def step(self, mcs):
if self.pixel_tracker_plugin is None:
return
# Test modification of field values outside of core routines
mcs_change_rate = 10e3
if mcs % mcs_change_rate == 0:
for cell in self.cell_list:
new_val = random.random()
for ptd in self.get_cell_pixel_list(cell):
self.yes_fluc_comp[ptd.pixel.x, ptd.pixel.y, 0] = new_val
self.no_fluc_comp[ptd.pixel.x, ptd.pixel.y, 0] = new_val
# Without this call, modifications (other than by core routines) to a field with a solver using
# FluctuationCompensator will likely cause numerical errors
CompuCell.updateFluctuationCompensators()
def step(self, mcs):
"""
type here the code that will run every frequency MCS
:param mcs: current Monte Carlo step
"""
if self.pixel_tracker_plugin is None:
return
f = CompuCell.getConcentrationField(self.simulator, "F1")
# Test modification of field values outside of core routines
mcs_change_value = 1e3
if mcs == mcs_change_value:
id_of_cell_to_change = 1
val_after_change = 1.0
for ptd in self.get_cell_pixel_list(
self.fetch_cell_by_id(id_of_cell_to_change)):
self.total_bias += val_after_change - f[ptd.pixel.x,
ptd.pixel.y, 0]
f[ptd.pixel.x, ptd.pixel.y, 0] = val_after_change
# Without this call, modifications (other than by core routines) to a field with a solver using
# FluctuationCompensator will likely cause numerical errors
CompuCell.updateFluctuationCompensators()
# Do testing
mcs_min_test = 0
if mcs >= mcs_min_test:
# For all periodic boundary conditions and any initial distribution, changes in these values
# over simulation time should only be due to the accumulation of rounding errors in the
# FluctuationCompensator algorithm
total_ptp_medium = 0.0
total_man_medium = 0.0
total_man = 0.0
for ptd in self.pixel_tracker_plugin.getMediumPixelSet():
total_ptp_medium += f[ptd.pixel.x, ptd.pixel.y, ptd.pixel.z]
for x, y, z in self.every_pixel():
this_f = f[x, y, z]
total_man += this_f
if not self.cell_field[x, y, z]:
total_man_medium += this_f
self.plot_win1.add_data_point('ConcPTPMed', mcs, total_ptp_medium)
self.plot_win1.add_data_point('ConcManMed', mcs, total_man_medium)
self.plot_win1.add_data_point('ConcMan', mcs, total_man)
# This compares the total concentration counted by looping over the lattice (the old way) to the total
# concentration counted from medium pixel sites returned by pixel tracker (a new feature).
# Pixel tracker should be able to keep track of medium sites for single- and multi-threaded simulations
# so that looping over the entire lattice is no longer necessary.
# If pixel tracker isn't properly tracking medium sites, then errors will be non-zero here
err = 0.0
if total_man_medium != 0:
err = total_ptp_medium / total_man_medium - 1.0
self.plot_win2.add_data_point('ErrorConcPTPMed', mcs, err)
if self.total_man_ini is None:
self.total_man_ini = total_man
# This compares the current total concentration to the initial total concentration. FluctuationCompensator
# should be able to keep these values the same (neglecting machine errors) for domains with all periodic
# boundary conditions, even when field values are modified outside of core routines (Metropolis and solver).
# For FluctuationCompensator to account for modifications to any field outside of core routines, call
# updateFluctuationCompensators after all modifications for a step have been made.
# If a field is modified outside of core routines and updateFluctuationCompensators is not called,
# then errors generated by FluctuationCompensator can be detected here
# To test this functionality, make changes to field values, add all differences in field values due to
# the changes to self.total_bias and see what happens when updateFluctuationCompensators is/isn't called
err = 0.0
if self.total_man_ini != 0:
err = (total_man - self.total_bias) / self.total_man_ini - 1.0
self.plot_win2.add_data_point('ErrorConcManMod', mcs, err)