```sh
       twine upload dist/xarray-{0.X.Y}*
       ```
     You will need to be listed as a package owner at
     https://pypi.python.org/pypi/xarray for this to work.
-11. Push your changes to master:
+14. Push your changes to master:
       ```sh
       git push upstream master
       git push upstream --tags
       ```
-12. Update the stable branch (used by ReadTheDocs) and switch back to master:
+15. Update the stable branch (used by ReadTheDocs) and switch back to master:
      ```sh
       git checkout stable
       git rebase master
@@ -83,7 +83,7 @@ upstream        https://github.com/pydata/xarray (push)
     It's OK to force push to 'stable' if necessary. (We also update the stable 
     branch with `git cherry-pick` for documentation only fixes that apply the 
     current released version.)
-13. Add a section for the next release {0.X.Y+1} to doc/whats-new.rst:
+16. Add a section for the next release {0.X.Y+1} to doc/whats-new.rst:
      ```
      .. _whats-new.{0.X.Y+1}:
 
@@ -109,19 +109,19 @@ upstream        https://github.com/pydata/xarray (push)
      Internal Changes
      ~~~~~~~~~~~~~~~~
      ```
-14. Commit your changes and push to master again:
+17. Commit your changes and push to master again:
       ```sh
       git commit -am 'New whatsnew section'
       git push upstream master
       ```
     You're done pushing to master!
-15. Issue the release on GitHub. Click on "Draft a new release" at
+18. Issue the release on GitHub. Click on "Draft a new release" at
     https://github.com/pydata/xarray/releases. Type in the version number
     and paste the release summary in the notes.
-16. Update the docs. Login to https://readthedocs.org/projects/xray/versions/
+19. Update the docs. Login to https://readthedocs.org/projects/xray/versions/
     and switch your new release tag (at the bottom) from "Inactive" to "Active".
     It should now build automatically.
-17. Issue the release announcement to mailing lists & Twitter. For bug fix releases, I 
+20. Issue the release announcement to mailing lists & Twitter. For bug fix releases, I 
     usually only email [email protected]. For major/feature releases, I will email a broader
     list (no more than once every 3-6 months):
       - [email protected]

ci/min_deps_check.py

@@ -23,9 +23,21 @@
     "pytest-env",
 }
 
-POLICY_MONTHS = {"python": 42, "numpy": 24, "pandas": 12, "scipy": 12}
-POLICY_MONTHS_DEFAULT = 6
-
+POLICY_MONTHS = {"python": 42, "numpy": 24, "setuptools": 42}
+POLICY_MONTHS_DEFAULT = 12
+POLICY_OVERRIDE = {
+    # dask < 2.9 has trouble with nan-reductions
+    # TODO remove this special case and the matching note in installing.rst
+    #      after January 2021.
+    "dask": (2, 9),
+    "distributed": (2, 9),
+    # setuptools-scm doesn't work with setuptools < 36.7 (Nov 2017).
+    # The conda metadata is malformed for setuptools < 38.4 (Jan 2018)
+    # (it's missing a timestamp which prevents this tool from working).
+    # TODO remove this special case and the matching note in installing.rst
+    #      after July 2021.
+    "setuptools": (38, 4),
+}
 has_errors = False
 
 
@@ -151,6 +163,11 @@ def process_pkg(
         policy_minor = minor
         policy_published_actual = published
 
+    try:
+        policy_major, policy_minor = POLICY_OVERRIDE[pkg]
+    except KeyError:
+        pass
+
     if (req_major, req_minor) < (policy_major, policy_minor):
         status = "<"
     elif (req_major, req_minor) > (policy_major, policy_minor):

ci/requirements/doc.yml

@@ -13,14 +13,18 @@ dependencies:
   - ipython
   - iris>=2.3
   - jupyter_client
+  - matplotlib-base=3.3.0
   - nbsphinx
   - netcdf4>=1.5
   - numba
   - numpy>=1.17
-  - pandas>=1.0
+  # FIXME https://github.com/pydata/xarray/issues/4287
+  # - pandas>=1.0
+  - pandas=1.0
   - rasterio>=1.1
   - seaborn
   - setuptools
-  - sphinx>=2.3
+  - sphinx=3.1
   - sphinx_rtd_theme>=0.4
+  - sphinx-autosummary-accessors
   - zarr>=2.4

ci/requirements/py36-bare-minimum.yml

@@ -10,4 +10,4 @@ dependencies:
   - pytest-env
   - numpy=1.15
   - pandas=0.25
-  - setuptools=41.2
+  - setuptools=38.4

ci/requirements/py36-min-all-deps.yml

@@ -43,7 +43,7 @@ dependencies:
   - rasterio=1.0
   - scipy=1.3
   - seaborn=0.9
-  - setuptools=41.2
+  - setuptools=38.4
   # - sparse  # See py36-min-nep18.yml
   - toolz=0.10
   - zarr=2.3

ci/requirements/py36-min-nep18.yml

@@ -15,6 +15,6 @@ dependencies:
   - pytest
   - pytest-cov
   - pytest-env
-  - scipy=1.2
-  - setuptools=41.2
+  - scipy=1.3
+  - setuptools=38.4
   - sparse=0.8

doc/api-hidden.rst

@@ -665,13 +665,10 @@
    backends.NetCDF4DataStore.encode
    backends.NetCDF4DataStore.encode_attribute
    backends.NetCDF4DataStore.encode_variable
-   backends.NetCDF4DataStore.get
    backends.NetCDF4DataStore.get_attrs
    backends.NetCDF4DataStore.get_dimensions
    backends.NetCDF4DataStore.get_encoding
    backends.NetCDF4DataStore.get_variables
-   backends.NetCDF4DataStore.items
-   backends.NetCDF4DataStore.keys
    backends.NetCDF4DataStore.load
    backends.NetCDF4DataStore.open
    backends.NetCDF4DataStore.open_store_variable
@@ -685,27 +682,20 @@
    backends.NetCDF4DataStore.store
    backends.NetCDF4DataStore.store_dataset
    backends.NetCDF4DataStore.sync
-   backends.NetCDF4DataStore.values
-   backends.NetCDF4DataStore.attrs
    backends.NetCDF4DataStore.autoclose
-   backends.NetCDF4DataStore.dimensions
    backends.NetCDF4DataStore.ds
    backends.NetCDF4DataStore.format
    backends.NetCDF4DataStore.is_remote
    backends.NetCDF4DataStore.lock
-   backends.NetCDF4DataStore.variables
 
    backends.H5NetCDFStore.close
    backends.H5NetCDFStore.encode
    backends.H5NetCDFStore.encode_attribute
    backends.H5NetCDFStore.encode_variable
-   backends.H5NetCDFStore.get
    backends.H5NetCDFStore.get_attrs
    backends.H5NetCDFStore.get_dimensions
    backends.H5NetCDFStore.get_encoding
    backends.H5NetCDFStore.get_variables
-   backends.H5NetCDFStore.items
-   backends.H5NetCDFStore.keys
    backends.H5NetCDFStore.load
    backends.H5NetCDFStore.open_store_variable
    backends.H5NetCDFStore.prepare_variable
@@ -718,39 +708,25 @@
    backends.H5NetCDFStore.store
    backends.H5NetCDFStore.store_dataset
    backends.H5NetCDFStore.sync
-   backends.H5NetCDFStore.values
-   backends.H5NetCDFStore.attrs
-   backends.H5NetCDFStore.dimensions
    backends.H5NetCDFStore.ds
-   backends.H5NetCDFStore.variables
 
    backends.PydapDataStore.close
-   backends.PydapDataStore.get
    backends.PydapDataStore.get_attrs
    backends.PydapDataStore.get_dimensions
    backends.PydapDataStore.get_encoding
    backends.PydapDataStore.get_variables
-   backends.PydapDataStore.items
-   backends.PydapDataStore.keys
    backends.PydapDataStore.load
    backends.PydapDataStore.open
    backends.PydapDataStore.open_store_variable
-   backends.PydapDataStore.values
-   backends.PydapDataStore.attrs
-   backends.PydapDataStore.dimensions
-   backends.PydapDataStore.variables
 
    backends.ScipyDataStore.close
    backends.ScipyDataStore.encode
    backends.ScipyDataStore.encode_attribute
    backends.ScipyDataStore.encode_variable
-   backends.ScipyDataStore.get
    backends.ScipyDataStore.get_attrs
    backends.ScipyDataStore.get_dimensions
    backends.ScipyDataStore.get_encoding
    backends.ScipyDataStore.get_variables
-   backends.ScipyDataStore.items
-   backends.ScipyDataStore.keys
    backends.ScipyDataStore.load
    backends.ScipyDataStore.open_store_variable
    backends.ScipyDataStore.prepare_variable
@@ -763,11 +739,7 @@
    backends.ScipyDataStore.store
    backends.ScipyDataStore.store_dataset
    backends.ScipyDataStore.sync
-   backends.ScipyDataStore.values
-   backends.ScipyDataStore.attrs
-   backends.ScipyDataStore.dimensions
    backends.ScipyDataStore.ds
-   backends.ScipyDataStore.variables
 
    backends.FileManager.acquire
    backends.FileManager.acquire_context

doc/api.rst

@@ -572,7 +572,9 @@ Universal functions
    With recent versions of numpy, dask and xarray, NumPy ufuncs are now
    supported directly on all xarray and dask objects. This obviates the need
    for the ``xarray.ufuncs`` module, which should not be used for new code
-   unless compatibility with versions of NumPy prior to v1.13 is required.
+   unless compatibility with versions of NumPy prior to v1.13 is
+   required. They will be removed once support for NumPy prior to
+   v1.17 is dropped.
 
 These functions are copied from NumPy, but extended to work on NumPy arrays,
 dask arrays and all xarray objects. You can find them in the ``xarray.ufuncs``

doc/computation.rst

@@ -188,9 +188,16 @@ a value when aggregating:
     r = arr.rolling(y=3, center=True, min_periods=2)
     r.mean()
 
+From version 0.17, xarray supports multidimensional rolling,
+
+.. ipython:: python
+
+    r = arr.rolling(x=2, y=3, min_periods=2)
+    r.mean()
+
 .. tip::
 
-   Note that rolling window aggregations are faster and use less memory when bottleneck_ is installed. This only applies to numpy-backed xarray objects.
+   Note that rolling window aggregations are faster and use less memory when bottleneck_ is installed. This only applies to numpy-backed xarray objects with 1d-rolling.
 
 .. _bottleneck: https://github.com/pydata/bottleneck/
 
@@ -227,9 +234,9 @@ windowed rolling, convolution, short-time FFT etc.
 .. ipython:: python
 
     # rolling with 2-point stride
-    rolling_da = r.construct("window_dim", stride=2)
+    rolling_da = r.construct(x="x_win", y="y_win", stride=2)
     rolling_da
-    rolling_da.mean("window_dim", skipna=False)
+    rolling_da.mean(["x_win", "y_win"], skipna=False)
 
 Because the ``DataArray`` given by ``r.construct('window_dim')`` is a view
 of the original array, it is memory efficient.
@@ -238,7 +245,7 @@ You can also use ``construct`` to compute a weighted rolling sum:
 .. ipython:: python
 
     weight = xr.DataArray([0.25, 0.5, 0.25], dims=["window"])
-    arr.rolling(y=3).construct("window").dot(weight)
+    arr.rolling(y=3).construct(y="window").dot(weight)
 
 .. note::
   numpy's Nan-aggregation functions such as ``nansum`` copy the original array.