PyQuante

I've been playing around with developing an improved interface to PyQuante of late. The old functions like rhf() and dft() had inconsistent requirements for input, and printed out inconsistent output. My first stab at the rewrite were the now defunct HFSolver(), UHFSolver(), etc., modules. But I didn't have enough flexibility to plug in different eigensolvers, for example, and have removed those functions in favor of the ones in PyQuante2. You can now import SCF() from PyQuante2, and then use that single function to run many different types of calculations. This is all still in development; in fact, the "2" in PyQuante2 refers to release 2.0, by which time I hope that we will have sorted out all of the issues remaining about better interfaces.

Everything is in the SVN development version, but this update was the major issue standing in the way of the 1.6 release, and so I anticipate making that release soon. Documentation is still incomplete, but snippets exist here and here, and the PyQuante/Tests/UnitSweet.py has many examples of how these functions can be used.

We've added some new basis sets (6-311G and friends, mostly). As of PyQuante 1.6, you can also import basis sets via

>>> solver = HFSolver(mol,basis="6-311G++(2d,2p)")

Noel O'Boyle has created a windows installer for PyQuante 1.5.1. Here's Noel's announcement from the mailing list:

Releases of PyQuante 1.5.1 for Pythons 2.4 and 2.5 on Windows are now
available for download from:
http://sourceforge.net/project/showfiles.php?group_id=43213&package_id=173119

Required:
(1) An installation of Python 2.4 or Python 2.5
(2) NumPy, available from http://numpy.scipy.org/ (click "Download NumPy").

Note:
(1) Make sure that the version of PyQuante and NumPy you download
matches the Python version (that is, "py2.4" for Python 2.4)

To test:
(1) Start a command prompt, and change directory to the Tests
subdirectory of PyQuante
(2) Run "python runalltests.py". Your output should be something like
the following:

>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

C:\PyQuante-1.5.1> C:\Python25\python runalltests.py

Energy of Be/OEP close to -7.98? ... ok
Energy of H2 (using Gaussians) close to -1.08? ... ok
Energy of He (using Gaussians) close to -2.855? ... ok
Energy of H2O close to -76.011751? ... ok
Energy of H2O (Mindo) close to -48.825159? ... ok
Energy of OH (Mindo) close to 18.128? ... ok

----------------------------------------------------------------------
Ran 6 tests in 23.589s

OK

I've rewritten the density matrix purification, and renamed the dmm.py module to DMP.py. Included is McWeeny purification, Canonical purification, trace correcting purification, and trace resetting purification. The Solvers.py module has tests that show how each of these functions can be called.

I've added a new module called Solvers.py to the SVN codebase that contains examples of using many different eigensolvers with PyQuante. PyQuante is small enough that the difference in speed between different eigensolvers is negligible. However, if you're interested in seeing how different solvers work, the module may interest you.

I've created a new set of web pages using RapidWeaver. Hopefully this will spur a new revolution in the PyQuante documentation.