3
���h�� � @ s, d dl Z d dlZd dlZd dlZd dlZddlmZ ddl m
Z
ddl m Z ddl m Z ddl m
Z
dd lmZ dd
lmZmZmZ dd lmZmZ dd lmZ dd
lmZmZm Z m Z ddl m Z m Z m!Z!m"Z"m#Z# ddl$m%Z% dd� Z&G dd� de'�Z(G dd� de'�Z)G dd� dee�Z*e j+e*dd� dS )� N� )� _PlotMethods� )�indexing)�groupby)�ops)�utils)�align)�
AbstractArray�BaseDataObject�squeeze)�DataArrayCoordinates�Indexes)�Dataset)� iteritems�
basestring� OrderedDict�zip)� as_variable�Variable�as_compatible_data�
Coordinate�default_index_coordinate)� format_itemc s |dk r<t j|� r<t|�t| �kr<tdt|�t| �f ��t� t�rL� f� � dkr�dd� tt| ��D �� |dk r�t|�t| �kr�t j|�r�t|j� �� n:x8t t
� |��D ]&\}\}}t ||d�j � }|j
� |< q�W t� �� n$x"� D ]}t|t�s�td| ��q�W t� }t j|��r@xl|j� D ] \}} t | |dd�||< �q W nB|dk �r�x6t
� |�D ](\}}t ||dd�}
|f|
_|
||< �qVW x0t
� | �D ]"\}} ||k�r�t|| �||< �q�W xD|j� D ]8\}} t� fd d
�| jD ���r�td || j� f ���q�W |� fS ) z*All the logic for creating a new DataArrayNz`coords is not dict-like, but it has %s items, which does not match the %s dimensions of the datac S s g | ] }d | �qS )zdim_%s� )�.0�nr r �5/tmp/pip-build-5_djhm0z/xray/xarray/core/dataarray.py�
<listcomp>% s z*_infer_coords_and_dims.<locals>.<listcomp>)�keyz dimension %s is not a stringT)r �copyc 3 s | ]}|� kV qd S )Nr )r �d)�dimsr r � <genexpr>D s z)_infer_coords_and_dims.<locals>.<genexpr>zZcoordinate %s has dimensions %s, but these are not a subset of the DataArray dimensions %s)r � is_dict_like�len�
ValueError�
isinstancer �range�list�keys� enumerater r Zto_coord�name�tuple� TypeErrorr �itemsr" r �any) �shape�coordsr" r �dim�coordr! Z
new_coords�k�v�var�sizer )r" r �_infer_coords_and_dims sF
r9 c @ s, e Zd Zdd� Zdd� Zdd� Zdd� Zd S )
� _LocIndexerc C s
|| _ d S )N)�
data_array)�selfr; r r r �__init__M s z_LocIndexer.__init__c sd �fdd�� t j|�r0t� fdd�t|�D ��S tj|�jj�}t� fdd�t �jj
|�D ��S d S )Nc s � j j| }tj||�S )N)r; �indexesr Zconvert_label_indexer)r3 �labels�index)r< r r �lookup_positionsQ s z0_LocIndexer._remap_key.<locals>.lookup_positionsc 3 s | ]\}}|� ||�fV qd S )Nr )r r3 r? )rA r r r# V s z)_LocIndexer._remap_key.<locals>.<genexpr>c 3 s | ]\}}� ||�V qd S )Nr )r r3 r? )rA r r r# [ s ) r r$ �dictr r �expanded_indexerr; �ndimr- r r" )r< r r )rA r< r �
_remap_keyP s
z_LocIndexer._remap_keyc C s | j | j|� S )N)r; rE )r< r r r r � __getitem__^ s z_LocIndexer.__getitem__c C s || j | j|�< d S )N)r; rE )r< r �valuer r r � __setitem__a s z_LocIndexer.__setitem__N)�__name__�
__module__� __qualname__r= rE rF rH r r r r r: L s r: c @ s e Zd ZdZdd� ZdS )�
_ThisArrayz�An instance of this object is used as the key corresponding to the
variable when converting arbitrary DataArray objects to datasets
c C s dS )Nz <this-array>r )r< r r r �__repr__i s z_ThisArray.__repr__N)rI rJ rK �__doc__rM r r r r rL e s rL c @ s� e Zd ZdZejZd�dd�Ze� Z dde fdd�Z
e fdd �Z e � Z
d
d � Ze fd d
�Zdd� Zd�dd�Zd�dd�Zedd� �Zejdd� �Zedd� �Zedd� �Zed d � �Zed d � �Zed d!� �Zed"d#� �Zd$d%� Z ed&d'� �Z e jd(d'� �Z ed)d*� �Z e jd+d*� �Z ed,d-� �Z d.d/� Z ed0d1� �Z!e!jd2d1� �Z!d3d4� Z"d5d6� Z#d7d8� Z$d9d:� Z%ed;d<� �Z&d=d>� Z'ed?d@� �Z(edAdB� �Z)e)jdCdB� �Z)edDdE� �Z*e*jdFdE� �Z*edGdH� �Z+edIdJ� �Z,d�dKdL�Z-dMdN� Z.dOdP� Z/d�dQdR�Z0dSdT� Z1d�dUdV�Z2dZ3edWdX� �Z4d�dYdZ�Z5d[d\� Z6d�d]d^�Z7d�d`da�Z8d�dbdc�Z9d�ddde�Z:d�dfdg�Z;dhdi� Z<djdk� Z=dldm� Z>dndo� Z?dpdq� Z@d�drds�ZAd�dtdu�ZBd�dwdx�ZCdydz� ZDd�d{d|�ZEd}d~� ZFd�dd��ZGd�d�� ZHd�d�d��ZIeJd�d�� �ZKd�d�� ZLeJd�d�� �ZMd�d�� ZNd�d�� ZOd�d�� ZPd�d�� ZQe� ZRd�d�d��ZSd�d�d��ZTeUd�d�� �ZVeUd�d�d���ZWeUd�d�� �ZXed�d�� �ZYd�d�d��ZZd�d�d��Z[d�d�� Z\d�d�� Z]ed�d�� �Z^ed�d�� �Z_dS )�� DataArraya� N-dimensional array with labeled coordinates and dimensions.
DataArray provides a wrapper around numpy ndarrays that uses labeled
dimensions and coordinates to support metadata aware operations. The API is
similar to that for the pandas Series or DataFrame, but DataArray objects
can have any number of dimensions, and their contents have fixed data
types.
Additional features over raw numpy arrays:
- Apply operations over dimensions by name: ``x.sum('time')``.
- Select or assign values by integer location (like numpy): ``x[:10]``
or by label (like pandas): ``x.loc['2014-01-01']`` or
``x.sel(time='2014-01-01')``.
- Mathematical operations (e.g., ``x - y``) vectorize across multiple
dimensions (known in numpy as "broadcasting") based on dimension names,
regardless of their original order.
- Keep track of arbitrary metadata in the form of a Python dictionary:
``x.attrs``
- Convert to a pandas Series: ``x.to_series()``.
Getting items from or doing mathematical operations with a DataArray
always returns another DataArray.
Attributes
----------
dims : tuple
Dimension names associated with this array.
values : np.ndarray
Access or modify DataArray values as a numpy array.
coords : dict-like
Dictionary of Coordinate objects that label values along each dimension.
name : str or None
Name of this array.
attrs : OrderedDict
Dictionary for holding arbitrary metadata.
NFc C sF |r.|}|dkst �|dks t �|dks,t �n�|dkr�t|dd�}t|tj�rX|jg}nNt|tj�rr|j|jg}n4t|tjt f�r�|g}n t|tj
�r�|j |j |j
g}|dkr�t|dt|dd��}|dkr�t|dd�}|dkr�t|dd�}|dkr�t|dd�}t|�}t|j||�\}}t||||dd�}|| _|| _|| _d| _dS ) a�
Parameters
----------
data : array_like
Values for this array. Must be an ``numpy.ndarray``, ndarray like,
or castable to an ``ndarray``. If a self-described xarray or pandas
object, attempts are made to use this array's metadata to fill in
other unspecified arguments. A view of the array's data is used
instead of a copy if possible.
coords : sequence or dict of array_like objects, optional
Coordinates (tick labels) to use for indexing along each dimension.
If dict-like, should be a mapping from dimension names to the
corresponding coordinates.
dims : str or sequence of str, optional
Name(s) of the the data dimension(s). Must be either a string (only
for 1D data) or a sequence of strings with length equal to the
number of dimensions. If this argument is omitted, dimension names
are taken from ``coords`` (if possible) and otherwise default to
``['dim_0', ... 'dim_n']``.
name : str or None, optional
Name of this array.
attrs : dict_like or None, optional
Attributes to assign to the new instance. By default, an empty
attribute dictionary is initialized.
encoding : dict_like or None, optional
Dictionary specifying how to encode this array's data into a
serialized format like netCDF4. Currently used keys (for netCDF)
include '_FillValue', 'scale_factor', 'add_offset', 'dtype',
'units' and 'calendar' (the later two only for datetime arrays).
Unrecognized keys are ignored.
Nr2 r" r, �attrs�encodingT)�fastpath)�AssertionError�getattrr' �pd�Seriesr@ � DataFrame�columns�Indexr �Panelr/ Z
major_axisZ
minor_axisr r9 r1 r � _variable�_coords�_name� _initialized) r< �datar2 r" r, rP rQ rR �variabler r r r= � s< !
zDataArray.__init__c C s@ |d kr| j }|d kr | j}|| jkr,| j}t| �|||dd�S )NT)r, rR )r` r\ �_DataArray__defaultr, �type)r< r` r2 r, r r r �_replace� s
zDataArray._replacec sF |j | j krd }n&t|j �� t� fdd�| jj� D ��}| j|||�S )Nc 3 s( | ] \}}t |j�� kr||fV qd S )N)�setr" )r r5 r6 )� allowed_dimsr r r# � s z5DataArray._replace_maybe_drop_dims.<locals>.<genexpr>)r" rd r r\ r/ rc )r< r` r, r2 r )re r �_replace_maybe_drop_dims� s
z"DataArray._replace_maybe_drop_dimsc C s | j | jdd�S )NF)r, � shallow_copy)�_to_dataset_whole�_DataArray__this_array)r< r r r �_to_temp_dataset� s zDataArray._to_temp_datasetc C s" |j j| j�}|j }| j|||�S )N)�
_variables�popri rc )r< �datasetr, r` r2 r r r �_from_temp_dataset� s z DataArray._from_temp_datasetc sH �fdd��t � �fdd��j� D ��}�jj� }|� = t||�j�S )Nc s � j | |i j| �}i |_|S )N)�loc�droprP )r3 �label�array)r< r r �subset� s z+DataArray._to_dataset_split.<locals>.subsetc s g | ]}t |��� |�f�qS r )�str)r rq )r3 rs r r r � s z/DataArray._to_dataset_split.<locals>.<listcomp>)r r> r2 �
to_datasetr rP )r< r3 � variablesr2 r )r3 r< rs r �_to_dataset_split� s
zDataArray._to_dataset_splitTc C sL |d kr| j }|d kr td��|| jkr0td��| jj|d�}| j||< |S )NzSunable to convert unnamed DataArray to a Dataset without providing an explicit namezUcannot create a Dataset from a DataArray with the same name as one of its coordinates)rg )r, r& r2 Z _to_datasetr` )r< r, rg rm r r r rh s
zDataArray._to_dataset_wholec C sZ |dk r*|| j kr*tjdtdd� |}d}|dk rL|dk rBtd��| j|�S | j|�S dS )a� Convert a DataArray to a Dataset.
Parameters
----------
dim : str, optional
Name of the dimension on this array along which to split this array
into separate variables. If not provided, this array is converted
into a Dataset of one variable.
name : str, optional
Name to substitute for this array's name. Only valid is ``dim`` is
not provided.
Returns
-------
dataset : Dataset
Nzuthe order of the arguments on DataArray.to_dataset has changed; you now need to supply ``name`` as a keyword argumentr )�
stacklevelz)cannot supply both dim and name arguments)r" �warnings�warn�
FutureWarningr. rw rh )r< r3 r, r r r ru s
zDataArray.to_datasetc C s | j S )z The name of this array.
)r] )r< r r r r, 2 s zDataArray.namec C s
|| _ d S )N)r] )r< rG r r r r, 8 s c C s | j S )N)r[ )r< r r r r` < s zDataArray.variablec C s | j jS )N)r` �dtype)r< r r r r| @ s zDataArray.dtypec C s | j jS )N)r` r1 )r< r r r r1 D s zDataArray.shapec C s | j jS )N)r` r8 )r< r r r r8 H s zDataArray.sizec C s | j jS )N)r` �nbytes)r< r r r r} L s zDataArray.nbytesc C s | j jS )N)r` rD )r< r r r rD P s zDataArray.ndimc C s
t | j�S )N)r% r` )r< r r r �__len__T s zDataArray.__len__c C s | j jS )z)The array's data as a dask or numpy array)r` r_ )r< r r r r_ W s zDataArray.datac C s || j _d S )N)r` r_ )r< rG r r r r_ \ s c C s | j jS )z#The array's data as a numpy.ndarray)r` �values)r< r r r r ` s zDataArray.valuesc C s || j _d S )N)r` r )r< rG r r r r e s c C s | j jS )N)r` �
_in_memory)r< r r r r� i s zDataArray._in_memoryc C s
| j j� S )zVConvert this variable to a pandas.Index. Only possible for 1D
arrays.
)r` �to_index)r< r r r r� m s zDataArray.to_indexc C s | j jS )z+Dimension names associated with this array.)r` r" )r< r r r r" s s zDataArray.dimsc C s t d��d S )NzMyou cannot assign dims on a DataArray. Use .rename() or .swap_dims() instead.)�AttributeError)r< rG r r r r" x s c C s0 t j|�r|S tj|| j�}tt| j|��S d S )N)r r$ r rC rD rB r r" )r< r r r r �_item_key_to_dict} s
zDataArray._item_key_to_dictc C sp t |t�rZddlm} y| j| }W n& tk
rJ || j|�\}}}Y nX | j||d�S | jf | j|��S d S )Nr )�_get_virtual_variable)r, ) r' r rm r� r\ �KeyErrorrf �iselr� )r< r r� r7 �_r r r rF � s
zDataArray.__getitem__c C s$ t |t�r|| j|< n
|| j|< d S )N)r' r r2 r` )r< r rG r r r rH � s
zDataArray.__setitem__c C s | j |= d S )N)r2 )r< r r r r � __delitem__� s zDataArray.__delitem__c C s | j | jgS )z:List of places to look-up items for attribute-style access)r2 rP )r< r r r �
_attr_sources� s zDataArray._attr_sourcesc C s
|| j kS )N)r\ )r< r r r r � __contains__� s zDataArray.__contains__c C s t | �S )z;Attribute for location based indexing like pandas.
)r: )r< r r r ro � s z
DataArray.locc C s | j jS )z6Dictionary storing arbitrary metadata with this array.)r` rP )r< r r r rP � s zDataArray.attrsc C s || j _d S )N)r` rP )r< rG r r r rP � s c C s | j jS )zWDictionary of format-specific settings for how this array should be
serialized.)r` rQ )r< r r r rQ � s zDataArray.encodingc C s || j _d S )N)r` rQ )r< rG r r r rQ � s c C s t | �S )zJOrderedDict of pandas.Index objects used for label based indexing
)r )r< r r r r> � s zDataArray.indexesc C s t | �S )z8Dictionary-like container of coordinate arrays.
)r
)r< r r r r2 � s zDataArray.coordsc C s� |r| rt d��|dkr.t| j�t| j� }| jj� j||�}|rb|rR|j| _q�| j|jd�S n"| j dkrtt d��| j
|| j < |S dS )az Given names of coordinates, reset them to become variables.
Parameters
----------
names : str or list of str, optional
Name(s) of non-index coordinates in this dataset to reset into
variables. By default, all non-index coordinates are reset.
drop : bool, optional
If True, remove coordinates instead of converting them into
variables.
inplace : bool, optional
If True, modify this dataset inplace. Otherwise, create a new
object.
Returns
-------
Dataset, or DataArray if ``drop == True``
zIcannot reset coordinates in-place on a DataArray without ``drop == True``N)r2 z<cannot reset_coords with drop=False on an unnamed DataArrray) r& rd r2 r" ru � reset_coordsrk r\ rc r, r` )r< �namesrp Zinplacerm r r r r� � s
zDataArray.reset_coordsc C s* | j � j� }| j|�}|j| _|j| _| S )a� Manually trigger loading of this array's data from disk or a
remote source into memory and return this array.
Normally, it should not be necessary to call this method in user code,
because all xarray functions should either work on deferred data or
load data automatically. However, this method can be necessary when
working with many file objects on disk.
)rj �loadrn r[ r\ )r< �ds�newr r r r� � s
zDataArray.loadc C s t jdtdd� | j� S )NzHthe DataArray method `load_data` has been deprecated; use `load` insteadr )rx )ry rz r{ r� )r< r r r � load_data� s
zDataArray.load_datac s6 | j j� d�}t� fdd�| jj� D ��}| j||�S )a
Returns a copy of this array.
If `deep=True`, a deep copy is made of all variables in the underlying
dataset. Otherwise, a shallow copy is made, so each variable in the new
array's dataset is also a variable in this array's dataset.
)�deepc 3 s" | ]\}}||j � d �fV qdS ))r� N)r )r r5 r6 )r� r r r# s z!DataArray.copy.<locals>.<genexpr>)r` r r r\ r/ rc )r< r� r` r2 r )r� r r s zDataArray.copyc C s | j dd�S )NF)r� )r )r< r r r �__copy__ s zDataArray.__copy__c C s | j dd�S )NT)r� )r )r< �memor r r � __deepcopy__ s zDataArray.__deepcopy__c C s | j jS )zYBlock dimensions for this array's data or None if it's not a dask
array.
)r` �chunks)r< r r r r� s zDataArray.chunksc C s6 t |ttf�r tt| j|��}| j� j|�}| j|�S )a� Coerce this array's data into a dask arrays with the given chunks.
If this variable is a non-dask array, it will be converted to dask
array. If it's a dask array, it will be rechunked to the given chunk
sizes.
If neither chunks is not provided for one or more dimensions, chunk
sizes along that dimension will not be updated; non-dask arrays will be
converted into dask arrays with a single block.
Parameters
----------
chunks : int, tuple or dict, optional
Chunk sizes along each dimension, e.g., ``5``, ``(5, 5)`` or
``{'x': 5, 'y': 5}``.
Returns
-------
chunked : xarray.DataArray
) r' r) r- rB r r" rj �chunkrn )r< r� r� r r r r� ! s zDataArray.chunkc K s | j � jf |�}| j|�S )z�Return a new DataArray whose dataset is given by integer indexing
along the specified dimension(s).
See Also
--------
Dataset.isel
DataArray.sel
)rj r� rn )r< �indexersr� r r r r� < s zDataArray.iselc K s | j f tj| |||d��S )z�Return a new DataArray whose dataset is given by selecting
index labels along the specified dimension(s).
See Also
--------
Dataset.sel
DataArray.isel
)�method� tolerance)r� r Zremap_label_indexers)r< r� r� r� r r r �selH s
z
DataArray.sel�pointsc K s" | j � jf d|i|��}| j|�S )z�Return a new DataArray whose dataset is given by pointwise integer
indexing along the specified dimension(s).
See Also
--------
Dataset.isel_points
r3 )rj � isel_pointsrn )r< r3 r� r� r r r r� T s zDataArray.isel_pointsc K s& | j � jf |||d�|��}| j|�S )z�Return a new DataArray whose dataset is given by pointwise selection
of index labels along the specified dimension(s).
See Also
--------
Dataset.sel_points
)r3 r� r� )rj �
sel_pointsrn )r< r3 r� r� r� r� r r r r� _ s
zDataArray.sel_pointsc s4 t � fdd�|jj� D ��}� jf |||d�|��S )aU Conform this object onto the indexes of another object, filling
in missing values with NaN.
Parameters
----------
other : Dataset or DataArray
Object with an 'indexes' attribute giving a mapping from dimension
names to pandas.Index objects, which provides coordinates upon
which to index the variables in this dataset. The indexes on this
other object need not be the same as the indexes on this
dataset. Any mis-matched index values will be filled in with
NaN, and any mis-matched dimension names will simply be ignored.
method : {None, 'nearest', 'pad'/'ffill', 'backfill'/'bfill'}, optional
Method to use for filling index values from other not found on this
data array:
* None (default): don't fill gaps
* pad / ffill: propgate last valid index value forward
* backfill / bfill: propagate next valid index value backward
* nearest: use nearest valid index value (requires pandas>=0.16)
tolerance : optional
Maximum distance between original and new labels for inexact
matches. The values of the index at the matching locations most
satisfy the equation ``abs(index[indexer] - target) <= tolerance``.
Requires pandas>=0.17.
copy : bool, optional
If `copy=True`, the returned array's dataset contains only copied
variables. If `copy=False` and no reindexing is required then
original variables from this array's dataset are returned.
Returns
-------
reindexed : DataArray
Another dataset array, with this array's data but coordinates from
the other object.
See Also
--------
DataArray.reindex
align
c 3 s$ | ] \}}|� j kr||fV qd S )N)r" )r r5 r6 )r< r r r# � s z)DataArray.reindex_like.<locals>.<genexpr>)r� r� r )rB r> r/ �reindex)r< �otherr� r� r r� r )r< r � reindex_likel s * zDataArray.reindex_likec K s& | j � jf |||d�|��}| j|�S )a� Conform this object onto a new set of indexes, filling in
missing values with NaN.
Parameters
----------
copy : bool, optional
If `copy=True`, the returned array's dataset contains only copied
variables. If `copy=False` and no reindexing is required then
original variables from this array's dataset are returned.
method : {None, 'nearest', 'pad'/'ffill', 'backfill'/'bfill'}, optional
Method to use for filling index values in ``indexers`` not found on
this data array:
* None (default): don't fill gaps
* pad / ffill: propgate last valid index value forward
* backfill / bfill: propagate next valid index value backward
* nearest: use nearest valid index value (requires pandas>=0.16)
tolerance : optional
Maximum distance between original and new labels for inexact
matches. The values of the index at the matching locations most
satisfy the equation ``abs(index[indexer] - target) <= tolerance``.
Requires pandas>=0.17.
**indexers : dict
Dictionary with keys given by dimension names and values given by
arrays of coordinates tick labels. Any mis-matched coordinate values
will be filled in with NaN, and any mis-matched dimension names will
simply be ignored.
Returns
-------
reindexed : DataArray
Another dataset array, with this array's data but replaced
coordinates.
See Also
--------
DataArray.reindex_like
align
)r� r� r )rj r� rn )r< r� r� r r� r� r r r r� � s (
zDataArray.reindexc C sL t j|�r<|j� }|j| j| j�}| j� j|�}| j||�S | j|d�S dS )aS Returns a new DataArray with renamed coordinates and/or a new name.
Parameters
----------
new_name_or_name_dict : str or dict-like
If the argument is dict-like, it it used as a mapping from old
names to new names for coordinates (and/or this array itself).
Otherwise, use the argument as the new name for this array.
Returns
-------
renamed : DataArray
Renamed array or array with renamed coordinates.
See Also
--------
Dataset.rename
DataArray.swap_dims
)r, N) r r$ r rl r, rj �renamern rc )r< Znew_name_or_name_dictZ name_dictr, rm r r r r� � s
zDataArray.renamec C s | j � j|�}| j|�S )a\ Returns a new DataArray with swapped dimensions.
Parameters
----------
dims_dict : dict-like
Dictionary whose keys are current dimension names and whose values
are new names. Each value must already be a coordinate on this
array.
inplace : bool, optional
If True, swap dimensions in-place. Otherwise, return a new object.
Returns
-------
renamed : Dataset
DataArray with swapped dimensions.
See Also
--------
DataArray.rename
Dataset.swap_dims
)rj � swap_dimsrn )r< Z dims_dictr� r r r r� � s zDataArray.swap_dimsc K s | j � jf |�}| j|�S )a�
Stack any number of existing dimensions into a single new dimension.
New dimensions will be added at the end, and the corresponding
coordinate variables will be combined into a MultiIndex.
Parameters
----------
**dimensions : keyword arguments of the form new_name=(dim1, dim2, ...)
Names of new dimensions, and the existing dimensions that they
replace.
Returns
-------
stacked : DataArray
DataArray with stacked data.
Example
-------
>>> arr = DataArray(np.arange(6).reshape(2, 3),
... coords=[('x', ['a', 'b']), ('y', [0, 1, 2])])
>>> arr
<xray.DataArray (x: 2, y: 3)>
array([[0, 1, 2],
[3, 4, 5]])
Coordinates:
* x (x) |S1 'a' 'b'
* y (y) int64 0 1 2
>>> stacked = arr.stack(z=('x', 'y'))
>>> stacked.indexes['z']
MultiIndex(levels=[[u'a', u'b'], [0, 1, 2]],
labels=[[0, 0, 0, 1, 1, 1], [0, 1, 2, 0, 1, 2]],
names=[u'x', u'y'])
See also
--------
DataArray.unstack
)rj �stackrn )r< Z
dimensionsr� r r r r� � s (zDataArray.stackc C s | j � j|�}| j|�S )a�
Unstack an existing dimension corresponding to a MultiIndex into
multiple new dimensions.
New dimensions will be added at the end.
Parameters
----------
dim : str
Name of the existing dimension to unstack.
Returns
-------
unstacked : DataArray
Array with unstacked data.
See also
--------
DataArray.stack
)rj �unstackrn )r< r3 r� r r r r� ) s zDataArray.unstackc G s | j j|� }| j|�S )a^ Return a new DataArray object with transposed dimensions.
Parameters
----------
*dims : str, optional
By default, reverse the dimensions. Otherwise, reorder the
dimensions to this order.
Returns
-------
transposed : DataArray
The returned DataArray's array is transposed.
Notes
-----
Although this operation returns a view of this array's data, it is
not lazy -- the data will be fully loaded.
See Also
--------
numpy.transpose
Dataset.transpose
)r` � transposerc )r< r" r` r r r r� A s zDataArray.transposec C s t | tt| j| j��|�S )a� Return a new DataArray object with squeezed data.
Parameters
----------
dim : None or str or tuple of str, optional
Selects a subset of the length one dimensions. If a dimension is
selected with length greater than one, an error is raised. If
None, all length one dimensions are squeezed.
Returns
-------
squeezed : DataArray
This array, but with with all or a subset of the dimensions of
length 1 removed.
Notes
-----
Although this operation returns a view of this array's data, it is
not lazy -- the data will be fully loaded.
See Also
--------
numpy.squeeze
)r rB r r" r1 )r< r3 r r r r \ s zDataArray.squeezec C s* t j|�r|g}| j� j||�}| j|�S )a� Drop coordinates or index labels from this DataArray.
Parameters
----------
labels : str
Names of coordinate variables or index labels to drop.
dim : str, optional
Dimension along which to drop index labels. By default (if
``dim is None``), drops coordinates rather than index labels.
Returns
-------
dropped : DataArray
)r Z is_scalarrj rp rn )r< r? r3 r� r r r rp w s
zDataArray.dropr0 c C s | j � j|||d�}| j|�S )ac Returns a new array with dropped labels for missing values along
the provided dimension.
Parameters
----------
dim : str
Dimension along which to drop missing values. Dropping along
multiple dimensions simultaneously is not yet supported.
how : {'any', 'all'}, optional
* any : if any NA values are present, drop that label
* all : if all values are NA, drop that label
thresh : int, default None
If supplied, require this many non-NA values.
Returns
-------
DataArray
)�how�thresh)rj �dropnarn )r< r3 r� r� r� r r r r� � s zDataArray.dropnac C s t j|�rtd��| j|�S )a} Fill missing values in this object.
This operation follows the normal broadcasting and alignment rules that
xarray uses for binary arithmetic, except the result is aligned to this
object (``join='left'``) instead of aligned to the intersection of
index coordinates (``join='inner'``).
Parameters
----------
value : scalar, ndarray or DataArray
Used to fill all matching missing values in this array. If the
argument is a DataArray, it is first aligned with (reindexed to)
this array.
Returns
-------
DataArray
zDcannot provide fill value as a dictionary with fillna on a DataArray)r r$ r. Z_fillna)r< rG r r r �fillna� s
zDataArray.fillnac K s | j j||||f|�}| j|�S )a Reduce this array by applying `func` along some dimension(s).
Parameters
----------
func : function
Function which can be called in the form
`f(x, axis=axis, **kwargs)` to return the result of reducing an
np.ndarray over an integer valued axis.
dim : str or sequence of str, optional
Dimension(s) over which to apply `func`.
axis : int or sequence of int, optional
Axis(es) over which to repeatedly apply `func`. Only one of the
'dim' and 'axis' arguments can be supplied. If neither are
supplied, then the reduction is calculated over the flattened array
(by calling `f(x)` without an axis argument).
keep_attrs : bool, optional
If True, the variable's attributes (`attrs`) will be copied from
the original object to the new one. If False (default), the new
object will be returned without attributes.
**kwargs : dict
Additional keyword arguments passed on to `func`.
Returns
-------
reduced : DataArray
DataArray with this object's array replaced with an array with
summarized data and the indicated dimension(s) removed.
)r` �reducerf )r< �funcr3 ZaxisZ
keep_attrs�kwargsr7 r r r r� � s zDataArray.reducec C s` dd� t jt jt jd�}y|| j }W n" tk
rH td| j ��Y nX || jf| jj� �� S )a� Convert this array into a pandas object with the same shape.
The type of the returned object depends on the number of DataArray
dimensions:
* 1D -> `pandas.Series`
* 2D -> `pandas.DataFrame`
* 3D -> `pandas.Panel`
Only works for arrays with 3 or fewer dimensions.
The DataArray constructor performs the inverse transformation.
c S s | S )Nr )�xr r r �<lambda>� s z%DataArray.to_pandas.<locals>.<lambda>)r r r � z<cannot convert arrays with %s dimensions into pandas objects) rU rV rW rZ rD r� r& r r> )r< Z constructors� constructorr r r � to_pandas� s
zDataArray.to_pandasc sf � dkr| j � � dkr td��tt| j| j��}d�| j�d�}|j|�}� �fdd�|jD �|_|S )a Convert this array and its coordinates into a tidy pandas.DataFrame.
The DataFrame is indexed by the Cartesian product of index coordinates
(in the form of a :py:class:`pandas.MultiIndex`).
Other coordinates are included as columns in the DataFrame.
NzNcannot convert an unnamed DataArray to a DataFrame: use the ``name`` parameterZ,__unique_name_identifier_z98xfz98xugfg73ho__)r, c s g | ]}|�kr� n|�qS r r )r �c)r, � unique_namer r r s z*DataArray.to_dataframe.<locals>.<listcomp>) r, r& r r r" r1 rh Z
_to_dataframerX )r< r, r" r� �dfr )r, r� r � to_dataframe� s
zDataArray.to_dataframec C s$ | j j� }tj| jjd�|| jd�S )z�Convert this array into a pandas.Series.
The Series is indexed by the Cartesian product of index coordinates
(in the form of a :py:class:`pandas.MultiIndex`).
r )r@ r, ���)r2 r� rU rV r Zreshaper, )r< r@ r r r � to_series s
zDataArray.to_seriesc C s t j| j�}tjj| j||d�S )a| Convert this array into a numpy.ma.MaskedArray
Parameters
----------
copy : bool
If True (default) make a copy of the array in the result. If False,
a MaskedArray view of DataArray.values is returned.
Returns
-------
result : MaskedArray
Masked where invalid values (nan or inf) occur.
)r_ �maskr )rU �isnullr �np�maZ MaskedArray)r< r r� r r r �to_masked_array s zDataArray.to_masked_arrayc C s& |j }tj||i�}tj|�}|| S )a6 Convert a pandas.Series into an xarray.DataArray.
If the series's index is a MultiIndex, it will be expanded into a
tensor product of one-dimensional coordinates (filling in missing values
with NaN). Thus this operation should be the inverse of the `to_series`
method.
)r, rU rW r Zfrom_dataframe)�clsZseriesr, r� r� r r r � from_series) s
zDataArray.from_seriesc C s ddl m} || �S )z1Convert this array into a cdms2.Variable
r )�to_cdms2)�convertr� )r< r� r r r r� 8 s zDataArray.to_cdms2c C s ddl m} ||�S )z:Convert a cdms2.Variable into an xarray.DataArray
r )�
from_cdms2)r� r� )r� r` r� r r r r� > s zDataArray.from_cdms2c s* � fdd�}t j| j|j|d�o(|| |�S )z(Helper function for equals and identicalc s t | j� �|j�S )N)rT r` )r� �y)�
compat_strr r �compatG s z%DataArray._all_compat.<locals>.compat)r� )r Z
dict_equivr2 )r< r� r� r� r )r� r � _all_compatE s zDataArray._all_compatc C s* y | j |d�S ttfk
r$ dS X dS )z�Two DataArrays are broadcast equal if they are equal after
broadcasting them against each other such that they have the same
dimensions.
See Also
--------
DataArray.equals
DataArray.identical
�broadcast_equalsFN)r� r. r� )r< r� r r r r� L s
zDataArray.broadcast_equalsc C s* y | j |d�S ttfk
r$ dS X dS )a� True if two DataArrays have the same dimensions, coordinates and
values; otherwise False.
DataArrays can still be equal (like pandas objects) if they have NaN
values in the same locations.
This method is necessary because `v1 == v2` for ``DataArray``
does element-wise comparisions (like numpy.ndarrays).
See Also
--------
DataArray.broadcast_equals
DataArray.identical
�equalsFN)r� r. r� )r< r� r r r r� [ s zDataArray.equalsc C s6 y| j |j ko| j|d�S ttfk
r0 dS X dS )z�Like equals, but also checks the array name and attributes, and
attributes on all coordinates.
See Also
--------
DataArray.broadcast_equals
DataArray.equal
� identicalFN)r, r� r. r� )r< r� r r r r� o s
zDataArray.identicalc C s0 t |d| j�}|| jks"|| jkr(| jS d S d S )Nr, )rT �_DataArray__default_namer, )r< r� Z
other_namer r r � _result_name� s zDataArray._result_namec C s | j j||�}| j|�S )N)r` �__array_wrap__rc )r< �obj�contextZnew_varr r r r� � s zDataArray.__array_wrap__c s t j� �� fdd��}|S )Nc s | j � | jjf|�|��S )N)r� r` r_ )r< �argsr� )�fr r r� � s z!DataArray._unary_op.<locals>.func)� functools�wraps)r� r� r )r� r � _unary_op� s zDataArray._unary_op�innerc s t j� �� ��fdd��}|S )Nc s� t |ttjf�rtS t|d�rZt| |�dd�\} }dd� t| j| j �D �}|rZt
d| ��t |d|�}t |dd �}�s�� | j |�n
� || j �}| j
j|�j}| j|�}| j|||�S ) Nr> F)�joinr c S s g | ]\}}|d kr|�qS )r r )r r! �sr r r r � s z6DataArray._binary_op.<locals>.func.<locals>.<listcomp>z-no overlapping labels for some dimensions: %sr` r2 )r' r r �GroupBy�NotImplemented�hasattrr r r" r1 r& rT r` r2 �mergerk r� rc )r< r� Z
empty_indexes�other_variable� other_coordsr` r2 r, )r� r� � reflexiver r r� � s
z"DataArray._binary_op.<locals>.func)r� r� )r� r� r� Zignored_kwargsr� r )r� r� r� r �
_binary_op� s zDataArray._binary_opc s t j� �� fdd��}|S )Nc
sT t |tj�rtd��t|dd �}t|d|�}| jj|�� � | j|� W d Q R X | S )NzNin-place operations between a DataArray and a grouped object are not permittedr2 r` )r' r r� r. rT r2 Z_merge_inplacer` )r< r� r� r� )r� r r r� � s z*DataArray._inplace_binary_op.<locals>.func)r� r� )r� r� r )r� r �_inplace_binary_op� s
z DataArray._inplace_binary_opc C s t | �S )a>
Access plotting functions
>>> d = DataArray([[1, 2], [3, 4]])
For convenience just call this directly
>>> d.plot()
Or use it as a namespace to use xarray.plot functions as
DataArray methods
>>> d.plot.imshow() # equivalent to xarray.plot.imshow(d)
)r )r< r r r �plot� s zDataArray.plot�2 c C sn g }x:t | j�D ],\}}|jdkr|jdj|t|j�d�� qW dj|�}t|�|krj|d|d � d }|S )aW
If the dataarray has 1 dimensional coordinates or comes from a slice
we can show that info in the title
Parameters
----------
truncate : integer
maximum number of characters for title
Returns
-------
title : string
Can be used for plot titles
r z {dim} = {v})r3 r6 z, Nr� z...) r r2 r8 �append�formatr r r� r% )r< �truncateZone_dimsr3 r4 �titler r r �_title_for_slice� s
zDataArray._title_for_slicer �upperc C s | j � j|||d�}| j|�S )a6 Calculate the n-th order discrete difference along given axis.
Parameters
----------
dim : str, optional
Dimension over which to calculate the finite difference.
n : int, optional
The number of times values are differenced.
label : str, optional
The new coordinate in dimension ``dim`` will have the
values of either the minuend's or subtrahend's coordinate
for values 'upper' and 'lower', respectively. Other
values are not supported.
Returns
-------
difference : same type as caller
The n-th order finite differnce of this object.
Examples
--------
>>> arr = xr.DataArray([5, 5, 6, 6], [[1, 2, 3, 4]], ['x'])
>>> arr.diff('x')
<xarray.DataArray (x: 3)>
array([0, 1, 0])
Coordinates:
* x (x) int64 2 3 4
>>> arr.diff('x', 2)
<xarray.DataArray (x: 2)>
array([ 1, -1])
Coordinates:
* x (x) int64 3 4
)r r3 rq )rj �diffrn )r< r3 r rq r� r r r r� � s #zDataArray.diffc K s | j jf |�}| j|�S )a� Shift this array by an offset along one or more dimensions.
Only the data is moved; coordinates stay in place. Values shifted from
beyond array bounds are replaced by NaN. This is consistent with the
behavior of ``shift`` in pandas.
Parameters
----------
**shifts : keyword arguments of the form {dim: offset}
Integer offset to shift along each of the given dimensions.
Positive offsets shift to the right; negative offsets shift to the
left.
Returns
-------
shifted : DataArray
DataArray with the same coordinates and attributes but shifted
data.
See also
--------
roll
Examples
--------
>>> arr = xr.DataArray([5, 6, 7], dims='x')
>>> arr.shift(x=1)
<xarray.DataArray (x: 3)>
array([ nan, 5., 6.])
Coordinates:
* x (x) int64 0 1 2
)r` �shiftrc )r< �shiftsr` r r r r� s "zDataArray.shiftc K s | j � jf |�}| j|�S )ay Roll this array by an offset along one or more dimensions.
Unlike shift, roll rotates all variables, including coordinates. The
direction of rotation is consistent with :py:func:`numpy.roll`.
Parameters
----------
**shifts : keyword arguments of the form {dim: offset}
Integer offset to rotate each of the given dimensions. Positive
offsets roll to the right; negative offsets roll to the left.
Returns
-------
rolled : DataArray
DataArray with the same attributes but rolled data and coordinates.
See also
--------
shift
Examples
--------
>>> arr = xr.DataArray([5, 6, 7], dims='x')
>>> arr.roll(x=1)
<xarray.DataArray (x: 3)>
array([7, 5, 6])
Coordinates:
* x (x) int64 2 0 1
)rj �rollrn )r< r� r� r r r r� 3 s zDataArray.rollc C s | j | jj�S )N)rc r` �real)r< r r r r� U s zDataArray.realc C s | j | jj�S )N)rc r` �imag)r< r r r r� Y s zDataArray.imag)NNNNNF)NT)NN)NFF)T)N)N)NN)r� )r� NN)NNT)NNT)N)N)r0 N)NNF)N)T)N)N)Fr� )r� )r r� )`rI rJ rK rN r ZDataArrayGroupByZ groupby_clsr= �objectra rc rf rL ri rj rn rw rh ru �propertyr, �setterr` r| r1 r8 r} rD r~ r_ r r� r� r" r� rF rH r� r� r� ro rP rQ r> r2 r� r� r� r r� r� �__hash__r� r� r� r� r� r� r� r� r� r� r� r� r� r rp r� r� r� r� r� r� r� � classmethodr� r� r� r� r� r� r� r� r� r� � staticmethodr� r� r� r� r� r� r� r� r� r� r r r r rO m s� %
D
%
/
, +
&%" rO �< )�priority),�
contextlibr� ry Znumpyr� ZpandasrU Z plot.plotr � r r r r � alignmentr �commonr
r r Z coordinatesr
r rm r Zpycompatr r r r r` r r r r r Z
formattingr r9 r� r: rL rO Z!inject_all_ops_and_reduce_methodsr r r r �<module> s< 4 z |