def getRDF(TrajName, Nbins, Cut = 20.0, Prefix = 'rdf', target_dir = os.getcwd(),
frame_start = 0, frame_stop = 260000, stepfreq = 1):
''' Makes a RDF from supplied sim trajectory
object. Only considers atoms within specified
Cut
'''
# Get un-mapped trajectory
Trj = getPickle(TrajName)
## Parse trajectory and make frame stepper
init = Trj[0]
BoxL = Trj.FrameData['BoxL']; BoxL = BoxL[0]
#check for non-periodic box 05/01/2015
if BoxL == 0:
BoxL = float(raw_input('BoxL = 0, found. Enter nonperiodic boxlength: '))
FrameRange = range(frame_start, frame_stop, stepfreq)
NFrames = len(FrameRange)
AtomTypes = Trj.AtomTypes
## Make bins
Bin_min = 1.00
Bin_max = Cut
Bin_delta = (Bin_max - Bin_min)/float(Nbins)
Bin_centers = np.zeros([Nbins], np.float64)
for i in range(Nbins):
Bin_centers[i] = Bin_min + (i+0.5)*Bin_delta
pb = sim.utility.ProgressBar(Text = 'Processing frame by frame...', Steps = NFrames)
g = np.zeros([Nbins, 3], np.float64) #1-MM #2-WW #3-MW
## Frame stepping
for frame in FrameRange:
Pos = Trj[frame]
# Call fortran routine to update histogram counts every frame
g = fortlib.find_rdf(g = g, bin_centers = Bin_centers, bin_delta = Bin_delta,
pos = Pos, atomtypes = AtomTypes, boxl = BoxL, atomtypedefs = atomtypedefs)
pb.Update(int(frame/stepfreq))
## Normalize the rdf and
rdf = {'bin_centers': Bin_centers, 'MM': '', 'WW': '', 'MW': ''}
N_mon = len(np.where(AtomTypes == monomer_atomtype)[0]); print N_mon
N_water = len(np.where(AtomTypes == water_atomtype)[0]); print N_water
print 'Normalizing g(r)...'
BoxVol = BoxL**3.
g /= NFrames
gofr = g * BoxVol
for j in range(Nbins):
r = Bin_centers[j] - 0.5*Bin_delta
next_r = Bin_centers[j] + 0.5*Bin_delta
gofr[j,:] /= (4.*np.pi/3.) * (next_r**3. - r**3.)
rdf['MM'] = {'g': gofr[:,0]/(N_mon * (N_mon - 1)/2.), 'count': g[:,0]}
rdf['WW'] = {'g' :gofr[:,1]/(N_water * (N_water - 1)/2.), 'count': g[:,0]}
rdf['MW'] = {'g': gofr[:,2]/(N_mon * N_water), 'count': g[:,0]}
# pickling data
pickleName = os.path.join(target_dir, Prefix + '.pickle')
pickle.dump(rdf, open(pickleName, 'w'))
def getLocalDensity(TrajName, LDAtomType, LDUpperCuts, Prefix = 'LD', target_dir = os.getcwd(), frame_start = 0, frame_stop = 260000, stepfreq = 1): ''' Calculates Local Density of central atomtype due to neighboring atom type and for different Upper Cuts. Currently used for estimating Local Densities of coarse grained SPC/E water model ''' # Get un-mapped trajectory Trj = getPickle(TrajName) ## Parse trajectory and make frame stepper init = Trj[0] BoxL = Trj.FrameData['BoxL']; BoxL = BoxL[0] FrameRange = range(frame_start, frame_stop, stepfreq) NFrames = len(FrameRange) AtomTypes = Trj.AtomTypes ## Make local density count array NCuts = len(LDUpperCuts) NAtom = len(AtomTypes) Central_atomtype = LDAtomType[0] Neigh_atomtype = LDAtomType[1] LD = np.zeros([NAtom, NCuts], np.float64) ## Precompute LD coefficients coeff = np.zeros([NCuts, 4]) for i, cut in enumerate(LDUpperCuts): coeff[i] = calcCoeff(cut) pb = sim.utility.ProgressBar(Text = 'Processing frame by frame...', Steps = NFrames) # Frame stepping for frame in FrameRange: Pos = Trj[frame] # Call fortran routine for per-frame calculation LD = fortlib.find_localdensity(ld = LD, pos = Pos, coeff = coeff, lduppercuts = LDUpperCuts, boxl = BoxL, atomtypes = AtomTypes, central_atomtype = Central_atomtype, neigh_atomtype = Neigh_atomtype) pb.Update(frame/stepfreq) ## Average over all the frames LD /= NFrames # select only the oxygens LD_oxygen = np.zeros([NAtom/3, NCuts], np.float64) print 'Selecing oxygen Local Densities...' count = 0 for i in range(0,NAtom-2,3): LD_oxygen[count, :] = LD[i,:] count += 1 # pickling data pickleName = os.path.join(target_dir, Prefix + '.pickle') pickle.dump((LDUpperCuts, LD_oxygen), open(pickleName, 'w'))
def getFirstShellWater(TrajName, LDUpperCuts, FirstShellCut = 6.0, Prefix = 'LD_wm', target_dir = os.getcwd(), frame_start = 0, frame_stop = 100, stepfreq = 100): ''' Calculates the local densities and first shell waters for each monomer for several different cutoffs. Also generates the monomer local density distribution. ''' # Get un-mapped trajectory Trj = getPickle(TrajName) ## Parse trajectory and make frame stepper init = Trj[0] BoxL = Trj.FrameData['BoxL']; BoxL = BoxL[0] FrameRange = range(frame_start, frame_stop, stepfreq) NFrames = len(FrameRange) AtomTypes = Trj.AtomTypes ## Make Local density count array NCuts = len(LDUpperCuts) N_mon = len(np.where(AtomTypes == 3)[0]) N_water = len(np.where(AtomTypes == 2)[0]) LD_mon = np.zeros([N_mon*NFrames,NCuts], np.float64) FirstShellWaters = np.zeros(N_mon*NFrames, np.float64) ## Precompute LD coefficients coeff = np.zeros([NCuts, 4], np.float64) for i, cut in enumerate(LDUpperCuts): coeff[i] = calcCoeff(cut) pb = sim.utility.ProgressBar(Text = 'Processing frame by frame...', Steps = NFrames) # Frame stepping count = 0 for frame in FrameRange: Pos = Trj[frame] # Call fortran subroutine for per-frame computation fsw = np.zeros(N_mon, np.float64) ld = np.zeros([N_mon, NCuts], np.float64) (ld, fsw) = fortlib.find_firstshellwaters(ld = ld, fsw = fsw, pos = Pos, atomtypes = AtomTypes, n_water = N_water, lduppercuts = LDUpperCuts, firstshellcut = FirstShellCut, coeff = coeff, boxl = BoxL, atomtypedefs = atomtypedefs) ## Average #LD_mon += ld #FirstShellWaters += fsw ## Per-frame FirstShellWaters[count:count+N_mon] = fsw for j in range(NCuts): LD_mon[count:count+N_mon,j] = ld[:,j] count += N_mon pb.Update(frame/stepfreq) #LD_mon /= NFrames #FirstShellWaters /= NFrames # pickling data pickleName = os.path.join(target_dir, Prefix + '.pickle') pickle.dump((LDUpperCuts, LD_mon, FirstShellWaters), open(pickleName, 'w'))
## Precompute LD coefficients coeff = np.zeros([NCuts, 4], np.float64) for i, cut in enumerate(LDUpperCuts): coeff[i] = calcCoeff(cut) pb = sim.utility.ProgressBar(Text = 'Processing frame by frame...', Steps = NFrames) # Frame stepping count = 0 for frame in FrameRange: Pos = Trj[frame] # Call fortran subroutine for per-frame computation fsw = np.zeros(NAtom, np.float64) ld = np.zeros([NAtom, NCuts], np.float64) (ld, fsw) = fortlib.find_spce_fsw(ld = ld, fsw = fsw, pos = Pos, atomtypes = AtomTypes, lduppercuts = LDUpperCuts, firstshellcut = FirstShellCut, coeff = coeff, boxl = BoxL) # pack per-frame data into global arrays count_o2 = 0 fsw_o2 = np.zeros(NWater, np.float64) ld_o2 = np.zeros([NWater,NCuts], np.float64) for i in range(0, NAtom,3): fsw_o2[count_o2] = fsw[i] ld_o2[count_o2,:] = ld[i,:] count_o2 += 1 FSW[count:count+NWater] = fsw_o2 for j in range(NCuts): LD[count:count+NWater,j] = ld_o2[:,j] count += NWater
