instruction_Linux_Ubuntu

Setting up SPRITE on Linux Ubuntu

Tested releases:

• Linux Ubuntu 22.04 LTS

Important notes:

• If you are a code developer and plan to make changes of the code and push the changes back to the repository on GitHub, we assume you have a GitHub account and you know how to set up SSH keys on your operating system and add your SSH keys to your GitHub account.
• We have prepared training data that can be directly used for some data-driven models in SPRITE. It is not mandatory to have the existing training data for using the data-driven models in SPRITE. But generating high-quality training data which are large in file size would take a few hours. We encourage users to download our prepared training data to get familiar with using the data-driven models in SPRITE. The training data are stored in a separate zip file called SPRITE-Data.

Follow the Terminal command line instructions below. We use username to denote your user name on your operating system. Replace username with your user name.

Install the prerequisite software packages

Install Git and related packages:

sudo apt install git gitk git-lfs

Download Anaconda to your Downloads folder. The example installation file name is Anaconda3-2023.07-1-Linux-x86_64.sh. Your installation file can be a more recent version than in this example (2023.07-1). Use the actual file name to replace the one in the command line instruction below.

cd ~/Downloads/
chmod +x Anaconda3-2023.07-1-Linux-x86_64.sh
./Anaconda3-2023.07-1-Linux-x86_64.sh

Obtain the SPRITE repository from GitHub

Go to a preferred directory, e.g., Downloads.

cd ~/Downloads/

If you are an end user, use the https links to download the SPRITE repository

git clone https://github.com/idaholab/SPRITE.git

If you are a code developer, use the SSH links to download the SPRITE repository

git clone [email protected]:idaholab/SPRITE.git

Obtain SPRITE-Data from Box

Download the SPIRITE-Data.zip file from https://inlbox.box.com/s/2d51ythhga9knwv0euuig215khbel3lg

Unzip SPRITE-Data.zip and make sure the SPRITE-Data folder is located alongside the SPRITE folder.

Create and activate a custom Conda environment for SPRITE

Go to the SPRITE folder, e.g.,

cd ~/Downloads/SPRITE

Create a new environment named sprite, type

conda create -n sprite python=3.9

To activate the sprite environment, type

conda activate sprite

You can verify if the sprite environment is ready by typing

conda env list

You should see the following command line output.

# conda environments:
#
base                     /home/username/anaconda3
sprite                *  /home/username/anaconda3/envs/SPRITE

The sprite environment is empty. To install the required packages in the sprite environment, type

pip install -r requirements.txt

Now go back to How to use section.