def basic_test_suite():
""" Simple testing of various functions in the script."""
print "read_lammps_data_file"
lattice, type_list = lmp_io.read_lammps_data_file("CG_first_interaction.lammps05")
print "read_lammps_rdf"
rdf_average = lmp_io.read_lammps_rdf("tmp.1.rdf", 3, 1)
print "read_CG_log_file"
final_pressure = read_CG_log_file("new_CG.prod1.log")
print "smooth_data"
smoothed = dm.smooth_data(rdf_average[1][1])
print "read_in_rdf_file"
rdf_array, numofbins, cutoff = read_in_rdf_file("rdf", 3, [[1,1],[1,2],[1,3],[2,2],[2,3],[3,3]])
print "read_in_interaction_files"
distance, potential, derivative, pot_file_list, cutoff = lmp_io.read_in_interaction_files("./pot", 3, [[1,1],[1,2],[1,3],[2,2],[2,3],[3,3]], 1)
def basic_test_suite():
""" Simple testing of various functions in the script."""
print "read_lammps_data_file"
lattice, type_list = lmp_io.read_lammps_data_file(
"CG_first_interaction.lammps05")
print "read_lammps_rdf"
rdf_average = lmp_io.read_lammps_rdf("tmp.1.rdf", 3, 1)
print "read_CG_log_file"
final_pressure = read_CG_log_file("new_CG.prod1.log")
print "smooth_data"
smoothed = dm.smooth_data(rdf_average[1][1])
print "read_in_rdf_file"
rdf_array, numofbins, cutoff = read_in_rdf_file(
"rdf", 3, [[1, 1], [1, 2], [1, 3], [2, 2], [2, 3], [3, 3]])
print "read_in_interaction_files"
distance, potential, derivative, pot_file_list, cutoff = lmp_io.read_in_interaction_files(
"./pot", 3, [[1, 1], [1, 2], [1, 3], [2, 2], [2, 3], [3, 3]], 1)
def production():
""" This script will create the next interation of coarse-grained potentials using the Iterative Boltzmann Inversion to
match to a user-supplied radial distribution function (normally from atomistic simulation). It will also attempt a correction
for the pressure. The script will also extrapolate the potentials at low distance values by fitting to a soft Lennard-Jones
potential. Note, this fitting is somewhat unstable (CurveFit.pm) and can cause the IBI to fail. """
print "ARGUMENTS TO THE IBI ARE: ", sys.argv
# user-supplied arguments to the IBI. Note, not all of these arguments are required depending on what analysis is need and files are provided.
lammps_input_file = "" # LAMMPS input file for the current CG iteration.
correct_rdf_base = "" # User-supplied Radial Distribution Function to match to (normally derived from atomistic simulation) - distance is column 1 and the RDF is column 3.
potential_base = "" # the file base-name for the potential energy files. The format is such: pot.<iteration_number>.new.<type1><type2>. In this case the base-name is "pot".
number = 0 # the current IBI iteration number
lammps_data_file = "" # LAMMPS CG data file
lammps_rdf_file = "" # the CG RDF file if calculated by LAMMPS - this is a series of snapshot values, which need to be averaged.
p_target = 1.0 # pressure target for the CG simulation.
p_flag = 0.0 # flag to indicate whether to apply pressure correction - set to one if a pressure target is set by the user.
CG_output_file = "" # LAMMPS thermodynamic log file for the current CG simulation. Used to calculate the current average CG pressure.
p_now = 0 # current CG pressure read from (and averaged) the CG lammps thermodynamic log file;
temperature = 300 # temperature the simulations are run at; default is 300K
LJ_file_flag = 0 # if this flag is set to one, the parameters used in the extrapolation by fitting to a Lennard-Jones potential are read from a file (called LJ_parameters) rather than computed from fitting to the potential / forces.
num_of_bins = 0
DeltaR = 0.0
number_of_arguments = len(sys.argv)
mode = "default"
num_of_types = 0
for i in xrange(0, number_of_arguments):
if sys.argv[i].lower() == "lammps_input_file":
lammps_input_file = sys.argv[i+1]
print "THE LAMMPS INPUT FILE IS ", lammps_input_file
elif sys.argv[i].lower() == "lammps_output_file":
lammps_output_file = sys.argv[i+1]
print "THE LAMMPS OUTPUT FILE IS ", lammps_input_file
elif sys.argv[i].lower() == "lammps_data_file":
lammps_data_file = sys.argv[i+1]
print "THE LAMMPS DATA FILE IS ", lammps_data_file
elif ((sys.argv[i] == "potential_base") or (sys.argv[i] == "potential")):
potential_base = sys.argv[i+1]
elif sys.argv[i].lower() == "lammps_rdf_file":
lammps_rdf_file = sys.argv[i+1]
print "THE RDFS WILL BE READ FROM LAMMPS OUTPUT", lammps_rdf_file
elif (sys.argv[i] == "correct_rdf_base"):
correct_rdf_base = sys.argv[i+1]
print "THE RDFS TO MATCH TO HAVE THE FILE BASE ", correct_rdf_base
elif ((sys.argv[i] == "number") or (sys.argv[i] == "current_number") or (sys.argv[i] == "iteration_number")):
number = int(sys.argv[i+1])
print "THE CURRENT ITERATION NUMBER IS ", number
elif ((sys.argv[i] == "pressure_flag") or (sys.argv[i] == "p_flag")):
p_flag = float(sys.argv[i+1])
print "THE PRESSURE FLAG is ", p_flag
elif ((sys.argv[i] == "pressure_target") or (sys.argv[i] == "p_target")):
p_target = float(sys.argv[i+1])
if abs(p_flag) < 0.00001:
p_flag = 1
print "THE PRESSURE TARGET is ", p_target
elif ((sys.argv[i] == "CG_log_file") or (sys.argv[i] == "CG_logfile")):
CG_output_file = sys.argv[i+1]
p_now = read_CG_log_file(CG_output_file, label="Press")
#TODO: this is only a temp hack!
print "THE CURRENT PRESSURE WILL BE CALCULATED FROM THE LOG FILE ", CG_output_file , p_now
elif (sys.argv[i] == "temperature"):
temperature = float(sys.argv[i+1])
elif (sys.argv[i] == "LJ_param_file"):
LJ_file_flag = 1
elif sys.argv[i].lower() == "numofbins":
num_of_bins = int(sys.argv[i+1])
print "THE NUMBER OF BINS IS ", num_of_bins
elif sys.argv[i].lower() == "deltar":
DeltaR = float(sys.argv[i+1])
print "DeltaR IS ", DeltaR
elif sys.argv[i] == "mode":
mode = sys.argv[i+1]
elif sys.argv[i].lower() == "numoftypes":
num_of_types = int(sys.argv[i+1])
# read in the lammps data file to identify the number of CG types and lattice parameters.
lattice, type_list = lmp_io.read_lammps_data_file(lammps_data_file)
num_of_types = len(type_list)
print "Num of types = ", num_of_types
#num_of_types = 4
number_of_types_array = np.zeros((num_of_types+1))
for n in xrange(1, num_of_types+1):
number_of_types_array[n] = len(type_list["%s" % n])
if mode=="pressure_correct":
num_of_bins, cutoff, offset = lmp_io.get_number_of_bins_and_cutoff(potential_base, 1)
print "Potential numofbins and cutoff:", num_of_bins, cutoff
pots = (potential_base.strip()).split('.')
atom1 = int(pots[-2])
atom2 = int(pots[-1])
print "ATOMS are:", atom1, atom2
potential = np.zeros((num_of_bins+1))
volume = float(lattice[0]) * float(lattice[1]) * float(lattice[2])
hist_rdf = lmp_io.read_lammps_rdf(lammps_rdf_file, num_of_types, number)
pressure_pot = calc_pressure_correction(hist_rdf[atom1][atom2], num_of_bins, DeltaR, number_of_types_array[atom1], number_of_types_array[atom2], abs(p_flag), p_now, p_target, volume, temperature)
old_distance, old_potential, old_derivative = lmp_io.read_in_interaction_file(potential_base, num_of_bins)
potential = apply_pressure_correction(old_potential, pressure_pot, num_of_bins+1, old_distance[1], DeltaR)
potential[0]=potential[1] # TODO: change this temporary workaround into something more systematic. The workaround reduces anomalies in the derivative at the start of the potential.
new_derivatives = da.derivatives(np.arange(offset, cutoff, DeltaR), potential)
print "dy lens:", num_of_bins, len(new_derivatives), len(np.arange(offset-DeltaR, cutoff, DeltaR)), len(potential)
write_pot_file("%s/pot.%d.new.%d.%d" % (os.path.dirname(lammps_output_file), number+1, atom1, atom2), new_derivatives[1:] , potential[1:], num_of_bins, DeltaR, atom1, atom2, num_of_bins, offset, smoothing="no", selection="no") #note: we use an offset here!
elif mode=="default":
# Either read an interaction list from a file in the atom_dir (useful if you want to parametrize only a subset of interactions), or generate one on the fly.
interaction_filename = os.path.dirname(correct_rdf_base) + "/interaction_list"
if os.path.exists(interaction_filename):
interaction_list = interaction_list_from_file(interaction_filename)
else:
interaction_list = generate_interaction_list(num_of_types)
first_array, num_of_bins, cutoff2 = read_in_rdf_file(correct_rdf_base, num_of_types, interaction_list) # read in the rdfs to match to.
print "THE CUTOFF in the RDF files is", cutoff2, ", with", len(first_array[1][1])-1, "number of bins ";
print "THIS IS ITERATION NUMBER", number
deltaR = cutoff2 / num_of_bins # bin spacing from the RDF
previous_position, previous_potential, previous_derivative, old_pot_files, cutoff = lmp_io.read_in_interaction_files(potential_base, num_of_types,
interaction_list, number)
num_of_bins = int(cutoff / deltaR)
print deltaR, cutoff2, num_of_bins, correct_rdf_base
print "THE CUTOFF in the POS FILES is", cutoff, "and number of bins are", num_of_bins
# read in the RDFs of the CG calculated by LAMMPS.
hist_rdf = lmp_io.read_lammps_rdf(lammps_rdf_file, num_of_types, number)
# print lammps_rdf_file, len(hist_rdf[1][1])
DeltaR = cutoff / num_of_bins
# calculate the IBI
compute_update(os.path.dirname(lammps_output_file), first_array, hist_rdf, previous_position, previous_potential, num_of_types, number, DeltaR, num_of_bins, number_of_types_array, lattice, LJ_file_flag, p_flag, p_now, p_target, temperature, interaction_list)
# modify the lammps input file, ready for the next iteration
modify_lammps_in_file(lammps_input_file, lammps_output_file, number, interaction_list, num_of_types)
else:
print "ERROR: mode is incorrectly set in IBI.py. Should be e.g., default or pressure_correct"
sys.exit()
def production():
""" This script will create the next interation of coarse-grained potentials using the Iterative Boltzmann Inversion to
match to a user-supplied radial distribution function (normally from atomistic simulation). It will also attempt a correction
for the pressure. The script will also extrapolate the potentials at low distance values by fitting to a soft Lennard-Jones
potential. Note, this fitting is somewhat unstable (CurveFit.pm) and can cause the IBI to fail. """
print "ARGUMENTS TO THE IBI ARE: ", sys.argv
# user-supplied arguments to the IBI. Note, not all of these arguments are required depending on what analysis is need and files are provided.
lammps_input_file = "" # LAMMPS input file for the current CG iteration.
correct_rdf_base = "" # User-supplied Radial Distribution Function to match to (normally derived from atomistic simulation) - distance is column 1 and the RDF is column 3.
potential_base = "" # the file base-name for the potential energy files. The format is such: pot.<iteration_number>.new.<type1><type2>. In this case the base-name is "pot".
number = 0 # the current IBI iteration number
lammps_data_file = "" # LAMMPS CG data file
lammps_rdf_file = "" # the CG RDF file if calculated by LAMMPS - this is a series of snapshot values, which need to be averaged.
p_target = 1.0 # pressure target for the CG simulation.
p_flag = 0.0 # flag to indicate whether to apply pressure correction - set to one if a pressure target is set by the user.
CG_output_file = "" # LAMMPS thermodynamic log file for the current CG simulation. Used to calculate the current average CG pressure.
p_now = 0 # current CG pressure read from (and averaged) the CG lammps thermodynamic log file;
temperature = 300 # temperature the simulations are run at; default is 300K
LJ_file_flag = 0 # if this flag is set to one, the parameters used in the extrapolation by fitting to a Lennard-Jones potential are read from a file (called LJ_parameters) rather than computed from fitting to the potential / forces.
num_of_bins = 0
DeltaR = 0.0
number_of_arguments = len(sys.argv)
mode = "default"
num_of_types = 0
for i in xrange(0, number_of_arguments):
if sys.argv[i].lower() == "lammps_input_file":
lammps_input_file = sys.argv[i + 1]
print "THE LAMMPS INPUT FILE IS ", lammps_input_file
elif sys.argv[i].lower() == "lammps_output_file":
lammps_output_file = sys.argv[i + 1]
print "THE LAMMPS OUTPUT FILE IS ", lammps_input_file
elif sys.argv[i].lower() == "lammps_data_file":
lammps_data_file = sys.argv[i + 1]
print "THE LAMMPS DATA FILE IS ", lammps_data_file
elif ((sys.argv[i] == "potential_base")
or (sys.argv[i] == "potential")):
potential_base = sys.argv[i + 1]
elif sys.argv[i].lower() == "lammps_rdf_file":
lammps_rdf_file = sys.argv[i + 1]
print "THE RDFS WILL BE READ FROM LAMMPS OUTPUT", lammps_rdf_file
elif (sys.argv[i] == "correct_rdf_base"):
correct_rdf_base = sys.argv[i + 1]
print "THE RDFS TO MATCH TO HAVE THE FILE BASE ", correct_rdf_base
elif ((sys.argv[i] == "number") or (sys.argv[i] == "current_number")
or (sys.argv[i] == "iteration_number")):
number = int(sys.argv[i + 1])
print "THE CURRENT ITERATION NUMBER IS ", number
elif ((sys.argv[i] == "pressure_flag") or (sys.argv[i] == "p_flag")):
p_flag = float(sys.argv[i + 1])
print "THE PRESSURE FLAG is ", p_flag
elif ((sys.argv[i] == "pressure_target")
or (sys.argv[i] == "p_target")):
p_target = float(sys.argv[i + 1])
if abs(p_flag) < 0.00001:
p_flag = 1
print "THE PRESSURE TARGET is ", p_target
elif ((sys.argv[i] == "CG_log_file") or (sys.argv[i] == "CG_logfile")):
CG_output_file = sys.argv[i + 1]
p_now = read_CG_log_file(CG_output_file, label="Press")
#TODO: this is only a temp hack!
print "THE CURRENT PRESSURE WILL BE CALCULATED FROM THE LOG FILE ", CG_output_file, p_now
elif (sys.argv[i] == "temperature"):
temperature = float(sys.argv[i + 1])
elif (sys.argv[i] == "LJ_param_file"):
LJ_file_flag = 1
elif sys.argv[i].lower() == "numofbins":
num_of_bins = int(sys.argv[i + 1])
print "THE NUMBER OF BINS IS ", num_of_bins
elif sys.argv[i].lower() == "deltar":
DeltaR = float(sys.argv[i + 1])
print "DeltaR IS ", DeltaR
elif sys.argv[i] == "mode":
mode = sys.argv[i + 1]
elif sys.argv[i].lower() == "numoftypes":
num_of_types = int(sys.argv[i + 1])
# read in the lammps data file to identify the number of CG types and lattice parameters.
lattice, type_list = lmp_io.read_lammps_data_file(lammps_data_file)
num_of_types = len(type_list)
print "Num of types = ", num_of_types
#num_of_types = 4
number_of_types_array = np.zeros((num_of_types + 1))
for n in xrange(1, num_of_types + 1):
number_of_types_array[n] = len(type_list["%s" % n])
if mode == "pressure_correct":
num_of_bins, cutoff, offset = lmp_io.get_number_of_bins_and_cutoff(
potential_base, 1)
print "Potential numofbins and cutoff:", num_of_bins, cutoff
pots = (potential_base.strip()).split('.')
atom1 = int(pots[-2])
atom2 = int(pots[-1])
print "ATOMS are:", atom1, atom2
potential = np.zeros((num_of_bins + 1))
volume = float(lattice[0]) * float(lattice[1]) * float(lattice[2])
hist_rdf = lmp_io.read_lammps_rdf(lammps_rdf_file, num_of_types,
number)
pressure_pot = calc_pressure_correction(hist_rdf[atom1][atom2],
num_of_bins, DeltaR,
number_of_types_array[atom1],
number_of_types_array[atom2],
abs(p_flag), p_now, p_target,
volume, temperature)
old_distance, old_potential, old_derivative = lmp_io.read_in_interaction_file(
potential_base, num_of_bins)
potential = apply_pressure_correction(old_potential, pressure_pot,
num_of_bins + 1, old_distance[1],
DeltaR)
potential[0] = potential[
1] # TODO: change this temporary workaround into something more systematic. The workaround reduces anomalies in the derivative at the start of the potential.
new_derivatives = da.derivatives(np.arange(offset, cutoff, DeltaR),
potential)
print "dy lens:", num_of_bins, len(new_derivatives), len(
np.arange(offset - DeltaR, cutoff, DeltaR)), len(potential)
write_pot_file(
"%s/pot.%d.new.%d.%d" %
(os.path.dirname(lammps_output_file), number + 1, atom1, atom2),
new_derivatives[1:],
potential[1:],
num_of_bins,
DeltaR,
atom1,
atom2,
num_of_bins,
offset,
smoothing="no",
selection="no") #note: we use an offset here!
elif mode == "default":
# Either read an interaction list from a file in the atom_dir (useful if you want to parametrize only a subset of interactions), or generate one on the fly.
interaction_filename = os.path.dirname(
correct_rdf_base) + "/interaction_list"
if os.path.exists(interaction_filename):
interaction_list = interaction_list_from_file(interaction_filename)
else:
interaction_list = generate_interaction_list(num_of_types)
first_array, num_of_bins, cutoff2 = read_in_rdf_file(
correct_rdf_base, num_of_types,
interaction_list) # read in the rdfs to match to.
print "THE CUTOFF in the RDF files is", cutoff2, ", with", len(
first_array[1][1]) - 1, "number of bins "
print "THIS IS ITERATION NUMBER", number
deltaR = cutoff2 / num_of_bins # bin spacing from the RDF
previous_position, previous_potential, previous_derivative, old_pot_files, cutoff = lmp_io.read_in_interaction_files(
potential_base, num_of_types, interaction_list, number)
num_of_bins = int(cutoff / deltaR)
print deltaR, cutoff2, num_of_bins, correct_rdf_base
print "THE CUTOFF in the POS FILES is", cutoff, "and number of bins are", num_of_bins
# read in the RDFs of the CG calculated by LAMMPS.
hist_rdf = lmp_io.read_lammps_rdf(lammps_rdf_file, num_of_types,
number)
# print lammps_rdf_file, len(hist_rdf[1][1])
DeltaR = cutoff / num_of_bins
# calculate the IBI
compute_update(os.path.dirname(lammps_output_file), first_array,
hist_rdf, previous_position, previous_potential,
num_of_types, number, DeltaR, num_of_bins,
number_of_types_array, lattice, LJ_file_flag, p_flag,
p_now, p_target, temperature, interaction_list)
# modify the lammps input file, ready for the next iteration
modify_lammps_in_file(lammps_input_file, lammps_output_file, number,
interaction_list, num_of_types)
else:
print "ERROR: mode is incorrectly set in IBI.py. Should be e.g., default or pressure_correct"
sys.exit()