search

choderalab / fahmunge

Tools for Munging Folding@Home datasets
MIT License
4 stars 6 forks source link

remove only water, NA, CL (not structural zincs!) for no-solvent #24

Closed rafwiewiora closed 8 years ago

rafwiewiora commented 8 years ago

Should fix https://github.com/choderalab/FAHMunge/issues/23 - only remove water and neutralizing ions for no-solvent trajs. I've only included NA and CL ions for removal - add others if need be.

Please review @jchodera

jchodera commented 8 years ago

We have to be a bit careful here: We should ideally also generate a reference PDB file in the root of the munged directory with all of the appropriate atoms in it, since we will need this for automated MSM construction (for @maxentile).

rafwiewiora commented 8 years ago

Ah, very good point. Let me add that!

rafwiewiora commented 8 years ago

Ok, added generation of the no-solvent PDBs for every RUN (in the same dir as the output h5's, i.e.. munged-with-time/11705/no-solvent/run0.pdb and munged-with-time/11705/no-solvent/run1.pdb.

Had to do these inside the strip_water() not the wrapper, so that we don't have all Workers checking for existence of these files and maybe there would be problem if multiple ones at start saw it wasn't there and all tried to write it.

Also some memory saving stuff in the wrapper - cut down the loaded trajectory to first frame only and del it before the stripping is done.

Looks good?

rafwiewiora commented 8 years ago

Btw tested on 2 trajectories - all ok.