2. Getting Started¶

2.1.1. Running HORTON as a Python library¶

There will be many examples in the following sections demonstrating how HORTON can be used when writing your own Python scripts. These scripts should all start with the following lines:

#!/usr/bin/env python

# Import the HORTON library
from horton import *

# Import some other stuff (optional)
import numpy as np, h5py as h5, matplotlib.pyplot as pt

# Actual Python script

This header is then followed by some Python code that does the actual computation of interest. In such a script, you basically write the main program to do your calculations, exploiting the components that HORTON offers. The HORTON library is designed such that all its features are as modular as possible, allowing you to combine them in various ways.

Before running your script, say run.py, we recommend that you to make it executable (this needs to be done only once for every script):

chmod +x run.py

Now, when your script has completed, you can run it as follows:

./run.py

Do not use horton.py as your script name; this will cause trouble when loading the horton library (due to a namespace collision).

2.1.2. Running HORTON as horton-*.py scripts¶

The built-in HORTON scripts all have the horton-*.py filename pattern. Through command line arguments, one can control the actual calculations performed by these scripts. Basic information on how to use each built-in script can be obtained by using the --help flag. For example,

horton-convert.py --help

2.2.1. Atomic Units¶

Internally, HORTON works exclusively in atomic units. If you want to convert a value from a different unit to atomic units, multiply it with the appropriate unit constant, e.g. the following snippet sets length to 5 Angstrom and prints it in atomic units:

length = 5*angstrom    # recording 5 angstrom in atomic units
print length

Conversely, if you want to print a value in a different unit than atomic units, divide it by the appropriate constant. For example, the following prints an energy of 0.001 Hartree in kJ/mol:

energy = 0.001            # recording energy in Hartree
print energy/kjmol        # printing energy in kJ/mol

An overview of all units can be found in horton.units.

There are two special cases:

1. Angles are in radians, but you can use the deg unit to work with degrees, for example, 90*deg and np.pi/2 are equivalent.
2. Temperatures are in Kelvin.

2.2.2. Array Indexing¶

All arrays and list-like objects in Python use zero-based indexing. This means that the first element of a vector is accessed as follows:

vector = np.array([1, 2, 3])
print vector[0]

This convention also applies to all array-like (and list-like) objects in HORTON, e.g. the first orbital in a Slater determinant has index 0.

2.2.3. Loading/Dumping Data from/to a File¶

All input and output of data in HORTON is managed through the IOData class. To load data, call the from_file() method of the IOData class, e.g.:

mol = IOData.from_file('water.xyz')

The information read from the file is accessible through attributes of the mol object. For example, the following prints the coordinates of the nuclei in Angstrom:

print mol.coordinates/angstrom

To write data into a file, first create an instance of the IOData class, then set the appropriate attributes, and write the content into a file by calling the to_file() method of the IOData class. For example, the following snippet creates a .xyz file for a Neon atom:

mol = IOData(title='Neon')
mol.coordinates = np.array([[0.0, 0.0, 0.0]])
mol.numbers = np.array([10])
mol.to_file('neon.xyz')

For a complete list of supported input/output file formats please refer to Data file formats (input and output); this includes a list of IOData attributes supported by each file format. A definition of all possible IOData attributes can be found in horton.io.iodata.IOData.

2.2.4. Periodic Table¶

HORTON has a periodic table of elements alongside several atomic properties that may come in handy in computations. For more details please refer to horton.periodic. The following example prints some information for Carbon atom:

print periodic[6].mass/amu                # The mass in atomic mass units
print periodic['C'].cov_radius/angstrom   # The covalent radius in angstrom
print periodic['c '].c6                   # The C6 coefficient in atomic units

As demonstrated above, you can be relatively sloppy with the index when referring to elements of the periodic table.

2.2.5. A Complete Example¶

This first example is kept very simple in order to illustrate the basics of a HORTON Python script. (It neither performs an electronic calculation nor does post-processing.) This example loads a .xyz file and computes the molecular mass. Finally, it writes the data read from the .xyz file and the calculated mass into a .h5 file, using HORTON’s internal data format.

#!/usr/bin/env python

from horton import *

# Load the molecule from an ``.xyz`` file.
mol = IOData.from_file(context.get_fn('test/water.xyz'))

# Compute the molecular mass
mass = 0.0
#   Loop over all atomic numbers
for number in mol.numbers:
    mass += periodic[number].mass

# Print the mass in the amu unit
print 'MOLECULAR MASS [amu]: %.5f' % (mass/amu)

# Store the mass in the IOData instance, in order to write it to the file
mol.mass = mass

# Write data in the mol object to a file in HORTON's internal HDF5-based
# file format.
mol.to_file('water.h5')

Note that the context.get_fn('test/water.xyz') expression is used to look up a data file from the HORTON data directory. If you want to use your own file, load the molecule as follows:

mol = IOData.from_file('your_file.xyz')