[MMTK] bulding water ice; problems with Peter's code

[Eugene Leitl]

Hi.

I am trying to build water ice (in fact, certain faces of it) with 
MMTK-2.0a5. As far as I see, there are two possibilites: using 
Peter McCluskey's MMTK scripts
http://www.rahul.net/pcm/mmtk/index.html , specifically modify
http://www.rahul.net/pcm/mmtk/Programs/hello.py and
translate/concatenate ice elementary cell (which I have as pdb 
coordinates), or use Lattice.py.

Unfortunately, Peter's scripts don't seem to run under the new version 
anymore. I've made some trivial fixes (reference to Amber94ForceField) 
but I'm stuck with the following error message:

eugene.leitl@lrz:~/download/mmtk/MMTK-2.0a5/Examples/peter > python hello.py
Traceback (innermost last):
  File "hello.py", line 5, in ?
    block1 = Group('diamond111')
NameError: Group

Peter, in case you've got some spare time (ha-ha), could you perhaps 
adjust the scripts to the new syntax? (It is somewhat difficult to
adapt other people's nonrunning code when one is learning both the 
language and the particular modeling package).

Also, I'm somewhat overwhelmed with Lattice.py. I have the following
data on water ice crystal cell:

space group I_h (no. 194 P6_{3/mmc})

unit cell constants:
a=b=4.516 Angstrom
c=  7.354 Angstrom
alpha=beta= 90 deg
gamma     = 120 deg

fractional coordinates:
O  (0.3333, 0.6667, 0.0629)
H_a(0.3333, 0.6667, 0.1989)
H_b(0.4551, 0.9102, 0.0182)

How do I use BravaisLattice for this?

Also, I am new to the Amber94 force field, having only used CHARMm
before (X-PLOR, EGO). Are there any dramatic differences?

What are your experiences with MMTK's numerical performance, btw?
 
TIA,

Eugene Leitl