Anaconda on the AIR-T¶

Conda is a powerful package and environment manager designed to simplify the installation and management of software dependencies across platforms. It enables users to create isolated environments for Python and non-Python packages, ensuring compatibility and reproducibility. For Deepwave's AIR-T, Conda is valuable for managing the complex dependencies of signal processing and machine learning tools, allowing streamlined deployment and consistent performance across development and production systems.

This tutorial will show you how to create an AirStack conda environment on the AIR-T. Conda comes pre-installed with a built-in local conda channels for the AIR-T drivers.

The following instructions apply to AirStack version 0.5.0+. For AirStack 0.4, see this tutorial.

Create AirStack Conda Environment¶

For the most up-to-date environment files for the AIR-T and information about creating custom conda environments that may include machine learning frameworks, please see our GitHub page here:

• https://github.com/deepwavedigital/airstack-examples

Clone the github repo¶

git clone https://github.com/deepwavedigital/airstack-examples.git

Select AirStack Version¶

Verify the AirStack version running on your device using:

SoapySDRUtil --find | grep -a driver_version

We will now create a conda environment called airstack using the airstack.yml file provided by Deepwave. You may modify this to fit your application.

Build Conda Environment¶

Run the following command on your AIR-T to create the environment from within the cloned github repository conda environment folder:

conda env create -f airstack.yml

Testing a New Conda Environment¶

Activate the conda environment and check that the current python version comes from a conda path:

$ conda activate airstack
$ which python
/home/<your_user>/.conda/envs/airstack/bin/python

Then run the hello_world.py example included in AirStack:

python /python/hello_world.py

Removing the Conda Environment¶

To remove the AirStack Conda environment, run the following command:

conda remove --name airstack --all