/pyquante2

Python Quantum Chemistry Reloaded. This is a rewrite of the standard PyQuante program to clean up things that have been nagging me.

Primary LanguagePythonOtherNOASSERTION

PyQuante2

PyQuante is an open-source suite of programs for developing quantum chemistry methods. The program is written in the Python programming language, but has many “rate-determining” modules also written in C for speed. The resulting code, though not as fast as Jaguar, NWChem, Gaussian, or MPQC, is much easier to understand and modify. The goal of this software is not necessarily to provide a working quantum chemistry program (although it will hopefully do that), but rather to provide a well-engineered set of tools so that scientists can construct their own quantum chemistry programs without going through the tedium of having to write every low-level routine.

Installation

I try to keep the required packages minimal (mostly just numpy). If you're building from source, you'll also need cython and a c compiler. However, I strongly recommend using conda.

Installation via conda

There are now conda installers for the full pyquante2 package on linux and mac, and a pure python version (that will be much slower) available on windows. To do this, just

% conda install -c rpmuller pyquante2

on linux/mac, or

% conda install -c rpmuller pyquante2_pure

on windows.

Why rewrite PyQuante?

PyQuante is a Quantum Chemistry Suite written in Python. Over the years, there are many things that I'm growing increasingly unsatisfied with:

  • I learned Python (essentially) as I was writing PyQuante, which means that I wasn't as good a programmer when I did most of the writing of the core modules as I am now.
  • Python was only at version 1.4-2.0 when I started writing. Many aspects of the language have changed substantially since then, especially the maturity of the numpy module.

I'm playing with PyQuante2 as a way of exploring more intelligent design. I'm slowly moving components over from PyQuante into PyQuante2. My goals are:

  • Clean code with a consistent style reflecting Python 2.7;
  • Easily understood code structure;
  • Much better test suite structure, use of nosetests as a matter of course.
  • Extensions
    • Written only in Cython
    • Not required for base code to function

Of course, it would be nice if this code were also much faster than PyQuante, but I'd be surprised if this happened automatically, and I'm prepared to deal with code performance at a later time.

There is an IPython notebook surveying some of the new features that can be viewed as a gist or on nbviewer.

What works so far:

  • Huzinaga and HGP integral methods, both in Python and with Cython wrappers to C
  • RHF, UHF wave functions
  • MP2 perturbation theory
  • A limited number of basis sets (STO-3G, 6-31G, 6-31G**)
  • Line and contour plotting
  • Basic IPython notebook support for some objects

Feel free to fork this if it interests you. The PyQuante code is still around, so I'm not rushing through the process, I'm just taking as much time as I feel I need to do this properly.