Examples

This document provides examples for the core functionality of PODPAC as Simple Tutorials and interactive Jupyter Notebooks.

See Supported Datasets for links to specific dataset examples in the repository.

Simple

To follow along, open a Python interpreter or Jupyter notebook in the Python environment where PODPAC is installed.

# activate the PODPAC environment, using anaconda
$ conda activate podpac

# open jupyter lab
$ jupyter lab

# or start a ipython interpreter
$ ipython

Array of Data

This example builds a DataSource Node for data located on a 11 x 21 lat x lon grid and evaluates this DataSource at arbitrary geospatial coordinates.

In the Python interpreter of Jupyter notebook, import podpac and numpy as np:

In [1]: import podpac
In [2]: import numpy as np

Define a matrix (np.ndarray) of data values on a 11 x 21 grid using numpy:

# mock grid data
In [3]: data = np.random.rand(11, 21)
In [4]: data
Out[4]:
array([[0.23698237, 0.59569972, 0.76340507, 0.88851687, 0.01036634,
        0.87147232, 0.70486927, 0.48485402, 0.11070812, 0.01146267,
        0.89649864, 0.59089579, 0.11195743, 0.58360194, 0.15225759,
        0.99246553, 0.31122967, 0.80974094, 0.00474486, 0.00650152,
        0.08999056],
        ...
])

Define lat and lon coordinates that define the axis coordinates for this grid. We’ll assume the short side of the grid are lat coordinates (length 11) and the long side are lon coordinates (length 21).

# mock coordinates for data
In [5]: lat = np.linspace(40, 50, 11)
In [6]: lon = np.linspace(-10, 10, 21)

Assemble PODPAC Coordinates from these coordinate values. Note the order of the dims keyword must match the shape of our data.

# create native coordinates for data
In [7]: coords = podpac.Coordinates([lat, lon], dims=['lat', 'lon'])
In [8]: coords
Out[8]: coords
Coordinates (EPSG:4326)
        lat: ArrayCoordinates1d(lat): Bounds[40.0, 50.0], N[11]
        lon: ArrayCoordinates1d(lon): Bounds[-10.0, 10.0], N[21]

Create a PODPAC Array Node from data and coords. An Array Node is a sub-class of DataSource Node.

# create node for data source
In [9]: node = podpac.data.Array(source=data, coordinates=coords)
In [10]: node
Out[10]:
Array DataSource
        source:
[[0.23698237 0.59569972 0.76340507 0.88851687 0.01036634 0.87147232
  0.70486927 0.48485402 0.11070812 0.01146267 0.89649864 0.59089579
  0.11195743 0.58360194 0.15225759 0.99246553 0.31122967 0.80974094
  0.00474486 0.00650152 0.08999056]
    ...]]
        coordinates:
                lat: ArrayCoordinates1d(lat): Bounds[40.0, 50.0], N[11]
                lon: ArrayCoordinates1d(lon): Bounds[-10.0, 10.0], N[21]
        interpolation: nearest

We’ve successfully created a DataSource Node that describes an 11 x 21 grid of data values with lat and lon Coordinates.

Evaluate the Node at arbitrary coordinates:

# Create coordinates to evaluate this node
In [11]: other_coords = podpac.Coordinates([42.2, 5.7], dims=['lat', 'lon'])

# Retrieve the datapoint from the array
In [12]: output = node.eval(other_coords)
In [13]: output
<xarray.UnitsDataArray (lat: 1, lon: 1)>
array([[0.734135]])
Coordinates:
  * lat      (lat) float64 42.2
  * lon      (lon) float64 5.7
Attributes:
    layer_style:  <podpac.core.style.Style object at 0x000001FA86896F60>
    crs:          EPSG:4326

SMAP Data Source

Note: This example uses the SMAP node, which requires a NASA Earth Data Account with OpenDAP access configured.

# import the library
import podpac

# Create a SMAP Node
node = podpac.datalib.smap.SMAP(username=<your_username>, password=<your_password>)

# Create coordinates to evaluate this node
coords = podpac.Coordinates(['2018-01-01 12:00:00', 0, 0], dims=['time', 'lat', 'lon'])

# Retrieve the datapoint from NSIDC's OpenDAP server
output = node.eval(coords)

Notebooks

Interactive PODPAC examples are distributed as example Jupyter notebooks hosted in the creare-com/podpac-examples repository.

Download

You can download the notebooks two ways:

$ git clone https://github.com/creare-com/podpac-examples.git

Run Jupyterlab

If using the standalone Window 10 Installation, run the run_podpac_jupyterlab.bat script by double-clicking its icon.

If using a different installation:

  • Make sure the optional notebook or all optional dependencies are installed

# via pip
$ pip install podpac[notebook]
$ pip install podpac[all]

# from source
$ pip install -e .[notebook]
$ pip install -e .[all]

  • Start a new JupyterLab session

$ cd <podpac-examples>
$ jupyter lab

  • Browse the example notebooks directory <podpac-examples>/notebooks/

  • Open a notebook that you want to run

  • From the JupyterLab menu, select Run --> Run All

    Note: Not all the examples will work due to authentication or private resources (with unpublished URLs) for development.