Re: NAMD trajectory format

From: Peter Freddolino (petefred_at_ks.uiuc.edu)
Date: Fri May 30 2008 - 08:28:08 CDT

On a related note, I know that someone had wrapped the catdcd 3.0
readers for python (http://code.google.com/p/mdanalysis/); a similar
python wrapper for the molfile plugins would probably be the ideal
solution to this and many other problems.
Best,
Peter

Axel Kohlmeyer wrote:
> On Thu, 29 May 2008, Konrad Hinsen wrote:
>
> KH> On May 28, 2008, at 16:05, Axel Kohlmeyer wrote:
> KH>
> KH> >it is not that difficult and you could use the fortran wrapper code
> KH> >that i contributed as a starting point. OTOH, you can also just link
> KH> >the individual subroutines as a static library. see:
> KH> >
> KH> >http://www.ks.uiuc.edu/Research/vmd/plugins/molfile/
> KH> >
> KH> >and particularly:
> KH> >
> KH> >http://www.ks.uiuc.edu/Research/vmd/plugins/doxygen/callmolfileplugins.html#callmolfileplugins
> KH>
> KH> Thanks, I will look at that for use in the future. For my current needs, I
> KH> am quite happy with my DCD parser written in Python, which seems to work
> KH> fine, at least for the one trajectory that I have for testing.
>
> dcd trajectories are nasty to parse in general because they are fortran
> unformatted. depending on whether the file was written on a machine
> with the same or different endianness and a (64-bit) compiler that
> uses either 32-bit (most) or 64-bit (some versions of gfortran and
> g95, depending on flags) record markers and 64bit default integers
> you get a different file. john has put a lot of effort into making
> the dcd reader in molfile particularly flexible and efficient.
>
> there is a lot of convenience in writing your own stuff, but it also
> adds inconsistencies (just look at how many broken/non-compliant .pdb
> format writers and readers there are). with computational tools becoming
> increasingly ubiquitous in our research it is high time to consolidate
> on a few tools/libraries and use and improve those instead of
> re-inventing the same old wheels over and over again.
>
> that is why, for instance, i wrote those fortran wrappers to
> molfile. instead of adding just one more file formats to the
> (fortran based) analysis code of a colleague, i added _all_ the
> formats that molfile supports and any future file format work
> from me can focus on improving the molfile library. this in turn
> will also help every other project that uses it at the same time.
>
> cheers,
> axel.
>
> KH> Thanks again for your help,
> KH> Konrad.
> KH>
> KH> PS: If anyone wants a copy of my DCD parser in Python, just ask!
> KH> --
> KH> ---------------------------------------------------------------------
> KH> Konrad Hinsen
> KH> Centre de Biophysique Moléculaire, CNRS Orléans
> KH> Synchrotron Soleil - Division Expériences
> KH> Saint Aubin - BP 48
> KH> 91192 Gif sur Yvette Cedex, France
> KH> Tel. +33-1 69 35 97 15
> KH> E-Mail: hinsen_at_cnrs-orleans.fr
> KH> Web: http://dirac.cnrs-orleans.fr/~hinsen/
> KH> ---------------------------------------------------------------------
> KH>
> KH>
> KH>
>
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:49:31 CST