Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add Coordinates.from_xindex method (+ refactor API doc) #10000

Merged
merged 12 commits into from
Feb 12, 2025
Merged

Add Coordinates.from_xindex method (+ refactor API doc) #10000

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
12 commits
Select commit Hold shift + click to select a range
734bffb
add Coordinates.from_xindex method
benbovy Jan 29, 2025
95938b0
doc: refactor Coordinates API reference
benbovy Jan 29, 2025
e110790
add tests
benbovy Jan 29, 2025
50d30fb
update what's new
benbovy Jan 29, 2025
ad757fa
fix doc build?
benbovy Jan 29, 2025
288c934
docstring tweaks
benbovy Jan 29, 2025
8113302
doc (api): add missing Coordinates.sizes property
benbovy Jan 29, 2025
07e6a37
update what's new (documentation)
benbovy Jan 29, 2025
ad78599
improve docstrings on Coordinates creation
benbovy Jan 30, 2025
7aff9ce
Merge branch 'main' into add-coordinate-from-xindex-method
benbovy Jan 31, 2025
0fbbd92
doc: move what's new entries after last release
benbovy Jan 31, 2025
86ceb91
Merge branch 'main' into add-coordinate-from-xindex-method
dcherian Feb 12, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
40 changes: 21 additions & 19 deletions doc/api-hidden.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,30 +9,12 @@
.. autosummary::
:toctree: generated/

Coordinates.from_pandas_multiindex
Coordinates.get
Coordinates.items
Coordinates.keys
Coordinates.values
Coordinates.dims
Coordinates.dtypes
Coordinates.variables
Coordinates.xindexes
Coordinates.indexes
Coordinates.to_dataset
Coordinates.to_index
Coordinates.update
Coordinates.assign
Coordinates.merge
Coordinates.copy
Coordinates.equals
Coordinates.identical

core.coordinates.DatasetCoordinates.get
core.coordinates.DatasetCoordinates.items
core.coordinates.DatasetCoordinates.keys
core.coordinates.DatasetCoordinates.values
core.coordinates.DatasetCoordinates.dims
core.coordinates.DatasetCoordinates.sizes
core.coordinates.DatasetCoordinates.dtypes
core.coordinates.DatasetCoordinates.variables
core.coordinates.DatasetCoordinates.xindexes
Expand Down Expand Up @@ -74,6 +56,7 @@
core.coordinates.DataArrayCoordinates.keys
core.coordinates.DataArrayCoordinates.values
core.coordinates.DataArrayCoordinates.dims
core.coordinates.DataArrayCoordinates.sizes
core.coordinates.DataArrayCoordinates.dtypes
core.coordinates.DataArrayCoordinates.variables
core.coordinates.DataArrayCoordinates.xindexes
Expand Down Expand Up @@ -104,6 +87,25 @@
core.weighted.DataArrayWeighted.obj
core.weighted.DataArrayWeighted.weights

core.coordinates.DataTreeCoordinates.get
core.coordinates.DataTreeCoordinates.items
core.coordinates.DataTreeCoordinates.keys
core.coordinates.DataTreeCoordinates.values
core.coordinates.DataTreeCoordinates.dims
core.coordinates.DataTreeCoordinates.sizes
core.coordinates.DataTreeCoordinates.dtypes
core.coordinates.DataTreeCoordinates.variables
core.coordinates.DataTreeCoordinates.xindexes
core.coordinates.DataTreeCoordinates.indexes
core.coordinates.DataTreeCoordinates.to_dataset
core.coordinates.DataTreeCoordinates.to_index
core.coordinates.DataTreeCoordinates.update
core.coordinates.DataTreeCoordinates.assign
core.coordinates.DataTreeCoordinates.merge
core.coordinates.DataTreeCoordinates.copy
core.coordinates.DataTreeCoordinates.equals
core.coordinates.DataTreeCoordinates.identical

core.accessor_dt.DatetimeAccessor.ceil
core.accessor_dt.DatetimeAccessor.floor
core.accessor_dt.DatetimeAccessor.round
Expand Down
101 changes: 80 additions & 21 deletions doc/api.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -67,6 +67,7 @@ Attributes
Dataset.attrs
Dataset.encoding
Dataset.indexes
Dataset.xindexes
Dataset.chunks
Dataset.chunksizes
Dataset.nbytes
Expand Down Expand Up @@ -275,6 +276,7 @@ Attributes
DataArray.attrs
DataArray.encoding
DataArray.indexes
DataArray.xindexes
DataArray.chunksizes

ndarray attributes
Expand Down Expand Up @@ -897,6 +899,84 @@ Methods copied from :py:class:`numpy.ndarray` objects, here applying to the data
.. DataTree.sortby
.. DataTree.broadcast_like

Coordinates
===========

Creating coordinates
--------------------

.. autosummary::
:toctree: generated/

Coordinates
Coordinates.from_xindex
Coordinates.from_pandas_multiindex

Attributes
----------

.. autosummary::
:toctree: generated/

Coordinates.dims
Coordinates.sizes
Coordinates.dtypes
Coordinates.variables
Coordinates.indexes
Coordinates.xindexes

Dictionary Interface
--------------------

Coordinates implement the mapping interface with keys given by variable names
and values given by ``DataArray`` objects.

.. autosummary::
:toctree: generated/

Coordinates.__getitem__
Coordinates.__setitem__
Coordinates.__delitem__
Coordinates.update
Coordinates.get
Coordinates.items
Coordinates.keys
Coordinates.values

Coordinates contents
--------------------

.. autosummary::
:toctree: generated/

Coordinates.to_dataset
Coordinates.to_index
Coordinates.assign
Coordinates.merge
Coordinates.copy

Comparisons
-----------

.. autosummary::
:toctree: generated/

Coordinates.equals
Coordinates.identical

Proxies
-------

Coordinates that are accessed from the ``coords`` property of Dataset, DataArray
and DataTree objects, respectively.

.. autosummary::
:toctree: generated/

core.coordinates.DatasetCoordinates
core.coordinates.DataArrayCoordinates
core.coordinates.DataTreeCoordinates

Universal functions
===================

Expand Down Expand Up @@ -1107,27 +1187,6 @@ Coder objects

coders.CFDatetimeCoder

Coordinates objects
===================

Dataset
-------

.. autosummary::
:toctree: generated/

core.coordinates.DatasetCoordinates
core.coordinates.DatasetCoordinates.dtypes

DataArray
---------

.. autosummary::
:toctree: generated/

core.coordinates.DataArrayCoordinates
core.coordinates.DataArrayCoordinates.dtypes

Plotting
========

Expand Down
11 changes: 10 additions & 1 deletion doc/whats-new.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,9 @@ v2025.02.0 (unreleased)

New Features
~~~~~~~~~~~~
- Added :py:meth:`Coordinates.from_xindex` as convenience for creating a new :py:class:`Coordinates` object
directly from an existing Xarray index object if the latter supports it (:pull:`10000`)
By `Benoit Bovy <https://github.com/benbovy>`_.
- Allow kwargs in :py:meth:`DataTree.map_over_datasets` and :py:func:`map_over_datasets` (:issue:`10009`, :pull:`10012`).
By `Kai Mühlbauer <https://github.com/kmuehlbauer>`_.
- support python 3.13 (no free-threading) (:issue:`9664`, :pull:`9681`)
Expand Down Expand Up @@ -52,7 +55,8 @@ Bug fixes

Documentation
~~~~~~~~~~~~~

- Better expose the :py:class:`Coordinates` class in API reference (:pull:`10000`)
By `Benoit Bovy <https://github.com/benbovy>`_.

Internal Changes
~~~~~~~~~~~~~~~~
Expand Down Expand Up @@ -131,6 +135,11 @@ New Features
:py:class:`pandas.DatetimeIndex` (:pull:`9965`). By `Spencer Clark
<https://github.com/spencerkclark>`_ and `Kai Mühlbauer
<https://github.com/kmuehlbauer>`_.
- :py:meth:`DatasetGroupBy.first` and :py:meth:`DatasetGroupBy.last` can now use ``flox`` if available. (:issue:`9647`)
By `Deepak Cherian <https://github.com/dcherian>`_.

Breaking changes
~~~~~~~~~~~~~~~~
- Adds shards to the list of valid_encodings in the zarr backend, so that
sharded Zarr V3s can be written (:issue:`9947`, :pull:`9948`).
By `Jacob Prince_Bieker <https://github.com/jacobbieker>`_
Expand Down
44 changes: 40 additions & 4 deletions xarray/core/coordinates.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -200,10 +200,17 @@ class Coordinates(AbstractCoordinates):

- returned via the :py:attr:`Dataset.coords`, :py:attr:`DataArray.coords`,
and :py:attr:`DataTree.coords` properties,
- built from Pandas or other index objects
(e.g., :py:meth:`Coordinates.from_pandas_multiindex`),
- built directly from coordinate data and Xarray ``Index`` objects (beware that
no consistency check is done on those inputs),
- built from Xarray or Pandas index objects
(e.g., :py:meth:`Coordinates.from_xindex` or
:py:meth:`Coordinates.from_pandas_multiindex`),
- built manually from input coordinate data and Xarray ``Index`` objects via
:py:meth:`Coordinates.__init__` (beware that no consistency check is done
on those inputs).

To create new coordinates from an existing Xarray ``Index`` object, use
:py:meth:`Coordinates.from_xindex` instead of
:py:meth:`Coordinates.__init__`. The latter is useful, e.g., for creating
coordinates with no default index.

Parameters
----------
Expand Down Expand Up @@ -352,6 +359,35 @@ def _construct_direct(
)
return obj

@classmethod
def from_xindex(cls, index: Index) -> Self:
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Seems a little inconsistent that the signature of this method only accepts one index, when the __init__ accepts multiple?

I'm also a bit unclear why you need to pass a Mapping from str names to indexes to __init__ but this method apparently doesn't require names?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ah yes I agree it is confusing and I'll clarify this.

__init__ is the low-level(-ish) constructor where we can pass any arbitrary mapping of coordinate variables (and optionally a mapping of Xarray indexes) that are just passed to a direct constructor with only little safety checks, whereas from_xindex creates a new Coordinates object from exactly one Xarray index via creating the variables and indexes mappings from the index object.

For example with a custom coordinate transform index (#9543):

class CustomIndex(CoordinateTransformIndex):
    ...

index = CustomIndex(...)

writing

coords = Coordinates.from_xindex(index)

is equivalent but safer and more convenient than writing

vars = index.create_variables()
indexes = {name: index for name in vars}
coords = Coordinates(coords=vars, indexes=indexes)

Alternatively to Coordinates.from_xindex() we could provide more specific constructors like Coordinates.from_transform() (alongside the already existing Coordinates.from_pandas_multiindex()), but Coordinates.from_xindex() is more generic and future-proof, i.e., it will just work with any custom index that can be created from anything other than pre-existing Xarray coordinates.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Okay thank you for the explanation!

"""Create Xarray coordinates from an existing Xarray index.

Parameters
----------
index : Index
Xarray index object. The index must support generating new
coordinate variables from itself.

Returns
-------
coords : Coordinates
A collection of Xarray indexed coordinates created from the index.

"""
variables = index.create_variables()

if not variables:
raise ValueError(
"`Coordinates.from_xindex()` only supports index objects that can generate "
"new coordinate variables from scratch. The given index (shown below) did not "
f"create any coordinate.\n{index!r}"
)

indexes = {name: index for name in variables}

return cls(coords=variables, indexes=indexes)

@classmethod
def from_pandas_multiindex(cls, midx: pd.MultiIndex, dim: Hashable) -> Self:
"""Wrap a pandas multi-index as Xarray coordinates (dimension + levels).
Expand Down
25 changes: 24 additions & 1 deletion xarray/tests/test_coordinates.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,7 @@
from __future__ import annotations

from collections.abc import Mapping

import numpy as np
import pandas as pd
import pytest
Expand All @@ -8,7 +10,7 @@
from xarray.core.coordinates import Coordinates
from xarray.core.dataarray import DataArray
from xarray.core.dataset import Dataset
from xarray.core.indexes import PandasIndex, PandasMultiIndex
from xarray.core.indexes import Index, PandasIndex, PandasMultiIndex
from xarray.core.variable import IndexVariable, Variable
from xarray.tests import assert_identical, source_ndarray

Expand Down Expand Up @@ -73,6 +75,27 @@ def test_init_dim_sizes_conflict(self) -> None:
with pytest.raises(ValueError):
Coordinates(coords={"foo": ("x", [1, 2]), "bar": ("x", [1, 2, 3, 4])})

def test_from_xindex(self) -> None:
idx = PandasIndex([1, 2, 3], "x")
coords = Coordinates.from_xindex(idx)

assert isinstance(coords.xindexes["x"], PandasIndex)
assert coords.xindexes["x"].equals(idx)

expected = PandasIndex(idx, "x").create_variables()
assert list(coords.variables) == list(expected)
assert_identical(expected["x"], coords.variables["x"])

def test_from_xindex_error(self) -> None:
class CustomIndexNoCoordsGenerated(Index):
def create_variables(self, variables: Mapping | None = None):
return {}

idx = CustomIndexNoCoordsGenerated()

with pytest.raises(ValueError, match=".*index.*did not create any coordinate"):
Coordinates.from_xindex(idx)

def test_from_pandas_multiindex(self) -> None:
midx = pd.MultiIndex.from_product([["a", "b"], [1, 2]], names=("one", "two"))
coords = Coordinates.from_pandas_multiindex(midx, "x")
Expand Down
Loading