Examples¶
Examples for how to use chemreps can also be found on GitHub under chemreps/examples as well as tried out on Binder.
Coulomb Matrix¶
The Coulomb Matrix representation was designed for the modeling of atomization energies and as a result requires information needed by the Hamiltonian, namely the set of Cartesian coordinates \({ R } _ { I }\) and nuclear charges \(Z _ { I }\). With this information the Coulomb matrix \(M\) is built
where the off-diagonal elements correspond to the Coulomb repulsion between \(I\) and \(J\). The expression for the diagonal elements is the result of an encoded polynomial fit to the atomic energies and nuclear charge.
DOI: 10.1103/PhysRevLett.108.058301
A Coulomb Matrix representation can be made below by importing coulomb_matrix from chemreps.coulomb_matrix and passing molecule file and size of the matrix as seen below:
from chemreps.coulomb_matrix import coulomb_matrix
mfile = '../data/sdf/butane.sdf'
cm = coulomb_matrix(mfile, size=14)
Bag of Bonds & other bagged representations¶
Bagged representations require the size of the bags and empty bags to be passed to the representation. The bag_sizes and bags can be made using BagMaker from chemreps.bagger if they are unknown or not yet made.
from chemreps.bagger import BagMaker
dataset = '../data/sdf/'
bagger = BagMaker('BoB', dataset)
bags = bagger.bags
bag_sizes = bagger.bag_sizes
Once the bags have been made, they can be passed to the representation function. Bag of Bonds (BoB) will be shown for this example, so bag_of_bonds is passed the newly created bags to create a BoB representation for the butane.sdf.
from chemreps.bag_of_bonds import bag_of_bonds
mfiles = '../data/sdf/butane.sdf'
rep = bag_of_bonds(mfiles, bagger.bags, bagger.bag_sizes)