Getting Started

RASCAL is available to download in PyPi and GitHub. To install RASCAL, it is recommended to create a new environment to avoid possible conflicts with its required dependencies.

(base) $ conda create --name rascal_env python==3.10
(base) $ conda activate rascal_env

Required dependencies

RASCAL runs with Python 3.10.

These are the dependencies of RASCAL:

Installation via PyPi

RASCAL can be installed via PyPi:

(rascal_env) $ pip install rascal-ties

Installation via GitHub

RASCAL can be used via GitHub:

(rascal_env) $ git clone https://github.com/alvaro-gc95/RASCAL

The GitHub repository also contains the following scripts:

  • multiple_runs_example.py to automatize running several configurations of similarity methods and pool sizes for various stations and variables. This can be configured through the config.yaml file

  • projection_example.py Mostly the same as multiple_runs_example.py, but including a split in training and testing periods for the PCA, and an added year as a projection onto the training period PCs

  • RASCAL_evaluation.ipynb a Jupyter Notebook to plot and validate the reconstructions