arkouda-xarray-seasonal-avgs-example

An example of how to use Arkouda's Array-API module as the backend for XArray.

Follows this example from the XArray Gallery.

Dependencies

• Chapel
• Arkouda
• Anaconda

Instructions

After completing the installation instructions for Chapel and Arkouda, go through the following steps to get the Jupyter notebook in this repo running:

1. clone and cd into this repo

2. create and activate the conda environment to install Xarray and set up some of Arkouda and XArray's dependencies

   • conda env create -f arkouda-xarray-env.yaml
   • conda activate arkouda-xarray-env

3. install the arkouda-xarray package into the environment

   • clone the repo
   • cd into arkouda-xarray and pip install .

4. install Arkouda into the environment

   • cd into your local Arkouda directory
   • pip install .

5. build the Arkouda server to support multidimensional arrays

   • in your local Arkouda directory, edit serverConfig.json, setting "max_array_dims": 3
   • build the server for Array API compatibility: ARKOUDA_CONFIG_FILE=ArrayAPIServerModules.cfg make (this will take some time)

6. start the Arkouda server

   • ./arkouda_server
   • (or ./arkouda_server -nl4 if Chapel is configured for multinode builds)

7. run the Jupyter notebook

   • as it runs, you should see the server executing commands in your terminal

8. stop the Arkouda server