This data set was used to examine the reactive dynamics and
spectroscopic properties of three beta-diketones: malonaldehyde (MA,
propandial), acetoacetaldehyde (AAA, 3- oxobutanal) and acetylacetone
(AcAc, pentan-2,4-dion), in a neural network based manner. The data
set contains different geometries for the three molecules as well as
for substructures (Amons [1]).

In total the data set contains 71208 geometries with reference
energies, forces and dipole moments calculated at the MP2/aug-cc-pVTZ
level of theory calculated using the MOLPRO [2].

For more details, see https://arxiv.org/abs/1911.09475


---------------------------------------------------------------------------------------
DATA SET FORMAT:

The data is saved in a compressed numpy binary file (.npz) and
contains a python dictionary with seven numpy arrays:

R: Cartesian Coordinates of atoms (in Angstrom [A]), (num_data, max_atoms, 3)

Q: Total charge (in elementary charges [e]), (num_data,)

D: Dipole moment vector with respect to the origin (in elementary charges times
   Angstrom [eA]), (num_data, 3)

E: Potential energy with respect to free atoms (in electronvolt [eV]), (num_data,)

F: Forces acting on atoms (in electronvolt per Angstrom [eV/A]), (num_data, max_atoms, 3)

Z: Atomic number of atoms, (num_data, max_atoms)

N: Number of atoms in structure (if N < max_atoms, the remaining entries are zero-padded)

It is important to note that the potential energies of the structures
are given with respect to free atoms. Thus, the constants listed below
are subtracted from the Molpro output for each occurence of the
corresponding atoms:

H: -0.499821176024 ha

C: -37.759560677467 ha

O: -74.959294147330 ha



---------------------------------------------------------------------------------------
ACCESS DATA SET:

The data set can be accessed using python:

>>> data = np.load("beta-diketones_71208.npz")

The different keywords of the python dictionary can be listed using

>>> data.files
>>>['Q', 'D', 'F', 'Z', 'R', 'E', 'N']

and the individual entries can be loaded using the appropriate
keyword, e.g. for the energy

>>> energies = data["E"]




---------------------------------------------------------------------------------------
[1] Huang, B.; von Lilienfeld, O.A., arXiv:1707.04146

[2] Werner, H.-J.; Knowles, P. J.; Knizia, G.; Manby, F. R.; Schütz, M.; et al.
    https://www.molpro.net, 2018.2.0