rascaline
=========

Rascaline is a library for the efficient computing of representations for
atomistic machine learning also called “descriptors” or “fingerprints”. These
representation can be used for atomistic machine learning (ML) models including
ML potentials, visualization or similarity analysis. You can learn how to use it
reading the `documentation <https://luthaf.fr/rascaline/>`_, and get the latest
version from the `github repository <https://github.com/Luthaf/rascaline>`_



.. grid:: 1 2 2 3
    :gutter: 1 1 2 3

    .. grid-item::
        .. card:: PCA/PCovR Visualization for the rattled GaAs training dataset
            :link: ../examples/gaas-map/gaas-map
            :link-type: doc
            :text-align: center
            :shadow: md

            .. image:: ../examples/gaas-map/images/thumb/sphx_glr_gaas-map_thumb.png
                :alt: This example uses rascaline and metatensor to compute structural properties for the structures in a training for a ML model. These are then used with simple dimensionality reduction algorithms (implemented in sklearn and skmatter) to obtain a simplified description of the dataset, that is then visualized using chemiscope.
                :class: gallery-img

                
    .. grid-item::
        .. card:: LPR analysis for amorphous silicon dataset
            :link: ../examples/lpr/lpr
            :link-type: doc
            :text-align: center
            :shadow: md

            .. image:: ../examples/lpr/images/thumb/sphx_glr_lpr_thumb.png
                :alt: In this tutorial, we calculate the SOAP descriptors of an amorphous silicon dataset using rascaline, then compute the local prediction rigidity (LPR) for the atoms of a "test" set before and after modifications to the "training" dataset has been made.
                :class: gallery-img

                
    .. grid-item::
        .. card:: Training the DOS with different Energy References
            :link: ../examples/dos-align/dos-align
            :link-type: doc
            :text-align: center
            :shadow: md

            .. image:: ../examples/dos-align/images/thumb/sphx_glr_dos-align_thumb.png
                :alt: This tutorial would go through the entire machine learning framework for the electronic density of states (DOS). It will cover the construction of the DOS and SOAP descriptors from ase Atoms and eigenenergy results. A simple neural network will then be constructed and the model parameters, along with the energy reference will be optimized during training. A total of three energy reference will be used, the average Hartree potential, the Fermi level, and an optimized energy reference starting from the Fermi level energy reference. The performance of each model is then compared.
                :class: gallery-img

                
    .. grid-item::
        .. card:: LODE Tutorial
            :link: ../examples/lode-linear/lode-linear
            :link-type: doc
            :text-align: center
            :shadow: md

            .. image:: ../examples/lode-linear/images/thumb/sphx_glr_lode-linear_thumb.png
                :alt: This tutorial explains how Long range equivariant descriptors can be constructed using rascaline and the resulting descriptors be used to construct a linear model with equisolve
                :class: gallery-img

                
    .. grid-item::
        .. card:: Generalized Convex Hull construction for the polymorphs of ROY
            :link: ../examples/roy-gch/roy-gch
            :link-type: doc
            :text-align: center
            :shadow: md

            .. image:: ../examples/roy-gch/images/thumb/sphx_glr_roy-gch_thumb.png
                :alt: This notebook analyzes the structures of 264 polymorphs of ROY, from Beran et Al, Chemical Science (2022), comparing the conventional density-energy convex hull with a Generalized Convex Hull (GCH) analysis (see Anelli et al., Phys. Rev. Materials (2018)). It uses features computed with rascaline and uses the directional convex hull function from scikit-matter to make the figure.
                :class: gallery-img

                
    .. grid-item::
        .. card:: Sample and Feature Selection with FPS and CUR
            :link: ../examples/sample-selection/sample-selection
            :link-type: doc
            :text-align: center
            :shadow: md

            .. image:: ../examples/sample-selection/images/thumb/sphx_glr_sample-selection_thumb.png
                :alt: In this tutorial we generate descriptors using rascaline, then select a subset of structures using both the farthest-point sampling (FPS) and CUR algorithms implemented in scikit-matter. Finally, we also generate a selection of the most important features using the same techniques.
                :class: gallery-img

                
    .. grid-item::
        .. card:: Periodic Hamiltonian learning
            :link: ../examples/periodic-hamiltonian/periodic-hamiltonian
            :link-type: doc
            :text-align: center
            :shadow: md

            .. image:: ../examples/periodic-hamiltonian/images/thumb/sphx_glr_periodic-hamiltonian_thumb.png
                :alt: This tutorial explains how to train a machine learning model for the electronic Hamiltonian of a periodic system. Even though we focus on periodic systems, the code and techniques presented here can be directly transferred to molecules.
                :class: gallery-img