[CP2K:6048] Re: script to convert GROMACS tpr to psf/prm

[Peter Mamonov]

Hi Das,

Here is the fix:
https://github.com/pmamonov/tpr2psf/commit/f402ff4cfc51c91ca29cb4efbb2ea4553826125f

Update your local repository and give it a try.

Regards,
Peter

В Wed, 7 Jan 2015 11:49:23 -0800 (PST)
Santanu Das <sant... at iiserb.ac.in> пишет:

> Hi,
>  I have tried to create prm and psf file from a gromacs tpr file with 
> *tpr2psf.py* , from the link https://github.com/pmamonov/tpr2psf 
> and I encounter the error 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> *FF parameters    Bonds: 93    Angles: 38    Proper dihedrals: 176    
> Ryckaert-Belleman dihedrals: 0    Improper dihedrals: 25Traceback
> (most recent call last):  File "tpr2psf.py", line 400, in <module>    
> tpr.print_ff(prm)  File "tpr2psf.py", line 342, in print_ff    print
> >>f, "%.5f 0 %.4f" % (imp.k*j2cal/2, imp.psi)AttributeError: STUFF
> >>instance has 
> no attribute 'psi'*
> 
> can anyone tell me how to overcome ?
> thank you
> 
> <https://github.com/pmamonov/tpr2psf>
> On Saturday, August 28, 2010 5:21:54 PM UTC+5:30, Peter wrote:
> >
> > Dear All, 
> > Some time ago I wrote a (python) code to convert GROMACS binary TPR 
> > files into CP2K readable PSF (topology) and PRM (FF parameters)
> > files. Unfortunately I didn't achieve exact match between energies
> > calculated by GROMACS and CP2K/FIST. Respective energies calculated
> > by GROMACS and CP2K/FIST match in case of bonds, angles and
> > sometimes dihedrals. Nonbonded interactions (Coulomb and VdW, both
> > long range and 1-4) are different for most of tested systems.
> > Gradients calculated by both codes for some large system correlate
> > with covariance coefficient 0.4. At some point I considered this
> > result satisfactory and moved on. For now I don't have an intention
> > to further improve this conversion tool, but probably others will
> > find it to be a good starting point for creating such a conversion
> > tool or probably try it as-is. So I put an archive with script,
> > auxillary library (modified libgmx wrapper by Martin Hofling
> > [http://oldwww.gromacs.org/pipermail/gmx-developers/
> > 2009-March/003183.html
> > <http://oldwww.gromacs.org/pipermail/gmx-developers/2009-March/003183.html>])
> > and some sample files here:
> > http://groups.google.com/group/cp2k/web/gmx2psf.tar.bz2 
> >
> > Best regards, 
> > Peter Mamonov 
> > Lomonosov Moscow State University 
> > Biological faculty 
> > biophysics dept. 
> >
> > P.S. Here is a help message from script, which provides some 
> > additional info: 
> >
> > ABOUT: 
> > This script  reads GROMACS .tpr file and produces .psf and .prm
> > files readable by  CP2K[ http://cp2k.berlios.de/] and probably some
> > other MM software. 
> > This script uses modified version of GROMACS library python
> > wrapper, written by Martin Hofling
> > [http://oldwww.gromacs.org/pipermail/gmx-
> > developers/2009-March/003183.html
> > <http://oldwww.gromacs.org/pipermail/gmx-developers/2009-March/003183.html>]. 
> >
> >
> > NB: This script IS NOT capable of exact conversion between tpr and
> > psf/ prm topology/force field representation. Respective energies 
> > calculated by GROMACS and CP2K/FIST match in case of bonds, angles
> > and sometimes dihedrals. Nonbonded interactions (Coulomb and VdW,
> > both long range and 1-4) are different for most of tested systems. 
> > Gradients calculated by both codes for some large system correlate 
> > with covariance coefficient 0.4. 
> >
> > REQUIREMENTS: 
> > * GROMACS dynamic libraries (libgmx*.so) 
> > * Modified version of GromPy wrapper (distributed  with this
> > script) 
> >
> > USAGE: tpr2psf.py system.tpr 
> > Will produce system.psf and system.prm 
> >
> >
> >
>