{"diffoscope-json-version": 1, "source1": "/srv/reproducible-results/rbuild-debian/r-b-build.zS6Y25Wu/b1/python-xarray_2025.03.1-8_arm64.changes", "source2": "/srv/reproducible-results/rbuild-debian/r-b-build.zS6Y25Wu/b2/python-xarray_2025.03.1-8_arm64.changes", "unified_diff": null, "details": [{"source1": "Files", "source2": "Files", "unified_diff": "@@ -1,3 +1,3 @@\n \n- 6590e92d3848374daeb0616d1d8247fb 5274924 doc optional python-xarray-doc_2025.03.1-8_all.deb\n+ 6de6a234e82666735ae0ed6abfdbee83 5274668 doc optional python-xarray-doc_2025.03.1-8_all.deb\n 290b2f0830b312db0f155225dfe879d5 820184 python optional python3-xarray_2025.03.1-8_all.deb\n"}, {"source1": "python-xarray-doc_2025.03.1-8_all.deb", "source2": "python-xarray-doc_2025.03.1-8_all.deb", "unified_diff": null, "details": [{"source1": "file list", "source2": "file list", "unified_diff": "@@ -1,3 +1,3 @@\n -rw-r--r-- 0 0 0 4 2025-05-05 09:04:03.000000 debian-binary\n--rw-r--r-- 0 0 0 7556 2025-05-05 09:04:03.000000 control.tar.xz\n--rw-r--r-- 0 0 0 5267176 2025-05-05 09:04:03.000000 data.tar.xz\n+-rw-r--r-- 0 0 0 7560 2025-05-05 09:04:03.000000 control.tar.xz\n+-rw-r--r-- 0 0 0 5266916 2025-05-05 09:04:03.000000 data.tar.xz\n"}, {"source1": "control.tar.xz", "source2": "control.tar.xz", "unified_diff": null, "details": [{"source1": "control.tar", "source2": "control.tar", "unified_diff": null, "details": [{"source1": "./control", "source2": "./control", "unified_diff": "@@ -1,13 +1,13 @@\n Package: python-xarray-doc\n Source: python-xarray\n Version: 2025.03.1-8\n Architecture: all\n Maintainer: Debian Science Maintainers Let\u2019s create a simple plot of 2-m air temperature in degrees Celsius: Write equations to calculate the vertical coordinate. These will be only evaluated when data is requested. Information about the ROMS vertical coordinate can be found (here)[https://www.myroms.org/wiki/Vertical_S-coordinate] In short, for The function we will apply is Plot the first timestep: We first have to come up with the weights, - calculate the month length for each monthly data record - calculate weights using Finally, we just need to multiply our weights by the In this example, the logical coordinates are Control the map projection parameters on multiple axes This example illustrates how to plot multiple maps and control their extent and aspect ratio. For more details see this discussion on github. Visualizing your datasets is quick and convenient: Note the automatic labeling with names and units. Our effort in adding metadata attributes has paid off! Many aspects of these figures are customizable: see Plotting. Note This method replicates the behavior of [3]:\n
\n-Error in callback <function _draw_all_if_interactive at 0xffff657b4040> (for post_execute), with arguments args (),kwargs {}:\n+Error in callback <function _draw_all_if_interactive at 0xffff711b0040> (for post_execute), with arguments args (),kwargs {}:\n
\n", "details": [{"source1": "html2text {}", "source2": "html2text {}", "unified_diff": "@@ -93,15 +93,15 @@\n File /usr/lib/python3/dist-packages/urllib3/connection.py:205, in\n HTTPConnection._new_conn(self)\n 204 except socket.gaierror as e:\n --> 205 raise NameResolutionError(self.host, self, e) from e\n 206 except SocketTimeout as e:\n \n NameResolutionError:
Add a lazilly calculated vertical coordinates\u00b6
\n Vtransform==2
as used in this example,np.interp
which expects 1D numpy arrays. This functionality is already implemented in xarray so we use that capability to make sure we are not making mistakes.[2]:\n
[3]:\n
[ ]:\n
\n", "details": [{"source1": "html2text {}", "source2": "html2text {}", "unified_diff": "@@ -92,15 +92,15 @@\n File /usr/lib/python3/dist-packages/urllib3/connection.py:205, in\n HTTPConnection._new_conn(self)\n 204 except socket.gaierror as e:\n --> 205 raise NameResolutionError(self.host, self, e) from e\n 206 except SocketTimeout as e:\n \n NameResolutionError:
Now for the heavy lifting:\u00b6
\n groupby('time.season')
Dataset
and sum along the time dimension. Creating a DataArray
for the month length is as easy as using the days_in_month
accessor on the time coordinate. The calendar type, in this case 'noleap'
, is automatically considered in this operation.x
and y
, while the physical coordinates are xc
and yc
, which represent the longitudes and latitudes of the data.[3]:\n
Multiple plots and map projections\u00b6
\n <xarray.Dataset> Size: 41kB\n Dimensions: (time: 731, location: 3)\n Coordinates:\n * time (time) datetime64[ns] 6kB 2000-01-01 2000-01-02 ... 2001-12-31\n * location (location) <U2 24B 'IA' 'IN' 'IL'\n Data variables:\n tmin (time, location) float64 18kB -8.037 -1.788 ... -1.346 -4.544\n- tmax (time, location) float64 18kB 12.98 3.31 6.779 ... 3.343 3.805
PandasIndex(Index(['IA', 'IN', 'IL'], dtype='object', name='location'))
Examine a dataset with pandas and seaborn\u00b6
\n Convert to a pandas DataFrame\u00b6
\n [2]:\n@@ -697,15 +697,15 @@\n
[5]:\n
\n-<seaborn.axisgrid.PairGrid at 0xffff37afaba0>\n+<seaborn.axisgrid.PairGrid at 0xffff4597aba0>\n
\n@@ -1110,26 +1110,26 @@\n [0. , 0. , 0. ],\n [0. , 0. , 0. ],\n [0. , 0.01612903, 0. ],\n [0.33333333, 0.35 , 0.23333333],\n [0.93548387, 0.85483871, 0.82258065]])\n Coordinates:\n * location (location) <U2 24B 'IA' 'IN' 'IL'\n- * month (month) int64 96B 1 2 3 4 5 6 7 8 9 10 11 12
array(['IA', 'IN', 'IL'], dtype='<U2')
array([ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12])
PandasIndex(Index(['IA', 'IN', 'IL'], dtype='object', name='location'))
PandasIndex(Index([1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], dtype='int64', name='month'))
[7]:\n
freeze.to_pandas().plot()\n
PandasIndex(Index(['IA', 'IN', 'IL'], dtype='object', name='location'))
[12]:\n
df = both.sel(time="2000").mean("location").reset_coords(drop=True).to_dataframe()\n df.head()\n", "details": [{"source1": "html2text {}", "source2": "html2text {}", "unified_diff": "@@ -142,15 +142,15 @@\n [4]:\n
array(['IA', 'IN', 'IL'], dtype='<U2')
array([ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12])
PandasIndex(Index(['IA', 'IN', 'IL'], dtype='object', name='location'))
PandasIndex(Index([1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12], dtype='int64', name='month'))
Plotting\u00b6
\n In [37]: data.plot()\n-Out[37]: <matplotlib.collections.QuadMesh at 0xffff3e2da7b0>\n+Out[37]: <matplotlib.collections.QuadMesh at 0xffff3ccba7b0>\n
\n
pandas\u00b6
\n", "details": [{"source1": "html2text {}", "source2": "html2text {}", "unified_diff": "@@ -253,15 +253,15 @@\n [0.37342613, 1.49497537, 1.33584385]])\n Coordinates:\n * x (x) int64 16B 10 20\n Dimensions without coordinates: y\n *\b**\b**\b**\b**\b* P\bPl\blo\bot\btt\bti\bin\bng\bg_\b?\b\u00b6 *\b**\b**\b**\b**\b*\n Visualizing your datasets is quick and convenient:\n In [37]: data.plot()\n-Out[37]: apply_ufunc
\", \"Compare weighted and unweighted mean temperature\", \"Blank template\", \"Calculating Seasonal Averages from Time Series of Monthly Means\", \"Working with Multidimensional Coordinates\", \"Visualization Gallery\", \"Toy weather data\", \"Gallery\", \"Frequently Asked Questions\", \"Getting Started\", \"Installation\", \"Quick overview\", \"Overview: Why xarray?\", \"Getting Help\", \"How do I \\u2026\", \"Xarray documentation\", \"Alternative chunked array types\", \"Integrating with duck arrays\", \"Extending xarray using accessors\", \"How to add a new backend\", \"How to create a custom index\", \"Xarray Internals\", \"Internal Design\", \"Interoperability of Xarray\", \"Time Coding\", \"Zarr Encoding Specification\", \"Development roadmap\", \"Tutorials and Videos\", \"Combining data\", \"Computation\", \"Parallel Computing with Dask\", \"Data Structures\", \"Working with numpy-like arrays\", \"GroupBy: Group and Bin Data\", \"Hierarchical data\", \"User Guide\", \"Indexing and selecting data\", \"Interpolating data\", \"Reading and writing files\", \"Configuration\", \"Working with pandas\", \"Plotting\", \"Reshaping and reorganizing data\", \"Terminology\", \"Testing your code\", \"Time series data\", \"Weather and climate data\", \"What\\u2019s New\"],\n \"titleterms\": {\n \"\": [13, 16, 55],\n \"0\": 55,\n \"01\": 55,\n \"02\": 55,\n"}]}, {"source1": "./usr/share/doc/python-xarray-doc/html/user-guide/computation.html", "source2": "./usr/share/doc/python-xarray-doc/html/user-guide/computation.html", "unified_diff": "@@ -934,16 +934,16 @@\n <xarray.Dataset> Size: 2kB\n Dimensions: (param: 10, cov_i: 10, cov_j: 10)\n Coordinates:\n * param (param) <U7 280B 'a0' 'xc0' ... 'xalpha1' 'yalpha1'\n * cov_i (cov_i) <U7 280B 'a0' 'xc0' ... 'xalpha1' 'yalpha1'\n * cov_j (cov_j) <U7 280B 'a0' 'xc0' ... 'xalpha1' 'yalpha1'\n Data variables:\n- curvefit_coefficients (param) float64 80B -0.659 4.858 ... 2.066 1.329\n- curvefit_covariance (cov_i, cov_j) float64 800B 5.662e+11 ... 6.911e-05\n+ curvefit_coefficients (param) float64 80B 3.0 1.004 1.003 ... 1.007 1.008\n+ curvefit_covariance (cov_i, cov_j) float64 800B 3.362e-05 ... 2.125e-05\n \n \n scipy.optimize.curve_fit()
.func(ds)
). This allows you to write pipelines for\n transforming your data (using \u201cmethod chaining\u201d) instead of writing hard to\n follow nested function calls:
# these lines are equivalent, but with pipe we can make the logic flow\n # entirely from left to right\n In [64]: plt.plot((2 * ds.temperature.sel(loc=0)).mean("instrument"))\n-Out[64]: [<matplotlib.lines.Line2D at 0xffff36c3afd0>]\n+Out[64]: [<matplotlib.lines.Line2D at 0xffff315e6fd0>]\n \n In [65]: (ds.temperature.sel(loc=0).pipe(lambda x: 2 * x).mean("instrument").pipe(plt.plot))\n-Out[65]: [<matplotlib.lines.Line2D at 0xffff36c3ad50>]\n+Out[65]: [<matplotlib.lines.Line2D at 0xffff315e6d50>]\n
Both pipe
and assign
replicate the pandas methods of the same names\n (DataFrame.pipe
and\n DataFrame.assign
).
With xarray, there is no performance penalty for creating new datasets, even if\n variables are lazily loaded from a file on disk. Creating new objects instead\n", "details": [{"source1": "html2text {}", "source2": "html2text {}", "unified_diff": "@@ -585,19 +585,19 @@\n There is also the pipe() method that allows you to use a method call with an\n external function (e.g., ds.pipe(func)) instead of simply calling it (e.g.,\n func(ds)). This allows you to write pipelines for transforming your data (using\n \u201cmethod chaining\u201d) instead of writing hard to follow nested function calls:\n # these lines are equivalent, but with pipe we can make the logic flow\n # entirely from left to right\n In [64]: plt.plot((2 * ds.temperature.sel(loc=0)).mean(\"instrument\"))\n-Out[64]: [ If you were a previous user of the prototype xarray-contrib/datatree package, this is different from what you\u2019re used to!\n In that package the data model was that the data stored in each node actually was completely unrelated. The data model is now slightly stricter.\n This allows us to provide features like Coordinate Inheritance. To demonstrate, let\u2019s first generate some example datasets which are not aligned with one another: Now we have a valid This is a useful way to organise our data because we can still operate on all the groups at once.\n For example we can extract all three timeseries at a specific lat-lon location: or compute the standard deviation of each timeseries to find out how it varies with sampling frequency: This helps to differentiate which variables are defined on the datatree node that you are currently looking at, and which were defined somewhere above it. We can also still perform all the same operations on the whole tree:# (drop the attributes just to make the printed representation shorter)\n In [89]: ds = xr.tutorial.open_dataset("air_temperature").drop_attrs()\n-ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff35702710>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n+ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff300b2710>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n \n \n In [90]: ds_daily = ds.resample(time="D").mean("time")\n KeyError: "No variable named 'time'. Variables on the dataset include ['foo', 'x', 'letters']"\n \n \n In [91]: ds_weekly = ds.resample(time="W").mean("time")\n@@ -1055,15 +1055,15 @@\n \u2514\u2500\u2500 Group: /b/B\n
DataTree
structure which contains all the data at each different time frequency, stored in a separate group.In [100]: dt.sel(lat=75, lon=300)\n-ValueError: Dimensions {'lat', 'lon'} do not exist. Expected one or more of set()\n+ValueError: Dimensions {'lon', 'lat'} do not exist. Expected one or more of set()\n
In [101]: dt.std(dim="time")\n ValueError: Dimension(s) 'time' do not exist. Expected one or more of set()\n
In [107]: print(dt["/daily"])\n KeyError: 'Could not find node at /daily'\n
In [108]: dt.sel(lat=[75], lon=[300])\n-ValueError: Dimensions {'lat', 'lon'} do not exist. Expected one or more of set()\n+ValueError: Dimensions {'lon', 'lat'} do not exist. Expected one or more of set()\n \n \n In [109]: dt.std(dim="time")\n ValueError: Dimension(s) 'time' do not exist. Expected one or more of set()\n
In [52]: ds = xr.tutorial.open_dataset("air_temperature")\n-ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff35702710>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n+ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff300b2710>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n \n \n # Define target latitude and longitude (where weather stations might be)\n In [53]: target_lon = xr.DataArray([200, 201, 202, 205], dims="points")\n \n In [54]: target_lat = xr.DataArray([31, 41, 42, 42], dims="points")\n \n@@ -697,15 +697,15 @@\n
To select and assign values to a portion of a DataArray()
you\n can use indexing with .loc
:
In [57]: ds = xr.tutorial.open_dataset("air_temperature")\n-ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff35701810>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n+ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff300b1810>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n \n \n # add an empty 2D dataarray\n In [58]: ds["empty"] = xr.full_like(ds.air.mean("time"), fill_value=0)\n AttributeError: 'Dataset' object has no attribute 'air'\n \n \n@@ -869,15 +869,15 @@\n
You can also assign values to all variables of a Dataset
at once:
In [83]: ds_org = xr.tutorial.open_dataset("eraint_uvz").isel(\n ....: latitude=slice(56, 59), longitude=slice(255, 258), level=0\n ....: )\n ....: \n-ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/eraint_uvz.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff357011d0>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n+ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/eraint_uvz.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff300b11d0>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n \n \n # set all values to 0\n In [84]: ds = xr.zeros_like(ds_org)\n NameError: name 'ds_org' is not defined\n \n \n", "details": [{"source1": "html2text {}", "source2": "html2text {}", "unified_diff": "@@ -476,15 +476,15 @@\n with a new shared output dimension name. In the example below, the selections\n of the closest latitude and longitude are renamed to an output dimension named\n \u201cpoints\u201d:\n In [52]: ds = xr.tutorial.open_dataset(\"air_temperature\")\n ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries\n exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by\n NameResolutionError(\": Failed to resolve 'github.com' ([Errno -3] Temporary failure\n+0xffff300b2710>: Failed to resolve 'github.com' ([Errno -3] Temporary failure\n in name resolution)\"))\n \n \n # Define target latitude and longitude (where weather stations might be)\n In [53]: target_lon = xr.DataArray([200, 201, 202, 205], dims=\"points\")\n \n In [54]: target_lat = xr.DataArray([31, 41, 42, 42], dims=\"points\")\n@@ -516,15 +516,15 @@\n *\b**\b**\b**\b**\b* A\bAs\bss\bsi\big\bgn\bni\bin\bng\bg v\bva\bal\blu\bue\bes\bs w\bwi\bit\bth\bh i\bin\bnd\bde\bex\bxi\bin\bng\bg_\b?\b\u00b6 *\b**\b**\b**\b**\b*\n To select and assign values to a portion of a DataArray() you can use indexing\n with .loc :\n In [57]: ds = xr.tutorial.open_dataset(\"air_temperature\")\n ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries\n exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by\n NameResolutionError(\": Failed to resolve 'github.com' ([Errno -3] Temporary failure\n+0xffff300b1810>: Failed to resolve 'github.com' ([Errno -3] Temporary failure\n in name resolution)\"))\n \n \n # add an empty 2D dataarray\n In [58]: ds[\"empty\"] = xr.full_like(ds.air.mean(\"time\"), fill_value=0)\n AttributeError: 'Dataset' object has no attribute 'air'\n \n@@ -678,15 +678,15 @@\n In [83]: ds_org = xr.tutorial.open_dataset(\"eraint_uvz\").isel(\n ....: latitude=slice(56, 59), longitude=slice(255, 258), level=0\n ....: )\n ....:\n ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries\n exceeded with url: /pydata/xarray-data/raw/master/eraint_uvz.nc (Caused by\n NameResolutionError(\": Failed to resolve 'github.com' ([Errno -3] Temporary failure\n+0xffff300b11d0>: Failed to resolve 'github.com' ([Errno -3] Temporary failure\n in name resolution)\"))\n \n \n # set all values to 0\n In [84]: ds = xr.zeros_like(ds_org)\n NameError: name 'ds_org' is not defined\n \n"}]}, {"source1": "./usr/share/doc/python-xarray-doc/html/user-guide/interpolation.html", "source2": "./usr/share/doc/python-xarray-doc/html/user-guide/interpolation.html", "unified_diff": "@@ -237,24 +237,24 @@\n ....: np.sin(np.linspace(0, 2 * np.pi, 10)),\n ....: dims="x",\n ....: coords={"x": np.linspace(0, 1, 10)},\n ....: )\n ....: \n \n In [17]: da.plot.line("o", label="original")\n-Out[17]: [<matplotlib.lines.Line2D at 0xffff601d9810>]\n+Out[17]: [<matplotlib.lines.Line2D at 0xffff56905810>]\n \n In [18]: da.interp(x=np.linspace(0, 1, 100)).plot.line(label="linear (default)")\n-Out[18]: [<matplotlib.lines.Line2D at 0xffff601d9450>]\n+Out[18]: [<matplotlib.lines.Line2D at 0xffff56905450>]\n \n In [19]: da.interp(x=np.linspace(0, 1, 100), method="cubic").plot.line(label="cubic")\n-Out[19]: [<matplotlib.lines.Line2D at 0xffff36c3ad50>]\n+Out[19]: [<matplotlib.lines.Line2D at 0xffff315e6d50>]\n \n In [20]: plt.legend()\n-Out[20]: <matplotlib.legend.Legend at 0xffff3e2db4d0>\n+Out[20]: <matplotlib.legend.Legend at 0xffff3ccbb4d0>\n
Additional keyword arguments can be passed to scipy\u2019s functions.
\n# fill 0 for the outside of the original coordinates.\n In [21]: da.interp(x=np.linspace(-0.5, 1.5, 10), kwargs={"fill_value": 0.0})\n@@ -439,15 +439,15 @@\n see Missing values.\n \n \n Example\u00b6
\n Let\u2019s see how interp()
works on real data.
\n # Raw data\n In [44]: ds = xr.tutorial.open_dataset("air_temperature").isel(time=0)\n-ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff601db890>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n+ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff56907890>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n \n \n In [45]: fig, axes = plt.subplots(ncols=2, figsize=(10, 4))\n \n In [46]: ds.air.plot(ax=axes[0])\n AttributeError: 'Dataset' object has no attribute 'air'\n \n", "details": [{"source1": "html2text {}", "source2": "html2text {}", "unified_diff": "@@ -153,26 +153,26 @@\n ....: np.sin(np.linspace(0, 2 * np.pi, 10)),\n ....: dims=\"x\",\n ....: coords={\"x\": np.linspace(0, 1, 10)},\n ....: )\n ....:\n \n In [17]: da.plot.line(\"o\", label=\"original\")\n-Out[17]: []\n+Out[17]: []\n \n In [18]: da.interp(x=np.linspace(0, 1, 100)).plot.line(label=\"linear\n (default)\")\n-Out[18]: []\n+Out[18]: []\n \n In [19]: da.interp(x=np.linspace(0, 1, 100), method=\"cubic\").plot.line\n (label=\"cubic\")\n-Out[19]: []\n+Out[19]: []\n \n In [20]: plt.legend()\n-Out[20]: \n+Out[20]: \n _\b[_\b._\b._\b/_\b__\bi_\bm_\ba_\bg_\be_\bs_\b/_\bi_\bn_\bt_\be_\br_\bp_\bo_\bl_\ba_\bt_\bi_\bo_\bn_\b__\bs_\ba_\bm_\bp_\bl_\be_\b1_\b._\bp_\bn_\bg_\b]\n Additional keyword arguments can be passed to scipy\u2019s functions.\n # fill 0 for the outside of the original coordinates.\n In [21]: da.interp(x=np.linspace(-0.5, 1.5, 10), kwargs={\"fill_value\": 0.0})\n Out[21]:\n Size: 80B\n array([ 0. , 0. , 0. , 0.814, 0.604, -0.604, -0.814, 0. , 0. ,\n@@ -338,15 +338,15 @@\n *\b**\b**\b**\b**\b* E\bEx\bxa\bam\bmp\bpl\ble\be_\b?\b\u00b6 *\b**\b**\b**\b**\b*\n Let\u2019s see how interp() works on real data.\n # Raw data\n In [44]: ds = xr.tutorial.open_dataset(\"air_temperature\").isel(time=0)\n ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries\n exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by\n NameResolutionError(\": Failed to resolve 'github.com' ([Errno -3] Temporary failure\n+0xffff56907890>: Failed to resolve 'github.com' ([Errno -3] Temporary failure\n in name resolution)\"))\n \n \n In [45]: fig, axes = plt.subplots(ncols=2, figsize=(10, 4))\n \n In [46]: ds.air.plot(ax=axes[0])\n AttributeError: 'Dataset' object has no attribute 'air'\n"}]}, {"source1": "./usr/share/doc/python-xarray-doc/html/user-guide/io.html", "source2": "./usr/share/doc/python-xarray-doc/html/user-guide/io.html", "unified_diff": "@@ -630,15 +630,15 @@\n ....: "y": pd.date_range("2000-01-01", periods=5),\n ....: "z": ("x", list("abcd")),\n ....: },\n ....: )\n ....: \n \n In [13]: ds.to_zarr("path/to/directory.zarr")\n-Out[13]: <xarray.backends.zarr.ZarrStore at 0xffff35b55c60>\n+Out[13]: <xarray.backends.zarr.ZarrStore at 0xffff30505c60>\n
\n \n (The suffix .zarr
is optional\u2013just a reminder that a zarr store lives\n there.) If the directory does not exist, it will be created. If a zarr\n store is already present at that path, an error will be raised, preventing it\n from being overwritten. To override this behavior and overwrite an existing\n store, add mode='w'
when invoking to_zarr()
.
\n@@ -660,17 +660,17 @@\n In [14]: ds_zarr = xr.open_zarr("path/to/directory.zarr")\n \n In [15]: ds_zarr\n Out[15]: \n <xarray.Dataset> Size: 264B\n Dimensions: (x: 4, y: 5)\n Coordinates:\n+ * x (x) int64 32B 10 20 30 40\n z (x) object 32B dask.array<chunksize=(4,), meta=np.ndarray>\n * y (y) datetime64[ns] 40B 2000-01-01 2000-01-02 ... 2000-01-05\n- * x (x) int64 32B 10 20 30 40\n Data variables:\n foo (x, y) float64 160B dask.array<chunksize=(4, 5), meta=np.ndarray>\n
\n \n \n Cloud Storage Buckets\u00b6
\n It is possible to read and write xarray datasets directly from / to cloud\n@@ -724,36 +724,36 @@\n \n In [18]: ds = xr.Dataset({"foo": ("x", dummies)}, coords={"x": np.arange(30)})\n \n In [19]: path = "path/to/directory.zarr"\n \n # Now we write the metadata without computing any array values\n In [20]: ds.to_zarr(path, compute=False)\n-Out[20]: Delayed('_finalize_store-9c2dca6b-5842-4f3d-84bb-0aa396adbcad')\n+Out[20]: Delayed('_finalize_store-21a4d77a-7afa-456e-972d-01a3d5cfaa3b')\n
Now, a Zarr store with the correct variable shapes and attributes exists that\n can be filled out by subsequent calls to to_zarr
.\n Setting region="auto"
will open the existing store and determine the\n correct alignment of the new data with the existing dimensions, or as an\n explicit mapping from dimension names to Python slice
objects indicating\n where the data should be written (in index space, not label space), e.g.,
# For convenience, we'll slice a single dataset, but in the real use-case\n # we would create them separately possibly even from separate processes.\n In [21]: ds = xr.Dataset({"foo": ("x", np.arange(30))}, coords={"x": np.arange(30)})\n \n # Any of the following region specifications are valid\n In [22]: ds.isel(x=slice(0, 10)).to_zarr(path, region="auto")\n-Out[22]: <xarray.backends.zarr.ZarrStore at 0xffff35b57010>\n+Out[22]: <xarray.backends.zarr.ZarrStore at 0xffff30506f80>\n \n In [23]: ds.isel(x=slice(10, 20)).to_zarr(path, region={"x": "auto"})\n-Out[23]: <xarray.backends.zarr.ZarrStore at 0xffff35b56dd0>\n+Out[23]: <xarray.backends.zarr.ZarrStore at 0xffff30506b90>\n \n In [24]: ds.isel(x=slice(20, 30)).to_zarr(path, region={"x": slice(20, 30)})\n-Out[24]: <xarray.backends.zarr.ZarrStore at 0xffff35f2cee0>\n+Out[24]: <xarray.backends.zarr.ZarrStore at 0xffff308dcee0>\n
Concurrent writes with region
are safe as long as they modify distinct\n chunks in the underlying Zarr arrays (or use an appropriate lock
).
As a safety check to make it harder to inadvertently override existing values,\n if you set region
then all variables included in a Dataset must have\n dimensions included in region
. Other variables (typically coordinates)\n@@ -816,28 +816,28 @@\n ....: "y": [1, 2, 3, 4, 5],\n ....: "t": pd.date_range("2001-01-01", periods=2),\n ....: },\n ....: )\n ....: \n \n In [30]: ds1.to_zarr("path/to/directory.zarr")\n-Out[30]: <xarray.backends.zarr.ZarrStore at 0xffff35c0c1f0>\n+Out[30]: <xarray.backends.zarr.ZarrStore at 0xffff305bc1f0>\n \n In [31]: ds2 = xr.Dataset(\n ....: {"foo": (("x", "y", "t"), np.random.rand(4, 5, 2))},\n ....: coords={\n ....: "x": [10, 20, 30, 40],\n ....: "y": [1, 2, 3, 4, 5],\n ....: "t": pd.date_range("2001-01-03", periods=2),\n ....: },\n ....: )\n ....: \n \n In [32]: ds2.to_zarr("path/to/directory.zarr", append_dim="t")\n-Out[32]: <xarray.backends.zarr.ZarrStore at 0xffff35c0c160>\n+Out[32]: <xarray.backends.zarr.ZarrStore at 0xffff305bc160>\n
Chunk sizes may be specified in one of three ways when writing to a zarr store:
\nFor example, let\u2019s say we\u2019re working with a dataset with dimensions\n ('time', 'x', 'y')
, a variable Tair
which is chunked in x
and y
,\n and two multi-dimensional coordinates xc
and yc
:
In [33]: ds = xr.tutorial.open_dataset("rasm")\n-ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/rasm.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff35c6b390>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n+ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/rasm.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff3061f390>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n \n \n In [34]: ds["Tair"] = ds["Tair"].chunk({"x": 100, "y": 100})\n KeyError: "No variable named 'Tair'. Variables on the dataset include ['foo', 'x']"\n \n \n In [35]: ds\n@@ -882,15 +882,15 @@\n foo (x) int64 240B 0 1 2 3 4 5 6 7 8 9 ... 21 22 23 24 25 26 27 28 29\n
These multi-dimensional coordinates are only two-dimensional and take up very little\n space on disk or in memory, yet when writing to disk the default zarr behavior is to\n split them into chunks:
\nIn [36]: ds.to_zarr("path/to/directory.zarr", mode="w")\n-Out[36]: <xarray.backends.zarr.ZarrStore at 0xffff35c0c5e0>\n+Out[36]: <xarray.backends.zarr.ZarrStore at 0xffff305bc5e0>\n \n In [37]: ! ls -R path/to/directory.zarr\n path/to/directory.zarr:\n foo x\tzarr.json\n \n path/to/directory.zarr/foo:\n c zarr.json\n@@ -1081,15 +1081,15 @@\n Ncdata\u00b6
\n Ncdata provides more sophisticated means of transferring data, including entire\n datasets. It uses the file saving and loading functions in both projects to provide a\n more \u201ccorrect\u201d translation between them, but still with very low overhead and not\n using actual disk files.
\n For example:
\n In [48]: ds = xr.tutorial.open_dataset("air_temperature_gradient")\n-ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature_gradient.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff6fac0a50>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n+ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature_gradient.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff664d0a50>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n \n \n In [49]: cubes = ncdata.iris_xarray.cubes_from_xarray(ds)\n NameError: name 'ncdata' is not defined\n \n \n In [50]: print(cubes)\n", "details": [{"source1": "html2text {}", "source2": "html2text {}", "unified_diff": "@@ -481,15 +481,15 @@\n ....: \"y\": pd.date_range(\"2000-01-01\", periods=5),\n ....: \"z\": (\"x\", list(\"abcd\")),\n ....: },\n ....: )\n ....:\n \n In [13]: ds.to_zarr(\"path/to/directory.zarr\")\n-Out[13]: \n+Out[13]: \n (The suffix .zarr is optional\u2013just a reminder that a zarr store lives there.)\n If the directory does not exist, it will be created. If a zarr store is already\n present at that path, an error will be raised, preventing it from being\n overwritten. To override this behavior and overwrite an existing store, add\n mode='w' when invoking to_zarr().\n DataArrays can also be saved to disk using the DataArray.to_zarr() method, and\n loaded from disk using the open_dataarray() function with engine='zarr'.\n@@ -507,17 +507,17 @@\n In [14]: ds_zarr = xr.open_zarr(\"path/to/directory.zarr\")\n \n In [15]: ds_zarr\n Out[15]:\n Size: 264B\n Dimensions: (x: 4, y: 5)\n Coordinates:\n+ * x (x) int64 32B 10 20 30 40\n z (x) object 32B dask.array\n * y (y) datetime64[ns] 40B 2000-01-01 2000-01-02 ... 2000-01-05\n- * x (x) int64 32B 10 20 30 40\n Data variables:\n foo (x, y) float64 160B dask.array\n *\b**\b**\b**\b* C\bCl\blo\bou\bud\bd S\bSt\bto\bor\bra\bag\bge\be B\bBu\buc\bck\bke\bet\bts\bs_\b?\b\u00b6 *\b**\b**\b**\b*\n It is possible to read and write xarray datasets directly from / to cloud\n storage buckets using zarr. This example uses the _\bg_\bc_\bs_\bf_\bs package to provide an\n interface to _\bG_\bo_\bo_\bg_\bl_\be_\b _\bC_\bl_\bo_\bu_\bd_\b _\bS_\bt_\bo_\br_\ba_\bg_\be.\n General _\bf_\bs_\bs_\bp_\be_\bc URLs, those that begin with s3:// or gcs:// for example, are\n@@ -562,35 +562,35 @@\n \n In [18]: ds = xr.Dataset({\"foo\": (\"x\", dummies)}, coords={\"x\": np.arange(30)})\n \n In [19]: path = \"path/to/directory.zarr\"\n \n # Now we write the metadata without computing any array values\n In [20]: ds.to_zarr(path, compute=False)\n-Out[20]: Delayed('_finalize_store-9c2dca6b-5842-4f3d-84bb-0aa396adbcad')\n+Out[20]: Delayed('_finalize_store-21a4d77a-7afa-456e-972d-01a3d5cfaa3b')\n Now, a Zarr store with the correct variable shapes and attributes exists that\n can be filled out by subsequent calls to to_zarr. Setting region=\"auto\" will\n open the existing store and determine the correct alignment of the new data\n with the existing dimensions, or as an explicit mapping from dimension names to\n Python slice objects indicating where the data should be written (in index\n space, not label space), e.g.,\n # For convenience, we'll slice a single dataset, but in the real use-case\n # we would create them separately possibly even from separate processes.\n In [21]: ds = xr.Dataset({\"foo\": (\"x\", np.arange(30))}, coords={\"x\": np.arange\n (30)})\n \n # Any of the following region specifications are valid\n In [22]: ds.isel(x=slice(0, 10)).to_zarr(path, region=\"auto\")\n-Out[22]: \n+Out[22]: \n \n In [23]: ds.isel(x=slice(10, 20)).to_zarr(path, region={\"x\": \"auto\"})\n-Out[23]: \n+Out[23]: \n \n In [24]: ds.isel(x=slice(20, 30)).to_zarr(path, region={\"x\": slice(20, 30)})\n-Out[24]: \n+Out[24]: \n Concurrent writes with region are safe as long as they modify distinct chunks\n in the underlying Zarr arrays (or use an appropriate lock).\n As a safety check to make it harder to inadvertently override existing values,\n if you set region then a\bal\bll\bl variables included in a Dataset must have dimensions\n included in region. Other variables (typically coordinates) need to be\n explicitly dropped and/or written in a separate calls to to_zarr with mode='a'.\n *\b**\b**\b**\b* Z\bZa\bar\brr\br C\bCo\bom\bmp\bpr\bre\bes\bss\bso\bor\brs\bs a\ban\bnd\bd F\bFi\bil\blt\bte\ber\brs\bs_\b?\b\u00b6 *\b**\b**\b**\b*\n@@ -636,28 +636,28 @@\n ....: \"y\": [1, 2, 3, 4, 5],\n ....: \"t\": pd.date_range(\"2001-01-01\", periods=2),\n ....: },\n ....: )\n ....:\n \n In [30]: ds1.to_zarr(\"path/to/directory.zarr\")\n-Out[30]: \n+Out[30]: \n \n In [31]: ds2 = xr.Dataset(\n ....: {\"foo\": ((\"x\", \"y\", \"t\"), np.random.rand(4, 5, 2))},\n ....: coords={\n ....: \"x\": [10, 20, 30, 40],\n ....: \"y\": [1, 2, 3, 4, 5],\n ....: \"t\": pd.date_range(\"2001-01-03\", periods=2),\n ....: },\n ....: )\n ....:\n \n In [32]: ds2.to_zarr(\"path/to/directory.zarr\", append_dim=\"t\")\n-Out[32]: \n+Out[32]: \n *\b**\b**\b**\b* S\bSp\bpe\bec\bci\bif\bfy\byi\bin\bng\bg c\bch\bhu\bun\bnk\bks\bs i\bin\bn a\ba z\bza\bar\brr\br s\bst\bto\bor\bre\be_\b?\b\u00b6 *\b**\b**\b**\b*\n Chunk sizes may be specified in one of three ways when writing to a zarr store:\n 1. Manual chunk sizing through the use of the encoding argument in\n Dataset.to_zarr():\n 2. Automatic chunking based on chunks in dask arrays\n 3. Default chunk behavior determined by the zarr library\n The resulting chunks will be determined based on the order of the above list;\n@@ -678,15 +678,15 @@\n For example, let\u2019s say we\u2019re working with a dataset with dimensions ('time',\n 'x', 'y'), a variable Tair which is chunked in x and y, and two multi-\n dimensional coordinates xc and yc:\n In [33]: ds = xr.tutorial.open_dataset(\"rasm\")\n ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries\n exceeded with url: /pydata/xarray-data/raw/master/rasm.nc (Caused by\n NameResolutionError(\": Failed to resolve 'github.com' ([Errno -3] Temporary failure\n+0xffff3061f390>: Failed to resolve 'github.com' ([Errno -3] Temporary failure\n in name resolution)\"))\n \n \n In [34]: ds[\"Tair\"] = ds[\"Tair\"].chunk({\"x\": 100, \"y\": 100})\n KeyError: \"No variable named 'Tair'. Variables on the dataset include ['foo',\n 'x']\"\n \n@@ -699,15 +699,15 @@\n * x (x) int64 240B 0 1 2 3 4 5 6 7 8 9 ... 21 22 23 24 25 26 27 28 29\n Data variables:\n foo (x) int64 240B 0 1 2 3 4 5 6 7 8 9 ... 21 22 23 24 25 26 27 28 29\n These multi-dimensional coordinates are only two-dimensional and take up very\n little space on disk or in memory, yet when writing to disk the default zarr\n behavior is to split them into chunks:\n In [36]: ds.to_zarr(\"path/to/directory.zarr\", mode=\"w\")\n-Out[36]: \n+Out[36]: \n \n In [37]: ! ls -R path/to/directory.zarr\n path/to/directory.zarr:\n foo x\tzarr.json\n \n path/to/directory.zarr/foo:\n c zarr.json\n@@ -874,15 +874,15 @@\n provide a more \u201ccorrect\u201d translation between them, but still with very low\n overhead and not using actual disk files.\n For example:\n In [48]: ds = xr.tutorial.open_dataset(\"air_temperature_gradient\")\n ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries\n exceeded with url: /pydata/xarray-data/raw/master/air_temperature_gradient.nc\n (Caused by NameResolutionError(\": Failed to resolve 'github.com' ([Errno -3] Temporary failure\n+0xffff664d0a50>: Failed to resolve 'github.com' ([Errno -3] Temporary failure\n in name resolution)\"))\n \n \n In [49]: cubes = ncdata.iris_xarray.cubes_from_xarray(ds)\n NameError: name 'ncdata' is not defined\n \n \n"}]}, {"source1": "./usr/share/doc/python-xarray-doc/html/user-guide/plotting.html", "source2": "./usr/share/doc/python-xarray-doc/html/user-guide/plotting.html", "unified_diff": "@@ -100,15 +100,15 @@\n In [3]: import matplotlib.pyplot as plt\n \n In [4]: import xarray as xr\n
\n \n For these examples we\u2019ll use the North American air temperature dataset.
\n In [5]: airtemps = xr.tutorial.open_dataset("air_temperature")\n-ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff6fac3d90>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n+ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff664d3d90>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n \n \n In [6]: airtemps\n NameError: name 'airtemps' is not defined\n \n \n # Convert to celsius\n@@ -445,15 +445,15 @@\n \n # Apply a nonlinear transformation to one of the coords\n In [50]: b.coords["lat"] = np.log(b.coords["lat"])\n KeyError: 'lat'\n \n \n In [51]: b.plot()\n-Out[51]: [<matplotlib.lines.Line2D at 0xffff35dfbb10>]\n+Out[51]: [<matplotlib.lines.Line2D at 0xffff307a7b10>]\n
\n \n
\n \n \n \n Other types of plot\u00b6
\n@@ -857,117 +857,117 @@\n * y (y) float64 88B 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0\n * z (z) int64 32B 0 1 2 3\n * w (w) <U5 80B 'one' 'two' 'three' 'five'\n Attributes:\n units: Aunits\n \n In [99]: ds.A.plot.scatter(x="y")\n-Out[99]: <matplotlib.collections.PathCollection at 0xffff35bfd550>\n+Out[99]: <matplotlib.collections.PathCollection at 0xffff30591160>\n
Same plot can be displayed using the dataset:
\nIn [100]: ds.plot.scatter(x="y", y="A")\n-Out[100]: <matplotlib.collections.PathCollection at 0xffff6ef31e50>\n+Out[100]: <matplotlib.collections.PathCollection at 0xffff65941e50>\n
Now suppose we want to scatter the A
DataArray against the B
DataArray
In [101]: ds.plot.scatter(x="A", y="B")\n-Out[101]: <matplotlib.collections.PathCollection at 0xffff6f020a50>\n+Out[101]: <matplotlib.collections.PathCollection at 0xffff65a30a50>\n
The hue
kwarg lets you vary the color by variable value
In [102]: ds.plot.scatter(x="A", y="B", hue="w")\n-Out[102]: <matplotlib.collections.PathCollection at 0xffff6f0fe0d0>\n+Out[102]: <matplotlib.collections.PathCollection at 0xffff65b120d0>\n
You can force a legend instead of a colorbar by setting add_legend=True, add_colorbar=False
.
In [103]: ds.plot.scatter(x="A", y="B", hue="w", add_legend=True, add_colorbar=False)\n-Out[103]: <matplotlib.collections.PathCollection at 0xffff3e273d90>\n+Out[103]: <matplotlib.collections.PathCollection at 0xffff3cc57d90>\n
In [104]: ds.plot.scatter(x="A", y="B", hue="w", add_legend=False, add_colorbar=True)\n-Out[104]: <matplotlib.collections.PathCollection at 0xffff3e351450>\n+Out[104]: <matplotlib.collections.PathCollection at 0xffff3cd2d450>\n
The markersize
kwarg lets you vary the point\u2019s size by variable value.\n You can additionally pass size_norm
to control how the variable\u2019s values are mapped to point sizes.
In [105]: ds.plot.scatter(x="A", y="B", hue="y", markersize="z")\n-Out[105]: <matplotlib.collections.PathCollection at 0xffff5ffe60d0>\n+Out[105]: <matplotlib.collections.PathCollection at 0xffff569f60d0>\n
The z
kwarg lets you plot the data along the z-axis as well.
In [106]: ds.plot.scatter(x="A", y="B", z="z", hue="y", markersize="x")\n-Out[106]: <mpl_toolkits.mplot3d.art3d.Path3DCollection at 0xffff3e32f890>\n+Out[106]: <mpl_toolkits.mplot3d.art3d.Path3DCollection at 0xffff3cd0b890>\n
Faceting is also possible
\nIn [107]: ds.plot.scatter(x="A", y="B", hue="y", markersize="x", row="x", col="w")\n-Out[107]: <xarray.plot.facetgrid.FacetGrid at 0xffff35bfc050>\n+Out[107]: <xarray.plot.facetgrid.FacetGrid at 0xffff30590050>\n
And adding the z-axis
\nIn [108]: ds.plot.scatter(x="A", y="B", z="z", hue="y", markersize="x", row="x", col="w")\n-Out[108]: <xarray.plot.facetgrid.FacetGrid at 0xffff6e854cd0>\n+Out[108]: <xarray.plot.facetgrid.FacetGrid at 0xffff6526ccd0>\n
For more advanced scatter plots, we recommend converting the relevant data variables\n to a pandas DataFrame and using the extensive plotting capabilities of seaborn
.
Visualizing vector fields is supported with quiver plots:
\nIn [109]: ds.isel(w=1, z=1).plot.quiver(x="x", y="y", u="A", v="B")\n-Out[109]: <matplotlib.quiver.Quiver at 0xffff357541a0>\n+Out[109]: <matplotlib.quiver.Quiver at 0xffff301041a0>\n
where u
and v
denote the x and y direction components of the arrow vectors. Again, faceting is also possible:
In [110]: ds.plot.quiver(x="x", y="y", u="A", v="B", col="w", row="z", scale=4)\n-Out[110]: <xarray.plot.facetgrid.FacetGrid at 0xffff6e66fd90>\n+Out[110]: <xarray.plot.facetgrid.FacetGrid at 0xffff65083d90>\n
scale
is required for faceted quiver plots.\n The scale determines the number of data units per arrow length unit, i.e. a smaller scale parameter makes the arrow longer.
Visualizing vector fields is also supported with streamline plots:
\nIn [111]: ds.isel(w=1, z=1).plot.streamplot(x="x", y="y", u="A", v="B")\n-Out[111]: <matplotlib.collections.LineCollection at 0xffff6e1ea350>\n+Out[111]: <matplotlib.collections.LineCollection at 0xffff64bfe350>\n
where u
and v
denote the x and y direction components of the vectors tangent to the streamlines.\n Again, faceting is also possible:
In [112]: ds.plot.streamplot(x="x", y="y", u="A", v="B", col="w", row="z")\n-Out[112]: <xarray.plot.facetgrid.FacetGrid at 0xffff35b995b0>\n+Out[112]: <xarray.plot.facetgrid.FacetGrid at 0xffff3054d5b0>\n
To follow this section you\u2019ll need to have Cartopy installed and working.
\nThis script will plot the air temperature on a map.
\nIn [113]: import cartopy.crs as ccrs\n \n In [114]: air = xr.tutorial.open_dataset("air_temperature").air\n-ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff6e7c1810>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n+ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff651d5810>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n \n \n In [115]: p = air.isel(time=0).plot(\n .....: subplot_kws=dict(projection=ccrs.Orthographic(-80, 35), facecolor="gray"),\n .....: transform=ccrs.PlateCarree(),\n .....: )\n .....: \n@@ -1024,24 +1024,24 @@\n In [121]: import xarray.plot as xplt\n \n In [122]: da = xr.DataArray(range(5))\n \n In [123]: fig, axs = plt.subplots(ncols=2, nrows=2)\n \n In [124]: da.plot(ax=axs[0, 0])\n-Out[124]: [<matplotlib.lines.Line2D at 0xffff6d861a90>]\n+Out[124]: [<matplotlib.lines.Line2D at 0xffff64271a90>]\n \n In [125]: da.plot.line(ax=axs[0, 1])\n-Out[125]: [<matplotlib.lines.Line2D at 0xffff6d861950>]\n+Out[125]: [<matplotlib.lines.Line2D at 0xffff64271950>]\n \n In [126]: xplt.plot(da, ax=axs[1, 0])\n-Out[126]: [<matplotlib.lines.Line2D at 0xffff6d861f90>]\n+Out[126]: [<matplotlib.lines.Line2D at 0xffff64271f90>]\n \n In [127]: xplt.line(da, ax=axs[1, 1])\n-Out[127]: [<matplotlib.lines.Line2D at 0xffff6d8625d0>]\n+Out[127]: [<matplotlib.lines.Line2D at 0xffff642725d0>]\n \n In [128]: plt.tight_layout()\n \n In [129]: plt.draw()\n
\n \n
\n@@ -1091,15 +1091,15 @@\n
The plot will produce an image corresponding to the values of the array.\n Hence the top left pixel will be a different color than the others.\n Before reading on, you may want to look at the coordinates and\n think carefully about what the limits, labels, and orientation for\n each of the axes should be.
\nIn [134]: a.plot()\n-Out[134]: <matplotlib.collections.QuadMesh at 0xffff6db29450>\n+Out[134]: <matplotlib.collections.QuadMesh at 0xffff64539450>\n
It may seem strange that\n the values on the y axis are decreasing with -0.5 on the top. This is because\n the pixels are centered over their coordinates, and the\n@@ -1122,57 +1122,57 @@\n .....: np.arange(20).reshape(4, 5),\n .....: dims=["y", "x"],\n .....: coords={"lat": (("y", "x"), lat), "lon": (("y", "x"), lon)},\n .....: )\n .....: \n \n In [139]: da.plot.pcolormesh(x="lon", y="lat")\n-Out[139]: <matplotlib.collections.QuadMesh at 0xffff6dc551d0>\n+Out[139]: <matplotlib.collections.QuadMesh at 0xffff646691d0>\n
Note that in this case, xarray still follows the pixel centered convention.\n This might be undesirable in some cases, for example when your data is defined\n on a polar projection (GH781). This is why the default is to not follow\n this convention when plotting on a map:
\nIn [140]: import cartopy.crs as ccrs\n \n In [141]: ax = plt.subplot(projection=ccrs.PlateCarree())\n \n In [142]: da.plot.pcolormesh(x="lon", y="lat", ax=ax)\n-Out[142]: <cartopy.mpl.geocollection.GeoQuadMesh at 0xffff6ddf7ed0>\n+Out[142]: <cartopy.mpl.geocollection.GeoQuadMesh at 0xffff6480bed0>\n \n In [143]: ax.scatter(lon, lat, transform=ccrs.PlateCarree())\n-Out[143]: <matplotlib.collections.PathCollection at 0xffff6dcc3ed0>\n+Out[143]: <matplotlib.collections.PathCollection at 0xffff646d3ed0>\n \n In [144]: ax.coastlines()\n-Out[144]: <cartopy.mpl.feature_artist.FeatureArtist at 0xffff35bfc590>\n+Out[144]: <cartopy.mpl.feature_artist.FeatureArtist at 0xffff30590590>\n \n In [145]: ax.gridlines(draw_labels=True)\n-Out[145]: <cartopy.mpl.gridliner.Gridliner at 0xffff37b5dd30>\n+Out[145]: <cartopy.mpl.gridliner.Gridliner at 0xffff3250dd30>\n
You can however decide to infer the cell boundaries and use the\n infer_intervals
keyword:
In [146]: ax = plt.subplot(projection=ccrs.PlateCarree())\n \n In [147]: da.plot.pcolormesh(x="lon", y="lat", ax=ax, infer_intervals=True)\n-Out[147]: <cartopy.mpl.geocollection.GeoQuadMesh at 0xffff6e3b6e90>\n+Out[147]: <cartopy.mpl.geocollection.GeoQuadMesh at 0xffff64dcae90>\n \n In [148]: ax.scatter(lon, lat, transform=ccrs.PlateCarree())\n-Out[148]: <matplotlib.collections.PathCollection at 0xffff6d941a90>\n+Out[148]: <matplotlib.collections.PathCollection at 0xffff64351a90>\n \n In [149]: ax.coastlines()\n-Out[149]: <cartopy.mpl.feature_artist.FeatureArtist at 0xffff6d941950>\n+Out[149]: <cartopy.mpl.feature_artist.FeatureArtist at 0xffff64351950>\n \n In [150]: ax.gridlines(draw_labels=True)\n-Out[150]: <cartopy.mpl.gridliner.Gridliner at 0xffff6d941f90>\n+Out[150]: <cartopy.mpl.gridliner.Gridliner at 0xffff64351f90>\n
Note
\nThe data model of xarray does not support datasets with cell boundaries\n@@ -1180,26 +1180,26 @@\n outside the xarray framework.
\nOne can also make line plots with multidimensional coordinates. In this case, hue
must be a dimension name, not a coordinate name.
In [151]: f, ax = plt.subplots(2, 1)\n \n In [152]: da.plot.line(x="lon", hue="y", ax=ax[0])\n Out[152]: \n-[<matplotlib.lines.Line2D at 0xffff6e3f1590>,\n- <matplotlib.lines.Line2D at 0xffff6e3f2210>,\n- <matplotlib.lines.Line2D at 0xffff6e184cd0>,\n- <matplotlib.lines.Line2D at 0xffff6e40cb90>]\n+[<matplotlib.lines.Line2D at 0xffff64e05590>,\n+ <matplotlib.lines.Line2D at 0xffff64e06210>,\n+ <matplotlib.lines.Line2D at 0xffff64b98cd0>,\n+ <matplotlib.lines.Line2D at 0xffff64c24b90>]\n \n In [153]: da.plot.line(x="lon", hue="x", ax=ax[1])\n Out[153]: \n-[<matplotlib.lines.Line2D at 0xffff6e40f390>,\n- <matplotlib.lines.Line2D at 0xffff6e40f250>,\n- <matplotlib.lines.Line2D at 0xffff6e40f110>,\n- <matplotlib.lines.Line2D at 0xffff6e40c190>,\n- <matplotlib.lines.Line2D at 0xffff6e40c050>]\n+[<matplotlib.lines.Line2D at 0xffff64c27390>,\n+ <matplotlib.lines.Line2D at 0xffff64c27250>,\n+ <matplotlib.lines.Line2D at 0xffff64c27110>,\n+ <matplotlib.lines.Line2D at 0xffff64c24190>,\n+ <matplotlib.lines.Line2D at 0xffff64c24050>]\n
Whilst coarsen
is normally used for reducing your data\u2019s resolution by applying a reduction function\n (see the page on computation),\n it can also be used to reorganise your data without applying a computation via construct()
.
Taking our example tutorial air temperature dataset over the Northern US
\nIn [56]: air = xr.tutorial.open_dataset("air_temperature")["air"]\n-ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff6e3f11d0>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n+ConnectionError: HTTPSConnectionPool(host='github.com', port=443): Max retries exceeded with url: /pydata/xarray-data/raw/master/air_temperature.nc (Caused by NameResolutionError("<urllib3.connection.HTTPSConnection object at 0xffff64e051d0>: Failed to resolve 'github.com' ([Errno -3] Temporary failure in name resolution)"))\n \n \n In [57]: air.isel(time=0).plot(x="lon", y="lat")\n NameError: name 'air' is not defined\n
To see an example of what each of these strategies might produce, you can call one followed by the .example()
method,\n which is a general hypothesis method valid for all strategies.
In [2]: import xarray.testing.strategies as xrst\n \n In [3]: xrst.variables().example()\n Out[3]: \n-<xarray.Variable (0: 1)> Size: 2B\n-array([0.], dtype=float16)\n+<xarray.Variable (\u011f: 3)> Size: 48B\n+array([-2.000e+000-7.065e-237j, -1.798e+308-2.225e-311j, nan -infj])\n \n In [4]: xrst.variables().example()\n Out[4]: \n-<xarray.Variable (4\u017f\u017d\u0170\u012a: 1)> Size: 4B\n-array([inf], dtype=float32)\n+<xarray.Variable (\u017d\u010dB\u0108X: 5, \u014e\u011e\u00ef\u00cc\u0128: 3, Z\u017d\u0109: 2)> Size: 120B\n+array([[[-2147472357, 470053658],\n+ [-2147483505, -783616614],\n+ [ 86016487, -2147450254]],\n+\n+ [[ 86016487, 86016487],\n+ [ 86016487, 86016487],\n+ [ 86016487, 86016487]],\n+\n+ [[ 86016487, 86016487],\n+ [ 86016487, 86016487],\n+ [ -337533136, 86016487]],\n+\n+ [[-2147483647, -2147483494],\n+ [ 86016487, -2147437085],\n+ [ 86016487, -2147454546]],\n+\n+ [[ 86016487, 86016487],\n+ [ 86016487, 1019695362],\n+ [-1751677034, 86016487]]], shape=(5, 3, 2), dtype=int32)\n \n In [5]: xrst.variables().example()\n Out[5]: \n-<xarray.Variable (\u017d\u015a\u0137\u00d1: 3)> Size: 6B\n-array([-15611, -15611, 32078], dtype=int16)\n+<xarray.Variable (\u017c\u017b\u00ef\u0124\u00e6: 5)> Size: 40B\n+array([ nan -infj, -2.000e+00-3.353e+16j, -6.252e+15+1.175e-38j, inf +nanj,\n+ 0.000e+00 -infj], dtype=complex64)\n+Attributes:\n+ \u0177\u00d1\u017d: True\n+ \u0169\u0177\u014f\u016e\u013b: \n+ \u00ec: [[b'\\xe1\\xa1\\x16']\\n [b'\\xe8m\\x03']]\n+ \u0128IY: 2\u017c\n+ 4: V\u012a\u013e\u00e5j\n+ 0\u017e\u00d8\u0115\u013c: nil\n
You can see that calling .example()
multiple times will generate different examples, giving you an idea of the wide\n range of data that the xarray strategies can generate.
In your tests however you should not use .example()
- instead you should parameterize your tests with the\n hypothesis.given()
decorator:
In [6]: from hypothesis import given\n@@ -121,114 +147,84 @@\n Xarray\u2019s strategies can accept other strategies as arguments, allowing you to customise the contents of the generated\n examples.
\n # generate a Variable containing an array with a complex number dtype, but all other details still arbitrary\n In [8]: from hypothesis.extra.numpy import complex_number_dtypes\n \n In [9]: xrst.variables(dtype=complex_number_dtypes()).example()\n Out[9]: \n-<xarray.Variable (0: 1)> Size: 8B\n-array([0.+0.j], dtype='>c8')\n+<xarray.Variable (\u015f\u00f66: 4)> Size: 64B\n+array([ inf+5.485e+15j, inf+1.900e+00j, inf +nanj,\n+ -2.225e-309-1.175e-38j])\n
\n \n This also works with custom strategies, or strategies defined in other packages.\n For example you could imagine creating a chunks
strategy to specify particular chunking patterns for a dask-backed array.
\n \n \n Fixing Arguments\u00b6
\n If you want to fix one aspect of the data structure, whilst allowing variation in the generated examples\n over all other aspects, then use hypothesis.strategies.just()
.
\n In [10]: import hypothesis.strategies as st\n \n # Generates only variable objects with dimensions ["x", "y"]\n In [11]: xrst.variables(dims=st.just(["x", "y"])).example()\n Out[11]: \n-<xarray.Variable (x: 5, y: 5)> Size: 25B\n-array([[ 80, -44, 19, -44, -31],\n- [-13, -44, -44, -44, -44],\n- [-51, -44, 31, -1, -44],\n- [-44, -44, -44, -44, -44],\n- [-92, -44, -44, -44, 111]], shape=(5, 5), dtype=int8)\n-Attributes:\n- \u0153\u0101: [b'f']\n- c: [-83]\n- \u017d\u00f9: True\n- \u0109f: None\n- \u016a\u017b\u00ed\u00c7\u0160: [b'\\xc7' b'\\xd7\\xddB?']\n- \u017d\u0173\u017do: None\n+<xarray.Variable (x: 1, y: 1)> Size: 1B\n+array([[0]], dtype=int8)\n
\n \n (This is technically another example of chaining strategies - hypothesis.strategies.just()
is simply a\n special strategy that just contains a single example.)
\n To fix the length of dimensions you can instead pass dims
as a mapping of dimension names to lengths\n (i.e. following xarray objects\u2019 .sizes()
property), e.g.
\n # Generates only variables with dimensions ["x", "y"], of lengths 2 & 3 respectively\n In [12]: xrst.variables(dims=st.just({"x": 2, "y": 3})).example()\n Out[12]: \n-<xarray.Variable (x: 2, y: 3)> Size: 96B\n-array([[6.992e+188-1.9j, 6.992e+188-1.9j, 6.992e+188-1.9j],\n- [6.992e+188-1.9j, 6.992e+188-1.9j, 6.992e+188-1.9j]])\n-Attributes:\n- \u00ed\u014e\u016a\u00f2: 1X\n- \u00e6cf: None\n- \u016f: \n- \u017b\u017e: None\n- \u00d4K\u00c6\u00e7\u00ff: [['\u00c6' '\\U0001a737\\x8c']]\n- \u017e9\u0151\u00c1\u00c9: ['fc\u00d4\\x92e' 'fc\u00d4\\x92e']\n- \u0148\u017d\u00f1M: None\n- : [['\\x0c\u00c7\u00b5u\\x1a\\U000bb2a0\u00e6' '#\ud883\ude29\\U000be7cba9\\x11\\x10\u00b5\\U00077b9a']]\n+<xarray.Variable (x: 2, y: 3)> Size: 12B\n+array([[25919, 57262, 3865],\n+ [34631, 32393, 39493]], dtype=uint16)\n
\n \n You can also use this to specify that you want examples which are missing some part of the data structure, for instance
\n # Generates a Variable with no attributes\n In [13]: xrst.variables(attrs=st.just({})).example()\n Out[13]: \n-<xarray.Variable (NULL: 4, \u017b\u017b\u0152\u0176\u017f: 3)> Size: 12B\n-array([[150, 0, 198],\n- [187, 62, 31],\n- [ 98, 150, 34],\n- [ 46, 165, 61]], shape=(4, 3), dtype=uint8)\n+<xarray.Variable (\u00f9\u017c: 4, \u017b\u0141\u0112\u013b\u015b: 2)> Size: 128B\n+array([[ 1.500e+00-1.175e-038j, 1.095e+24+1.411e+016j],\n+ [ nan-1.100e+000j, inf +nanj],\n+ [-1.900e+00+3.333e-001j, -8.641e-69+1.000e-005j],\n+ [ 1.900e+00 +infj, 1.175e-38+1.113e-308j]])\n
\n \n Through a combination of chaining strategies and fixing arguments, you can specify quite complicated requirements on the\n objects your chained strategy will generate.
\n In [14]: fixed_x_variable_y_maybe_z = st.fixed_dictionaries(\n ....: {"x": st.just(2), "y": st.integers(3, 4)}, optional={"z": st.just(2)}\n ....: )\n ....: \n \n In [15]: fixed_x_variable_y_maybe_z.example()\n-Out[15]: {'x': 2, 'y': 3, 'z': 2}\n+Out[15]: {'x': 2, 'y': 4}\n \n In [16]: special_variables = xrst.variables(dims=fixed_x_variable_y_maybe_z)\n \n In [17]: special_variables.example()\n Out[17]: \n-<xarray.Variable (x: 2, y: 4, z: 2)> Size: 128B\n-array([[[ 5.e-01+1.401e-45j, 5.e-01+1.401e-45j],\n- [ 5.e-01+1.401e-45j, 5.e-01+1.401e-45j],\n- [ 5.e-01+1.401e-45j, 5.e-01+1.401e-45j],\n- [-0.e+00+9.007e+15j, 5.e-01+1.401e-45j]],\n-\n- [[ 1.e-05 +infj, 5.e-01+1.401e-45j],\n- [ 5.e-01+1.401e-45j, 5.e-01+1.401e-45j],\n- [ 5.e-01+1.401e-45j, 5.e-01+1.401e-45j],\n- [ 5.e-01+1.401e-45j, 5.e-01+1.401e-45j]]], shape=(2, 4, 2), dtype=complex64)\n+<xarray.Variable (x: 2, y: 4)> Size: 32B\n+array([[4154222754, 13797, 16898, 1830933917],\n+ [4154222754, 194, 3765082830, 8648]], dtype=uint32)\n Attributes:\n- X\u017b\u0114\u0170q: \n+ \u00bc\u012a\u0121\u017e\u0142: {}\n+ False: {'\u00f0\u0139\u0136\u0129\u00c8': None, '4g\u00db\u00c5\u0158': array([b'\\xf0B\\xa3'], dtype='|S16'), '...\n \n In [18]: special_variables.example()\n Out[18]: \n-<xarray.Variable (x: 2, y: 3, z: 2)> Size: 192B\n-array([[[-2.086e+03 -infj, -2.086e+03 -infj],\n- [-2.086e+03 -infj, -2.086e+03 -infj],\n- [-2.086e+03 -infj, -2.086e+03 -infj]],\n-\n- [[-2.086e+03 -infj, -4.912e+16 +nanj],\n- [-2.086e+03 -infj, inf+5.633e-36j],\n- [-2.086e+03 -infj, nan-6.104e-05j]]], shape=(2, 3, 2))\n+<xarray.Variable (x: 2, y: 3)> Size: 6B\n+array([[0, 0, 0],\n+ [0, 0, 0]], dtype=int8)\n
\n \n Here we have used one of hypothesis\u2019 built-in strategies hypothesis.strategies.fixed_dictionaries()
to create a\n strategy which generates mappings of dimension names to lengths (i.e. the size
of the xarray object we want).\n This particular strategy will always generate an x
dimension of length 2, and a y
dimension of\n length either 3 or 4, and will sometimes also generate a z
dimension of length 2.\n By feeding this strategy for dictionaries into the dims
argument of xarray\u2019s variables()
strategy,\n@@ -329,65 +325,43 @@\n ....: array_strategy_fn=xps.arrays,\n ....: dtype=xps.scalar_dtypes(),\n ....: )\n ....: \n \n In [32]: xp_variables.example()\n Out[32]: \n-<xarray.Variable (\u00eb\u017a: 4, \u016e\u0166: 5, \u00deIJgZ: 1)> Size: 160B\n-array([[[ 9223372036854775807],\n- [ 9223372036854775807],\n- [-9223372036854761127],\n- [-9223372036854712296],\n- [ 555449524449720797]],\n-\n- [[ 9223372036854775807],\n- [-9223372036854775608],\n- [-1693137102084182800],\n- [-9223372036854730810],\n- [-9223372036854757140]],\n-\n- [[-9223372036854767111],\n- [-5146262954906203667],\n- [ 9223372036854775807],\n- [-9223372036854766292],\n- [-9223372035969212156]],\n-\n- [[ 9223372036854775807],\n- [-1226473953007738667],\n- [-9223372036746988332],\n- [ 9223372036854775807],\n- [-9223372036854763286]]], shape=(4, 5, 1))\n+<xarray.Variable (0: 1)> Size: 1B\n+array([0], dtype=int8)\n
Another array API-compliant duck array library would replace the import, e.g. import cupy as cp
instead.
A common task when testing xarray user code is checking that your function works for all valid input dimensions.\n We can chain strategies to achieve this, for which the helper strategy unique_subset_of()
\n is useful.
It works for lists of dimension names
\nIn [33]: dims = ["x", "y", "z"]\n \n In [34]: xrst.unique_subset_of(dims).example()\n-Out[34]: ['y', 'x']\n+Out[34]: ['x', 'y', 'z']\n \n In [35]: xrst.unique_subset_of(dims).example()\n-Out[35]: ['y', 'z', 'x']\n+Out[35]: ['z', 'y']\n
as well as for mappings of dimension names to sizes
\nIn [36]: dim_sizes = {"x": 2, "y": 3, "z": 4}\n \n In [37]: xrst.unique_subset_of(dim_sizes).example()\n-Out[37]: {'x': 2, 'y': 3, 'z': 4}\n+Out[37]: {'y': 3, 'x': 2}\n \n In [38]: xrst.unique_subset_of(dim_sizes).example()\n-Out[38]: {'x': 2, 'z': 4, 'y': 3}\n+Out[38]: {'x': 2, 'y': 3}\n
This is useful because operations like reductions can be performed over any subset of the xarray object\u2019s dimensions.\n For example we can write a pytest test that tests that a reduction gives the expected result when applying that reduction\n along any possible valid subset of the Variable\u2019s dimensions.
\nimport numpy.testing as npt\n \n", "details": [{"source1": "html2text {}", "source2": "html2text {}", "unified_diff": "@@ -28,26 +28,54 @@\n To see an example of what each of these strategies might produce, you can call\n one followed by the .example() method, which is a general hypothesis method\n valid for all strategies.\n In [2]: import xarray.testing.strategies as xrst\n \n In [3]: xrst.variables().example()\n Out[3]:\n- Size: 2B\n-array([0.], dtype=float16)\n+ Size: 48B\n+array([-2.000e+000-7.065e-237j, -1.798e+308-2.225e-311j, nan -\n+infj])\n \n In [4]: xrst.variables().example()\n Out[4]:\n- Size: 4B\n-array([inf], dtype=float32)\n+ Size: 120B\n+array([[[-2147472357, 470053658],\n+ [-2147483505, -783616614],\n+ [ 86016487, -2147450254]],\n+\n+ [[ 86016487, 86016487],\n+ [ 86016487, 86016487],\n+ [ 86016487, 86016487]],\n+\n+ [[ 86016487, 86016487],\n+ [ 86016487, 86016487],\n+ [ -337533136, 86016487]],\n+\n+ [[-2147483647, -2147483494],\n+ [ 86016487, -2147437085],\n+ [ 86016487, -2147454546]],\n+\n+ [[ 86016487, 86016487],\n+ [ 86016487, 1019695362],\n+ [-1751677034, 86016487]]], shape=(5, 3, 2), dtype=int32)\n \n In [5]: xrst.variables().example()\n Out[5]:\n- Size: 6B\n-array([-15611, -15611, 32078], dtype=int16)\n+ Size: 40B\n+array([ nan -infj, -2.000e+00-3.353e+16j, -6.252e+15+1.175e-38j,\n+inf +nanj,\n+ 0.000e+00 -infj], dtype=complex64)\n+Attributes:\n+ \u0177\u00d1\u017d: True\n+ \u0169\u0177\u014f\u016e\u013b:\n+ \u00ec: [[b'\\xe1\\xa1\\x16']\\n [b'\\xe8m\\x03']]\n+ \u0128IY: 2\u017c\n+ 4: V\u012a\u013e\u00e5j\n+ 0\u017e\u00d8\u0115\u013c: nil\n You can see that calling .example() multiple times will generate different\n examples, giving you an idea of the wide range of data that the xarray\n strategies can generate.\n In your tests however you should not use .example() - instead you should\n parameterize your tests with the hypothesis.given() decorator:\n In [6]: from hypothesis import given\n In [7]: @given(xrst.variables())\n@@ -59,113 +87,83 @@\n customise the contents of the generated examples.\n # generate a Variable containing an array with a complex number dtype, but all\n other details still arbitrary\n In [8]: from hypothesis.extra.numpy import complex_number_dtypes\n \n In [9]: xrst.variables(dtype=complex_number_dtypes()).example()\n Out[9]:\n- Size: 8B\n-array([0.+0.j], dtype='>c8')\n+ Size: 64B\n+array([ inf+5.485e+15j, inf+1.900e+00j, inf +nanj,\n+ -2.225e-309-1.175e-38j])\n This also works with custom strategies, or strategies defined in other\n packages. For example you could imagine creating a chunks strategy to specify\n particular chunking patterns for a dask-backed array.\n *\b**\b**\b**\b* F\bFi\bix\bxi\bin\bng\bg A\bAr\brg\bgu\bum\bme\ben\bnt\bts\bs_\b?\b\u00b6 *\b**\b**\b**\b*\n If you want to fix one aspect of the data structure, whilst allowing variation\n in the generated examples over all other aspects, then use\n hypothesis.strategies.just().\n In [10]: import hypothesis.strategies as st\n \n # Generates only variable objects with dimensions [\"x\", \"y\"]\n In [11]: xrst.variables(dims=st.just([\"x\", \"y\"])).example()\n Out[11]:\n- Size: 25B\n-array([[ 80, -44, 19, -44, -31],\n- [-13, -44, -44, -44, -44],\n- [-51, -44, 31, -1, -44],\n- [-44, -44, -44, -44, -44],\n- [-92, -44, -44, -44, 111]], shape=(5, 5), dtype=int8)\n-Attributes:\n- \u0153\u0101: [b'f']\n- c: [-83]\n- \u017d\u00f9: True\n- \u0109f: None\n- \u016a\u017b\u00ed\u00c7\u0160: [b'\\xc7' b'\\xd7\\xddB?']\n- \u017d\u0173\u017do: None\n+ Size: 1B\n+array([[0]], dtype=int8)\n (This is technically another example of chaining strategies -\n hypothesis.strategies.just() is simply a special strategy that just contains a\n single example.)\n To fix the length of dimensions you can instead pass dims as a mapping of\n dimension names to lengths (i.e. following xarray objects\u2019 .sizes() property),\n e.g.\n # Generates only variables with dimensions [\"x\", \"y\"], of lengths 2 & 3\n respectively\n In [12]: xrst.variables(dims=st.just({\"x\": 2, \"y\": 3})).example()\n Out[12]:\n- Size: 96B\n-array([[6.992e+188-1.9j, 6.992e+188-1.9j, 6.992e+188-1.9j],\n- [6.992e+188-1.9j, 6.992e+188-1.9j, 6.992e+188-1.9j]])\n-Attributes:\n- \u00ed\u014e\u016a\u00f2: 1X\n- \u00e6cf: None\n- \u016f:\n- \u017b\u017e: None\n- \u00d4K\u00c6\u00e7\u00ff: [['\u00c6' '\\U0001a737\\x8c']]\n- \u017e9\u0151\u00c1\u00c9: ['fc\u00d4\\x92e' 'fc\u00d4\\x92e']\n- \u0148\u017d\u00f1M: None\n- : [['\\x0c\u00c7\u00b5u\\x1a\\U000bb2a0\u00e6' '#\ud883\ude29\\U000be7cba9\\x11\\x10\u00b5\\U00077b9a']]\n+ Size: 12B\n+array([[25919, 57262, 3865],\n+ [34631, 32393, 39493]], dtype=uint16)\n You can also use this to specify that you want examples which are missing some\n part of the data structure, for instance\n # Generates a Variable with no attributes\n In [13]: xrst.variables(attrs=st.just({})).example()\n Out[13]:\n- Size: 12B\n-array([[150, 0, 198],\n- [187, 62, 31],\n- [ 98, 150, 34],\n- [ 46, 165, 61]], shape=(4, 3), dtype=uint8)\n+ Size: 128B\n+array([[ 1.500e+00-1.175e-038j, 1.095e+24+1.411e+016j],\n+ [ nan-1.100e+000j, inf +nanj],\n+ [-1.900e+00+3.333e-001j, -8.641e-69+1.000e-005j],\n+ [ 1.900e+00 +infj, 1.175e-38+1.113e-308j]])\n Through a combination of chaining strategies and fixing arguments, you can\n specify quite complicated requirements on the objects your chained strategy\n will generate.\n In [14]: fixed_x_variable_y_maybe_z = st.fixed_dictionaries(\n ....: {\"x\": st.just(2), \"y\": st.integers(3, 4)}, optional={\"z\": st.just\n (2)}\n ....: )\n ....:\n \n In [15]: fixed_x_variable_y_maybe_z.example()\n-Out[15]: {'x': 2, 'y': 3, 'z': 2}\n+Out[15]: {'x': 2, 'y': 4}\n \n In [16]: special_variables = xrst.variables(dims=fixed_x_variable_y_maybe_z)\n \n In [17]: special_variables.example()\n Out[17]:\n- Size: 128B\n-array([[[ 5.e-01+1.401e-45j, 5.e-01+1.401e-45j],\n- [ 5.e-01+1.401e-45j, 5.e-01+1.401e-45j],\n- [ 5.e-01+1.401e-45j, 5.e-01+1.401e-45j],\n- [-0.e+00+9.007e+15j, 5.e-01+1.401e-45j]],\n-\n- [[ 1.e-05 +infj, 5.e-01+1.401e-45j],\n- [ 5.e-01+1.401e-45j, 5.e-01+1.401e-45j],\n- [ 5.e-01+1.401e-45j, 5.e-01+1.401e-45j],\n- [ 5.e-01+1.401e-45j, 5.e-01+1.401e-45j]]], shape=(2, 4, 2),\n-dtype=complex64)\n+ Size: 32B\n+array([[4154222754, 13797, 16898, 1830933917],\n+ [4154222754, 194, 3765082830, 8648]], dtype=uint32)\n Attributes:\n- X\u017b\u0114\u0170q:\n+ \u00bc\u012a\u0121\u017e\u0142: {}\n+ False: {'\u00f0\u0139\u0136\u0129\u00c8': None, '4g\u00db\u00c5\u0158': array([b'\\xf0B\\xa3'], dtype='|S16'),\n+'...\n \n In [18]: special_variables.example()\n Out[18]:\n- Size: 192B\n-array([[[-2.086e+03 -infj, -2.086e+03 -infj],\n- [-2.086e+03 -infj, -2.086e+03 -infj],\n- [-2.086e+03 -infj, -2.086e+03 -infj]],\n-\n- [[-2.086e+03 -infj, -4.912e+16 +nanj],\n- [-2.086e+03 -infj, inf+5.633e-36j],\n- [-2.086e+03 -infj, nan-6.104e-05j]]], shape=(2, 3, 2))\n+ Size: 6B\n+array([[0, 0, 0],\n+ [0, 0, 0]], dtype=int8)\n Here we have used one of hypothesis\u2019 built-in strategies\n hypothesis.strategies.fixed_dictionaries() to create a strategy which generates\n mappings of dimension names to lengths (i.e. the size of the xarray object we\n want). This particular strategy will always generate an x dimension of length\n 2, and a y dimension of length either 3 or 4, and will sometimes also generate\n a z dimension of length 2. By feeding this strategy for dictionaries into the\n dims argument of xarray\u2019s variables() strategy, we can generate arbitrary\n@@ -259,60 +257,38 @@\n ....: array_strategy_fn=xps.arrays,\n ....: dtype=xps.scalar_dtypes(),\n ....: )\n ....:\n \n In [32]: xp_variables.example()\n Out[32]:\n- Size: 160B\n-array([[[ 9223372036854775807],\n- [ 9223372036854775807],\n- [-9223372036854761127],\n- [-9223372036854712296],\n- [ 555449524449720797]],\n-\n- [[ 9223372036854775807],\n- [-9223372036854775608],\n- [-1693137102084182800],\n- [-9223372036854730810],\n- [-9223372036854757140]],\n-\n- [[-9223372036854767111],\n- [-5146262954906203667],\n- [ 9223372036854775807],\n- [-9223372036854766292],\n- [-9223372035969212156]],\n-\n- [[ 9223372036854775807],\n- [-1226473953007738667],\n- [-9223372036746988332],\n- [ 9223372036854775807],\n- [-9223372036854763286]]], shape=(4, 5, 1))\n+ Size: 1B\n+array([0], dtype=int8)\n Another array API-compliant duck array library would replace the import, e.g.\n import cupy as cp instead.\n *\b**\b**\b**\b* T\bTe\bes\bst\bti\bin\bng\bg o\bov\bve\ber\br S\bSu\bub\bbs\bse\bet\bts\bs o\bof\bf D\bDi\bim\bme\ben\bns\bsi\bio\bon\bns\bs_\b?\b\u00b6 *\b**\b**\b**\b*\n A common task when testing xarray user code is checking that your function\n works for all valid input dimensions. We can chain strategies to achieve this,\n for which the helper strategy unique_subset_of() is useful.\n It works for lists of dimension names\n In [33]: dims = [\"x\", \"y\", \"z\"]\n \n In [34]: xrst.unique_subset_of(dims).example()\n-Out[34]: ['y', 'x']\n+Out[34]: ['x', 'y', 'z']\n \n In [35]: xrst.unique_subset_of(dims).example()\n-Out[35]: ['y', 'z', 'x']\n+Out[35]: ['z', 'y']\n as well as for mappings of dimension names to sizes\n In [36]: dim_sizes = {\"x\": 2, \"y\": 3, \"z\": 4}\n \n In [37]: xrst.unique_subset_of(dim_sizes).example()\n-Out[37]: {'x': 2, 'y': 3, 'z': 4}\n+Out[37]: {'y': 3, 'x': 2}\n \n In [38]: xrst.unique_subset_of(dim_sizes).example()\n-Out[38]: {'x': 2, 'z': 4, 'y': 3}\n+Out[38]: {'x': 2, 'y': 3}\n This is useful because operations like reductions can be performed over any\n subset of the xarray object\u2019s dimensions. For example we can write a pytest\n test that tests that a reduction gives the expected result when applying that\n reduction along any possible valid subset of the Variable\u2019s dimensions.\n import numpy.testing as npt\n \n \n"}]}, {"source1": "./usr/share/doc/python-xarray-doc/html/whats-new.html", "source2": "./usr/share/doc/python-xarray-doc/html/whats-new.html", "unified_diff": "@@ -8191,15 +8191,15 @@\n New xray.Dataset.where
method for masking xray objects according\n to some criteria. This works particularly well with multi-dimensional data:
\n In [45]: ds = xray.Dataset(coords={"x": range(100), "y": range(100)})\n \n In [46]: ds["distance"] = np.sqrt(ds.x**2 + ds.y**2)\n \n In [47]: ds.distance.where(ds.distance < 100).plot()\n-Out[47]: <matplotlib.collections.QuadMesh at 0xffff35df8b90>\n+Out[47]: <matplotlib.collections.QuadMesh at 0xffff307a79d0>\n
\n \n
\n \n \n Added new methods xray.DataArray.diff
and xray.Dataset.diff
\n for finite difference calculations along a given axis.
\n", "details": [{"source1": "html2text {}", "source2": "html2text {}", "unified_diff": "@@ -5286,15 +5286,15 @@\n * New xray.Dataset.where method for masking xray objects according to some\n criteria. This works particularly well with multi-dimensional data:\n In [45]: ds = xray.Dataset(coords={\"x\": range(100), \"y\": range(100)})\n \n In [46]: ds[\"distance\"] = np.sqrt(ds.x**2 + ds.y**2)\n \n In [47]: ds.distance.where(ds.distance < 100).plot()\n- Out[47]: \n+ Out[47]: \n _\b[_\b__\bi_\bm_\ba_\bg_\be_\bs_\b/_\bw_\bh_\be_\br_\be_\b__\be_\bx_\ba_\bm_\bp_\bl_\be_\b._\bp_\bn_\bg_\b]\n * Added new methods xray.DataArray.diff and xray.Dataset.diff for finite\n difference calculations along a given axis.\n * New xray.DataArray.to_masked_array convenience method for returning a\n numpy.ma.MaskedArray.\n In [48]: da = xray.DataArray(np.random.random_sample(size=(5, 4)))\n \n"}]}]}]}]}]}