D@X: ** DONE Support projects with files on different file systems Depending on storage criteria, an entire project may not be hosted on a single file systems accessible by nodes in a cluster. The solution is to abstract over file locations. This is accomplished by a Location base class that supports multiple protocols for resolving a file to the local machine. In the case of a local file, the resolution is trivial. However, if the file is on a different machine, there may be different ways to do so.

One protocol currently supported is the use of Chirp. By running a Chirp server on myhost.domain one could add the file to the current project like so

+BEGIN_SRC python

contains paths to project data

mylisting = 'datalist.dat' locations = dax.Location.from_file(mylisting)

idnum = 42 project = dax.Project('prefix', 'group', 'platform', idnum) project.load_locations(dax.read_cannonical, locations)

+END_SRC

Furthermore, Location objects implement the enter and exit methods allowing the use of with-statements to fetch the remote file and automatically clean up:

+BEGIN_SRC python

url = 'chirp://myhost.domain:port/path/to/RUNXXXX/CLONEYYYY/GENZZZZ' with dax.Location.location(url) as name: print 'Remote file', url, 'resolved to local', name

+END_SRC

M@X: ** DONE Transfer further arguments to gmx Use the [[https://github.com/badi/curry][curry]] module to partially apply functions that can still be pickled.

Example:

+BEGIN_SRC python

import dax, max, rax, gmx, curry

daxproj = dax.Project() raxproj = rax.Project()

pdb = 'protein.pdb' ndx = 'system.ndx' func = curry.curry(gmx.g_rms, s=pdb, n=ndx) data = daxproj.get_files('*.xtc', files=True) mapper = max.Mapper(func) mapper.process(data, raxproj)

raxproj.write()

+END_SRC

GMX ** DONE Provide functions that abstract can be curried with arguements before passing to a M@X Mapper.