xarray_example.rst

Rendering XArray data with Vapor¶

Vapor supports render XArray data in a similar fashion to Numpy data In order to pass XArray data to Vapor, create a data set of the type vapor.dataset.PYTHON This can also be done with the convenience function Session.CreatePythonDataset(). You can add XArray variables as vapor variables to that dataset by using Dataset.AddXArrayData.

These variables can then be rendered normally using any of Vapor's renderers.

In [1]:

import example_utils
from vapor import session, renderer, dataset, camera
import xarray as xr
import numpy as np
from math import cos, sin, pi

ses = session.Session()
data = ses.CreatePythonDataset()

Warning: sysroot "/Applications/Xcode_12.4.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk" not found (ignoring for now).

Vapor 3.9.2
Python 3.9.19 (/Users/pearse/miniconda3/envs/readTheDocs)
OpenGL 4.1 Metal - 83.1

Below we generate an XArray variable that consists of a curvilinear 2D grid. You can also open an existing dataset with xarray.open_dataset and use variables in that dataset.

In [2]:

def gen2d(w,h,f):
    """Generate a 2D grid of size (w,h) by evaluating f(x,y) for every x,y coordinate"""
    ay = []
    for y in range(0,h):
        ax = []
        for x in range(0,w):
            ax += [f(x,y)]
        ay += [ax]
    return ay

w = h = 8
curveVar = xr.DataArray(
    np.random.randn(8, 8),
    dims=("x", "y"),
    coords={
        "x_coord": xr.DataArray(gen2d(8,8,lambda x,y: cos(y/(h-1)*pi) * (x+(w+1))), dims=("x", "y")),
        "y_coord": xr.DataArray(gen2d(8,8,lambda x,y: sin(y/(h-1)*pi) * (x+(w+1))), dims=("x", "y"))
    })

data.AddXArrayData("variable_name", curveVar)

In [3]:

# Create a renderer for the data

ren = data.NewRenderer(renderer.WireFrameRenderer)
ren.SetVariableName("variable_name")

In [4]:

# Show the rendering

ses.GetCamera().ViewAll()
ses.Show()

In [5]:

ses.DeleteRenderer(ren)

Below we generate an XArray variable that consists of a curvilinear 3D grid.

In [6]:

def gen3d(w,h,d,f):
    """Generate a 3D grid of size (w,h,d) by evaluating f(x,y,z) for every x,y,z coordinate"""
    az = []
    for z in range(0,d):
        ay = []
        for y in range(0,h):
            ax = []
            for x in range(0,w):
                ax += [f(x,y,z)]
            ay += [ax]
        az += [ay]
    return az


w = h = d = 8
curveVar = xr.DataArray(
    gen3d(w,h,d,lambda x,y,z: z*w*h + y*w + x),
    dims=("x", "y", "z"),
    coords={
        "x_coord": xr.DataArray(gen2d(w,h,lambda x,y: x), dims=("x", "y")),
        "y_coord": xr.DataArray(gen2d(w,h,lambda x,y: y), dims=("x", "y")),
        "z_coord": xr.DataArray(gen3d(w,h,d,lambda x,y,z: z), dims=("x", "y", "z")),
    })

data.AddXArrayData("variable_3d", curveVar)

In [7]:

# Create a renderer for the data

ren = data.NewRenderer(renderer.WireFrameRenderer)
ren.SetVariableName("variable_3d")

In [8]:

# Show the rendering

ses.GetCamera().ViewAll()
ses.Show()