def prada1(
self, window=None, granularity=1.2, bins=20, ybins=10, segment=None, delta=None, clusters=True, verbose=False
):
if self.dimensions != 1:
print "# Method not implemented yet for more than 1D."
return
bins, mmx, mmn, delta = pyn_math.parameters_bins(self.traj, bins, segment, delta)
rv_min = 0
rv_max = 0
if mmn > self.traj.min():
rv_min = 1
bins += 1
if mmx < self.traj.max():
rv_max = 1
bins += 1
if verbose:
if rv_min:
print "# Extra node for x <", mmn
if rv_max:
print "# Extra node for x >", mmx
traj_aux = f_kin_anal.prada1(
ybins, bins, mmn, mmx, delta, rv_min, rv_max, self.traj, window, self.particles, self.frames
)
ranges = pyn_math.build_ranges(traj_aux)
self.network, self.traj_nodes = kinetic_network(traj_aux, ranges=ranges, traj_out=True, verbose=verbose)
del (traj_aux)
self.__offset__ = window
if clusters:
self.network.symmetrize(new=False, verbose=verbose)
self.network.mcl(granularity=granularity, pruning=True, verbose=verbose)
num_nodes = self.network.num_nodes
aux_list = numpy.empty(num_nodes, dtype=int, order="F")
for ii in range(num_nodes):
aux_list[ii] = self.network.node[ii].cluster
new_num_frames = self.traj_nodes.shape[0]
self.traj_clusters = f_kin_anal.trajnodes2trajclusters(
aux_list, self.traj_nodes, num_nodes, new_num_frames, self.particles
)
del (num_nodes, new_num_frames, aux_list)
self.__type_clusters__ = "prada1"
def prada2(self,window=None,granularity=1.2,bins=10,ybins=10,sbins=10,segment=None,delta=None,clusters=True,verbose=False):
ref_max=self.traj.max()
ref_min=self.traj.min()
if segment==None:
opt_range=0
mmx=ref_max
mmn=ref_min
else:
opt_range=1
mmn=segment[0]
mmx=segment[1]
if delta!=None:
opt=1
else:
delta=1.0 # Its given by gannas function
opt=2
bins,mmx,mmn,delta=pyn_math.parameters_bins(opt_range,opt,bins,mmn,mmx,delta)
traj_aux=f_trajs.prada2(ybins,sbins,bins,mmn,mmx,delta,self.traj,window,self.particles,self.frames)
ranges=pyn_math.build_ranges(traj_aux)
self.network,self.traj_nodes=kinetic_network(traj_aux,ranges=ranges,traj_out=True,verbose=False)
del(traj_aux)
self.__offset__=window
if clusters:
self.network.symmetrize(new=False,verbose=verbose)
self.network.mcl(granularity=granularity,pruning=True,verbose=verbose)
num_nodes=self.network.num_nodes
aux_list=numpy.empty(num_nodes,dtype=int,order='F')
for ii in range(num_nodes):
aux_list[ii]=self.network.node[ii].cluster
new_num_frames=self.traj_nodes.shape[0]
self.traj_clusters=f_trajs.trajnodes2trajclusters(aux_list,self.traj_nodes,num_nodes,new_num_frames,self.particles,self.dimensions)
del(num_nodes,new_num_frames,aux_list)
self.__type_clusters__='prada2'
def prada1_largo(
self,
window=None,
granularity=1.2,
bins=20,
ybins=10,
segment=None,
delta=None,
extra_min=False,
extra_max=False,
ram=4,
increment=1,
clusters=True,
verbose=False,
):
if self.dimensions != 1:
print "# Method not implemented yet for more than 1D."
return
bins, mmx, mmn, delta = pyn_math.parameters_bins(False, bins, segment, delta)
rv_min = 0
rv_max = 0
if extra_min:
rv_min = 1
bins += 1
if extra_max:
rv_max = 1
bins += 1
if verbose:
if rv_min:
print "# Extra node for x <", mmn
if rv_max:
print "# Extra node for x >", mmx
if not self.__tr_mode_in_file__:
"# Error: the trajectory must be in a file"
return
self.network = network(kinetic=True, verbose=False)
self.file_traj.open()
mmram = ram * 1024 * 1024 * 1024
iterations = int(mmram / (self.particles * bins * 8))
b_frame = window * increment
e_frame = self.file_traj.frames - 1 - window * increment
first_period = 1
salida = 1
if (b_frame + iterations * increment) > e_frame:
iterations = (e_frame - b_frame) / increment
while iterations > 0:
print b_frame + iterations * increment
traj_aux = f_kin_anal.prada1_infile(
self.file_traj.unit,
ybins,
bins,
segment[0],
delta,
rv_min,
rv_max,
b_frame,
iterations,
increment,
window,
self.particles,
self.dimensions,
)
ranges = pyn_math.build_ranges(traj_aux)
network_per = kinetic_network(traj_aux, ranges=ranges, traj_out=False, verbose=False)
self.network.merge_net(network_per, verbose=True)
del (traj_aux)
del (network_per)
b_frame += iterations * increment
if (b_frame + iterations * increment) > e_frame:
iterations = (e_frame - b_frame) / increment
# cerrar unidad
self.file_traj.close()
self.__offset__ = window
if clusters:
print "# Entra a clusters"
self.network.symmetrize(new=False, verbose=verbose)
self.network.mcl(granularity=granularity, pruning=True, verbose=verbose)
# num_nodes=self.network.num_nodes
# aux_list=numpy.empty(num_nodes,dtype=int,order='F')
# for ii in range(num_nodes):
# aux_list[ii]=self.network.node[ii].cluster
#
# new_num_frames=self.traj_nodes.shape[0]
# self.traj_clusters=f_kin_anal.trajnodes2trajclusters(aux_list,self.traj_nodes,num_nodes,new_num_frames,self.particles)
#
# del(num_nodes,new_num_frames,aux_list)
self.__type_clusters__ = "prada1"
