Implement AMBER formatted text format about mdtraj HOT 12 CLOSED

What exactly does "formatted text" mean? Is there a specific format in mind, like the xyz format?

from mdtraj.

AMBER can write trajectories in a couple of different formats:

1. Formatted text output (concatenated atom positions in a fixed text format):
   http://ambermd.org/formats.html#trajectory
2. NetCDF format (preferred now):
   http://ambermd.org/netcdf/nctraj.html

from mdtraj.

John, do you have a small short traj on each of the formats that you could upload to be part of the test suite?

-Robert
Sent from my iPhone.

On Sat, Apr 20, 2013 at 2:40 PM, jchodera [email protected]
wrote:

AMBER can write trajectories in a couple of different formats:

1. Formatted text output (concatenated atom positions in a fixed text format):
   http://ambermd.org/formats.html#trajectory
2. NetCDF format (preferred now):

http://ambermd.org/netcdf/nctraj.html

Reply to this email directly or view it on GitHub:
https://github.com/rmcgibbo/mdtraj/issues/27#issuecomment-16711834

from mdtraj.

I'm happy to generate some new examples of multiple trajectory formats for the same small system (e.g. alanine dipeptide) once we get our Amber license updated here at MSKCC, but I'll send along something I've dug up in the meantime.

from mdtraj.

Yeah, the trans only need to be a dozen frames long or so, this would be just for correctness testing, not performance benchmarks. And if you have the identical trajectory in two or more formats that would be optimal for testing.

-Robert
Sent from my iPhone.

On Sat, Apr 20, 2013 at 2:58 PM, jchodera [email protected]
wrote:

I'm happy to generate some new examples of multiple trajectory formats for the same small system (e.g. alanine dipeptide) once we get our Amber license updated here at MSKCC, but I'll send along something I've dug up in the meantime.

Reply to this email directly or view it on GitHub:
https://github.com/rmcgibbo/mdtraj/issues/27#issuecomment-16712098

from mdtraj.

Agreed. I'll see if I can produce this on one of the machines we have access to where Amber is installed.

from mdtraj.

@jchodera: I've added NetCDF support in #31.

from mdtraj.

@jchodera: do you really want formatted text support? The format is garbage.

from mdtraj.

@rmcgibbo: I agree the AMBER formatted text trajectory format is awful, but it's been around since the early days of AMBER in the 1990s. There's a ton of existing data in that format, and it takes extra effort to force AMBER to write NetCDF trajectories instead. So I think we're stuck supporting it, but should encourage people to use the NetCDF trajectory format whenever possible.

It could be worse---there's also a binary ("unformatted") output option that isn't even platform-portable! That one we should ignore---it's just a disaster and nobody uses it.

from mdtraj.

If it's something you'd like, feel free to prepare a pull request. I started writing a reader, but I don't think it correctly deals with variable box sizes.

from mdtraj.

OK, I'll try to find some time to tackle that over the weekend.

It's a pain. The number of box dimension variables has to be autodetected
based on the number of atoms.

Best,

John

On Wed, May 1, 2013 at 1:35 PM, Robert McGibbon [email protected]:

If it's something you'd like, feel free to prepare a pull request. I
started writing a reader https://gist.github.com/rmcgibbo/5498132, but
I don't think it correctly deals with variable box sizes.

—
Reply to this email directly or view it on GitHubhttps://github.com/rmcgibbo/mdtraj/issues/27#issuecomment-17304887
.

John D. Chodera
Assistant Faculty Member, Computational Biology
Memorial Sloan-Kettering Cancer Center
email: [email protected]
office: 646.888.3400
fax: 510.280.3760
mobile: 415.867.7384
url: http://www.choderalab.org

from mdtraj.

ok. done with last pr.

from mdtraj.