Marugoto Installation and Recommended Project Setup

Installing and setting up marugoto

clone the marugoto directory from github to your local folder(referred to as $marugoto = ~/Documents/) via:
```
git clone https://github.com/KatherLab/marugoto
```
set up your local virtual environment in your (folder referenced to as $env = ~/Documents/venvs/)

python -m venv venv_name
Activate your environment

source $env/venv_name/bin/activate
To find out which cuda version is appropriate, run nvidia-smi to view the cuda version in the top right-hand corner:
Install Pytorch by using the command provided by the website after selecting the appropriate settings.
Confirm cuda is installed correctly by python and then in shell
```
import torch
torch.cuda.is_available()
```
This should return True
Go to $marugoto/marugoto/ and do

pip install .
marugoto should now be correctly installed and ready for use.

If your OS is installed on a smaller local drive with data stored on larger mounted drives, we recommend storing project data between the two drive as follows. If only a single drive is available, then all data will need to be stored there.

local drive mounted HDD

marugoto WSI’s

Scripts Tiles

Results Features
A good structure is to have a folder for each main project documents with a symbolic link, scripts and experiment results located within it (see below). To create a symbolic link, using the terminal when inside the project folder, run ln -s ~/Documents/marugoto/marugoto