def main():
global data2d
global As
# First I need to turn the assignments matrix into a 1D list of assignments
sys.stdout = os.fdopen(sys.stdout.fileno(),'w',0)
print "Reading in Assignments... from %s " % options.assFN
As = io.loadh(options.assFN)['arr_0'].astype(int)
print "Reading in data... from %s " % options.dataFN
try:
f = io.loadh( options.dataFN )
try:
data2d = f['arr_0']
except:
data2d = f['Data']
except:
data = load(options.dataFN)
proj = Project.load_from( options.projFN )
data2d = msmTools.reshapeRawData( data, proj )
print "Calculating averages for:"
pool = mp.Pool(options.procs)
clusters = range( As.max() + 1)
result = pool.map_async(calcAvg,clusters[:])
result.wait()
sol = result.get()
sol = array(sol)
savetxt(options.outFN, sol)
return
import os, sys, re
from scipy.optimize import curve_fit
import matplotlib.pyplot as plt
import numpy as np
import warnings
warnings.filterwarnings('ignore','Warning: overflow encountered in exp')
def f( x , a, b ):
return np.exp( - b * x )
print "Loading data"
proj = Serializer.Serializer.LoadFromHDF( options.proj_FN )
data = dataIO.readData( options.raw_FN )
data2d = msmTools.reshapeRawData( data, proj )
print "Calculating autocorrelations"
Autos = [ autocorrelate.fft_autocorrelate( trj[ np.where( trj != -1 ) ] ) for trj in data2d ]
print "Fitting the data to single exponentials"
Fits = [ curve_fit( f, np.arange( len( corr ) ), corr )[0] for corr in Autos ]
outName = '.'.join( options.out_FN.split('.')[:-1] )
Fits = np.array( Fits )
np.savetxt( outName + '.dat', Fits )
print "Plotting some fits"
indList = np.random.permutation( np.arange( len( Autos ) ) )
def main():
# Need to construct the list of trajframes to use. First look at what the values are for the ones already done.
OldTrajFrames = []
for tempDir in dirs2check:
tempList = [ fn for fn in os.listdir( tempDir ) if re.search( 'trj\d+_frm\d+$', fn ) ]
tempTrajFrames = [ re.search( 'trj(\d+)_frm(\d+)$', fn ).groups() for fn in tempList ]
tempTrajFrames = [ [ int( a ), int( b ) ] for (a,b) in tempTrajFrames ]
OldTrajFrames.extend( tempTrajFrames )
Xdat = dataIO.readData( options.x_dat )
Ydat = dataIO.readData( options.y_dat )
uniqX = unique( Xdat )
uniqY = unique( Ydat )
diffX = abs( uniqX[1:] - uniqX[:-1] ).min()
diffY = abs( uniqY[1:] - uniqY[:-1] ).min()
Nx = 1. / diffX
Ny = 1. / diffY
Xdat = Xdat * ( Nx )
Ydat = Ydat * ( Ny )
if len( Xdat.shape ) > 1:
Xdat = Xdat[:,0]
if len( Ydat.shape ) > 1:
Ydat = Ydat[:,0]
Xdat2D = msmTools.reshapeRawData( Xdat.astype(int), Proj )
Ydat2D = msmTools.reshapeRawData( Ydat.astype(int), Proj )
x_interval = [ float( i ) for i in options.x_int.split(',') ]
y_interval = [ float( i ) for i in options.y_int.split(',') ]
x_range = arange( int( x_interval[0] * Nx ), int( x_interval[1] * Nx ) + 1 )
y_range = arange( int( y_interval[0] * Ny ), int( y_interval[1] * Ny ) + 1 )
#print x_range, y_range
#print Xdat.max(), Ydat.max()
#print OldTrajFrames
# Now construct the list of traj frames to use in the analysis
TrajFrames = []
print "Finding conformations..."
for Xi in x_range:
for Yi in y_range:
whichTrajFrames = array( where( ( Xdat2D == Xi ) * ( Ydat2D == Yi) ) ).T
if len( whichTrajFrames ) <= 4:
ToAddTrajFrames = [ list( i ) for i in whichTrajFrames ]
else:
ToAddTrajFrames = [ list( pair ) for pair in whichTrajFrames if list( pair ) in OldTrajFrames ]
# The above list contains pairs for this x,y for which the simulations have already been done.
if len( ToAddTrajFrames ) > 4:
ToAddTrajFrames = [ list( pair ) for pair in random.permutation( ToAddTrajFrames )[:4] ]
while len( ToAddTrajFrames ) < 4:
randPair = whichTrajFrames[ random.randint( len( whichTrajFrames ) ) ]
randPair = list( randPair )
# print '\t', randPair, ToAddTrajFrames
if randPair in ToAddTrajFrames:
continue
else:
ToAddTrajFrames.append( randPair )
TrajFrames.extend( ToAddTrajFrames )
# print Xi, Yi, ToAddTrajFrames, [ ( Xdat2D[ tuple(pair) ], Ydat2D[ tuple(pair) ] ) for pair in ToAddTrajFrames ]
print TrajFrames
print "Running the simulations..."
run( TrajFrames )
from pyschwancr import dataIO, msmTools import os, sys, re Ass = Serializer.LoadData( args.ass_FN ) Proj = Serializer.LoadFromHDF( args.proj_FN ) if (Ass.max()+1) > 100: print "You have %d states... This is going to be a large pdf file..." % (Ass.max()+1) pp = PdfPages( args.out_FN ) Data = dataIO.readData( args.data_FN ) if len(Data.shape) == 1: Data = msmTools.reshapeRawData( Data, Proj ) x0 = 0 x1 = int(Data.max()+1) if args.x_lbl != None: x_lbl = ' '.join( args.x_lbl ) else: x_lbl = 'Data in State' for i in range( Ass.max() + 1 ): figure() state_dat = Data[ np.where( Ass == i ) ] hist( state_dat, bins=50, log=True,histtype='step',lw=2,range=(x0,x1) )
matplotlib.use("pdf")
from matplotlib.pyplot import *
from pyschwancr import dataIO, msmTools
import os, sys, re
print "Loading Data"
Xdat = dataIO.readData(options.x_FN)[:, 0]
Ydat = dataIO.readData(options.y_FN)[:, 0]
Proj = Serializer.Serializer.LoadFromHDF(options.proj_FN)
Ass = Serializer.LoadData(options.ass_FN)
FCperState = dataIO.readData(options.fc_FN)
print Xdat.shape, Ydat.shape
Xdat = msmTools.reshapeRawData(Xdat, Proj)
Ydat = msmTools.reshapeRawData(Ydat, Proj)
keptAssInd = np.where(Ass != -1)
Xdat = Xdat[keptAssInd]
Ydat = Ydat[keptAssInd]
FCdat = FCperState[Ass[keptAssInd]]
# The above is all the data, but now we need to average the FC's for conformations at the same point in the projection.
if Xdat.shape[0] != Ydat.shape[0] or Ydat.shape[0] != FCdat.shape[0]:
print "Data is not the same shape! %d, %d, %d" % (Xdat.shape[0], Ydat.shape[0], FCdat.shape[0])
nx = len(np.unique(Xdat))
ny = len(np.unique(Ydat))
Z = np.zeros((nx, ny))
