def extract_wrapped(simname, format_spec=['id', 'type', 'xs', 'ys', 'zs']):
"""Extract coordinates of all atoms from wrapped trajectory files with
format 'id','type','xs','ys','zs'
Args:
simname (str): name of the LAMMPS simulation.
Returns:
coord (dict): Dictionary with the number of timesteps as keys and
coordinates of all atoms at the corresponding timestep as
a pandas dataframe. The format of keys is:
'timestep_<timestep>'
For ex:
'timestep_0'
"""
dump_wrapped, dump_unwrapped, dump_def1,dump_def2,\
dump_def3, log_file=fileIO.retrieve_different_filetypes(simname)
path = fileIO.findInSubdirectory(dump_wrapped)
unwrap = pd.read_csv(path, header=None, index_col=False)
index = 0
natoms_notfound = True
while natoms_notfound:
try:
if unwrap.iloc[index].str.split()[0][1] == 'NUMBER':
natoms_notfound = False
else:
index += 1
except IndexError:
index += 1
natoms = int(unwrap.iloc[index + 1].str.split()[0][0])
natoms = int(unwrap.iloc[index + 1].str.split()[0][0])
index = 0
timestep = 0
coord = {}
while index < len(unwrap):
line = unwrap.iloc[index].str.split()
try:
if line[0][0] == 'ITEM:' and line[0][1] == 'ATOMS':
length = len(unwrap.iloc[index + 1].str.split()[0])
df2 = unwrap.iloc[index + 1:index + natoms + 1]
df2 = df2[0].str.split(' ', length - 1, expand=True)
#Split based on separator - expensive
num2str = lambda x: float(x)
#convert all elements from str to float
df2 = df2.applymap(num2str)
#apply num2str to every element - expensive
df2.columns = format_spec
#add corresponding column labels
df2 = df2.sort_values(by=['id'])
#sort based on atom id so that future operations are easy.
key = 'timestep_' + str(timestep)
#save in the corresponding dictionary.
coord[key] = df2
index = index + natoms
timestep += 1
else:
index += 1
except IndexError:
index += 1
return coord
def read_boxsize_generic(simname, nc, dp, last_timestep=173):
"""Reads the box size of the simulation from the unwrapped coordinates
trajectory.
Args:
simname (str): name of the LAMMPS simulation.
Returns:
prop_constant (np array): an array containing the proportionality
constant pertaining to the increase in size of
the box (along x, y and z) at every timestep.
"""
dump_wrapped, dump_unwrapped, dump_def1, dump_def2,\
dump_def3, log_file=fileIO.retrieve_different_filetypes(simname)
path = fileIO.findInSubdirectory(dump_unwrapped)
unwrap = pd.read_csv(path, header=None, index_col=False)
index = 0
timestep = 1
prev_index = -(nc * dp + 4)
next_x_add = nc * dp + 9
next_y_add = nc * dp + 10
next_z_add = nc * dp + 11
ref_size_x = (
float(unwrap.iloc[prev_index + next_x_add].str.split()[0][1]) -
float(unwrap.iloc[prev_index + next_x_add].str.split()[0][0]))
print(ref_size_x)
ref_size_y = (
float(unwrap.iloc[prev_index + next_y_add].str.split()[0][1]) -
float(unwrap.iloc[prev_index + next_y_add].str.split()[0][0]))
ref_size_z = (
float(unwrap.iloc[prev_index + next_z_add].str.split()[0][1]) -
float(unwrap.iloc[prev_index + next_z_add].str.split()[0][0]))
prop_constant = [None] * last_timestep
while timestep <= last_timestep:
curr_size_x = (
float(unwrap.iloc[prev_index + next_x_add].str.split()[0][1]) -
float(unwrap.iloc[prev_index +
next_x_add].str.split()[0][0])) / ref_size_x
curr_size_y = (
float(unwrap.iloc[prev_index + next_y_add].str.split()[0][1]) -
float(unwrap.iloc[prev_index +
next_y_add].str.split()[0][0])) / ref_size_y
curr_size_z = (
float(unwrap.iloc[prev_index + next_z_add].str.split()[0][1]) -
float(unwrap.iloc[prev_index +
next_z_add].str.split()[0][0])) / ref_size_z
px = round(curr_size_x, 2)
py = round(curr_size_y, 2)
pz = round(curr_size_z, 2)
prop_constant[timestep - 1] = [px, py, pz]
prev_index += next_x_add
timestep += 1
return prop_constant
def test_retrieve_filetypes(self):
"""Asserts file type returns for the given file"""
rootname = 'test_file'
dump_wrapped = rootname + '.lammpstrj'
dump_unwrapped = rootname + '.unwrapped.lammpstrj'
dump_def1 = rootname + '.def1.txt'
dump_def2 = rootname + '.def2.txt'
dump_def3 = rootname + '.def3.txt'
log_file = 'log.' + rootname + '.txt'
self.assertEqual(fileIO.retrieve_different_filetypes('test_file'),
(dump_wrapped, dump_unwrapped, dump_def1, dump_def2,
dump_def3, log_file))
def extract_log_thermo(simname):
"""Read and return the thermodynamic data from a LAMMMPS simulation file.
Args:
simname (str): name of the simulation.
Returns:
log_thermo (dict): Dictionary containing the number of runs in the current
file and corresponding pandas dataframe containing all
thermodynamic output regarding that run file.
"""
dump_wrapped, dump_unwrapped, dump_def1, dump_def2,\
dump_def3, log_file=fileIO.retrieve_different_filetypes(simname)
path = fileIO.findInSubdirectory(log_file)
df3 = pd.read_csv(path,
sep='nevergonnahappen',
engine='python',
index_col=False,
header=None)
natoms = []
log_thermo = {}
run_num = 0
for index in df3.index:
line = df3.iloc[index].str.split()
try:
if line[0][0] == 'Per' and line[0][1] == 'MPI':
try:
float(df3.iloc[index + 3].str.split()[0][0])
#To make sure we do not include "run 0" runs.
col_names = df3.iloc[index + 1].str.split()[0]
run_num += 1
#%%%%%%%%%%%%% FIND RUN and THERMO numbers %%%%%%%%%%%%%%%
run_index = index - 1
while df3.iloc[run_index].str.split()[0][0] != 'run':
run_index -= 1
run = df3.iloc[run_index].str.split()[0][1]
thermo_index = run_index - 1
while df3.iloc[thermo_index].str.split()[0][0] != 'thermo':
thermo_index -= 1
thermo = df3.iloc[thermo_index].str.split()[0][1]
output_num = int(int(run) / int(thermo))
natoms.append(df3.iloc[index + output_num +
3].str.split()[0][-2])
#%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
#%%%%%%%%%%%%%%%% Read thermo data %%%%%%%%%%%%%%%%%%%%%%%
df3.iloc[index + 2:index + output_num + 3].to_csv(
r'.//test.csv', index=False, header=None)
#write dummy
# file
df2 = pd.read_csv('./test.csv',
delim_whitespace=True,
names=col_names,
dtype=np.float64)
#read dummy file
key = 'run_' + str(run_num)
#save in the corresponding dictionary.
log_thermo[key] = df2
#%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
except ValueError:
pass
except IndexError:
pass
return log_thermo
def extract_def(simname, syntax='18f'):
"""
This function reads the given LAMMPS simulation name (if present somewhere
in the current directory or subdirectories) and returns a numpy array
containing the file's output (log, lammpstrj, def1, def2).
*def* files are usually created by Raiter during deformations. They are
clutter-free and contain only numbers. This function will NOT work for log
files. For log files, please check extract_log_thermo function.
Args:
simname (str): name of the *def* file
syntax (str): the format specification of the requested def file. Default
is '18f'
Returns:
def1 (dataframe): Pandas dataframe containing the output from the *def1*
file.
def2 (dataframe): Pandas dataframe containing output from the *def2* file.
"""
dump_wrapped, dump_unwrapped, dump_def1,dump_def2,\
dump_def3, log_file=fileIO.retrieve_different_filetypes(simname)
path = fileIO.findInSubdirectory(dump_def1)
if syntax == '18f':
column_names = [
'strain', 'pxx', 'pyy', 'pzz', 'lx', 'ly', 'lz', 'temp', 'epair',
'ebond', 'eangle', 'edihed', 'ecoul', 'evdwl', 'etotal', 'pe',
'ke', 'density'
]
elif syntax == 'old':
column_names = [''] #FILL IN LATER
def1 = pd.read_csv(path,
delim_whitespace=True,
skiprows=1,
dtype=np.float64,
names=column_names,
index_col=False) #Read and skip the first line
path = fileIO.findInSubdirectory(dump_def2)
if syntax == '18f':
column_names = [
"strain", "c_meansquarep[1]", "c_meansquarep[2]",
"c_meansquarep[3]", "c_meansquarep[4]", "c_meansquaren[1]",
"c_meansquaren[2]", "c_meansquaren[3]", "c_meansquaren[4]",
"c_meansquarei[1]", "c_meansquarei[2]", "c_meansquarei[3]",
"c_meansquarei[4]", "c_meansquareu[1]", "c_meansquareu[2]",
"c_meansquareu[3]", "c_meansquareu[4]", "c_nongaussp[1]",
"c_nongaussp[2]", "c_nongaussp[3]", "c_nongaussn[1]",
"c_nongaussn[2]", "c_nongaussn[3]", "c_nongaussi[1]",
"c_nongaussi[2]", "c_nongaussi[3]", "c_nongaussu[1]",
"c_nongaussu[2]", "c_nongaussu[3]", "c_msdall[1]", "c_msdall[2]",
"c_msdall[3]", "c_msdall[4]", "c_ngpall[1]", "c_ngpall[2]",
"c_ngpall[3]"
]
def2 = pd.read_csv(path,
delim_whitespace=True,
skiprows=1,
dtype=np.float64,
names=column_names,
index_col=False)
return def1, def2
def extract_unwrapped(simname,
format_spec=['id', 'mol', 'type', 'xu', 'yu', 'zu'],
first_only=False,
last_only=False,
boxsize=False,
boxsize_whole=False,
start_end=False):
"""Extract coordinates of all atoms from unwrapped trajectory files with
format 'id','mol','type','xu','yu','zu'
Args:
simname (str): name of the LAMMPS simulation.
Returns:
coord (dict): Dictionary with the number of timesteps as keys and
coordinates of all atoms at the corresponding timestep as
a pandas dataframe.
"""
dump_wrapped, dump_unwrapped, dump_def1,dump_def2,\
dump_def3, log_file=fileIO.retrieve_different_filetypes(simname)
path = fileIO.findInSubdirectory(dump_unwrapped)
unwrap = pd.read_csv(path, header=None, index_col=False)
index = 0
natoms_notfound = True
while natoms_notfound:
try:
if unwrap.iloc[index].str.split()[0][1] == 'NUMBER':
natoms_notfound = False
else:
index += 1
except IndexError:
index += 1
natoms = int(unwrap.iloc[index + 1].str.split()[0][0])
first_timestep_lines = natoms + 9
index = 0
timestep = 0
coord = {}
if first_only:
while index < first_timestep_lines:
line = unwrap.iloc[index].str.split()
try:
if line[0][0] == 'ITEM:' and line[0][1] == 'ATOMS':
length = len(unwrap.iloc[index + 1].str.split()[0])
df2 = unwrap.iloc[index + 1:index + natoms + 1]
df2 = df2[0].str.split(' ', length - 1, expand=True)
#Split based on separator - expensive
num2str = lambda x: float(x)
#convert all elements from str to float
df2 = df2.applymap(num2str)
#apply num2str to every element - expensive
df2.columns = format_spec
#add corresponding column labels
df2 = df2.sort_values(by=['id'])
#sort based on atom id so that future operations
# are easy.
key = 'timestep_' + str(timestep)
#save in the corresponding dictionary.
coord[key] = df2
index = index + natoms
timestep += 1
else:
index += 1
except IndexError:
index += 1
elif last_only:
index = len(unwrap) - natoms - 15 #15 is an arbitrary chosen value
while index < len(unwrap):
line = unwrap.iloc[index].str.split()
try:
if line[0][0] == 'ITEM:' and line[0][1] == 'ATOMS':
length = len(unwrap.iloc[index + 1].str.split()[0])
df2 = unwrap.iloc[index + 1:index + natoms + 1]
df2 = df2[0].str.split(' ', length - 1, expand=True)
#Split based on separator - expensive
num2str = lambda x: float(x)
#convert all elements from str to float
df2 = df2.applymap(num2str)
#apply num2str to every element - expensive
df2.columns = format_spec
#add corresponding column labels
df2 = df2.sort_values(by=['id'])
#sort based on atom id so that future operations
# are easy.
key = 'timestep_' + str(timestep)
#save in the corresponding dictionary.
coord[key] = df2
index = index + natoms
timestep += 1
else:
index += 1
except IndexError:
index += 1
else:
while index < len(unwrap):
line = unwrap.iloc[index].str.split()
try:
if line[0][0] == 'ITEM:' and line[0][1] == 'ATOMS':
length = len(unwrap.iloc[index + 1].str.split()[0])
df2 = unwrap.iloc[index + 1:index + natoms + 1]
df2 = df2[0].str.split(' ', length - 1, expand=True)
#Split based on separator - expensive
num2str = lambda x: float(x)
#convert all elements from str to float
df2 = df2.applymap(num2str)
#apply num2str to every element - expensive
df2.columns = format_spec
#add corresponding column labels
df2 = df2.sort_values(by=['id'])
#sort based on atom id so that future operations
# are easy.
key = 'timestep_' + str(timestep)
#save in the corresponding dictionary.
coord[key] = df2
index = index + natoms
timestep += 1
else:
index += 1
except IndexError:
index += 1
if boxsize:
index = 0
boxsizev = [None, None, None]
if first_only:
while index < first_timestep_lines:
line = unwrap.iloc[index].str.split()
try:
if line[0][0] == 'ITEM:' and line[0][1] == 'BOX':
df2 = unwrap.iloc[index + 1:index + 2]
df2 = df2[0].str.split(' ').values
boxsizev[0] = float(df2[0][1]) - float(df2[0][0])
dfy = unwrap.iloc[index + 2:index + 3]
dfy = dfy[0].str.split(' ').values
boxsizev[1] = float(dfy[0][1]) - float(dfy[0][0])
dfz = unwrap.iloc[index + 3:index + 4]
dfz = dfz[0].str.split(' ').values
boxsizev[2] = float(dfz[0][1]) - float(dfz[0][0])
if start_end:
return coord, boxsizev, float(df2[0][1]), float(
df2[0][0]), float(dfy[0][1]), float(
dfy[0][0]), float(dfz[0][1]), float(
dfz[0][0])
else:
return coord, boxsizev
else:
index += 1
except IndexError:
index += 1
if last_only:
index = len(unwrap) - natoms - 20 #15 is an arbitrary chosen value
while index < len(unwrap):
line = unwrap.iloc[index].str.split()
try:
if line[0][0] == 'ITEM:' and line[0][1] == 'BOX':
df2 = unwrap.iloc[index + 1:index + 2]
df2 = df2[0].str.split(' ').values
boxsizev[0] = float(df2[0][1]) - float(df2[0][0])
dfy = unwrap.iloc[index + 2:index + 3]
dfy = dfy[0].str.split(' ').values
boxsizev[1] = float(dfy[0][1]) - float(dfy[0][0])
dfz = unwrap.iloc[index + 3:index + 4]
dfz = dfz[0].str.split(' ').values
boxsizev[2] = float(dfz[0][1]) - float(dfz[0][0])
return coord, boxsizev
else:
index += 1
except IndexError:
index += 1
if boxsize_whole:
index = 0
timestep = 0
boxsize_whole_dict = {}
boxsizev = [None, None, None]
while index < len(unwrap):
line = unwrap.iloc[index].str.split()
try:
if line[0][0] == 'ITEM:' and line[0][1] == 'BOX':
df2 = unwrap.iloc[index + 1:index + 2]
df2 = df2[0].str.split(' ').values
boxsizev[0] = float(df2[0][1]) - float(df2[0][0])
dfy = unwrap.iloc[index + 2:index + 3]
dfy = dfy[0].str.split(' ').values
boxsizev[1] = float(dfy[0][1]) - float(dfy[0][0])
dfz = unwrap.iloc[index + 3:index + 4]
dfz = dfz[0].str.split(' ').values
boxsizev[2] = float(dfz[0][1]) - float(dfz[0][0])
key = 'timestep_' + str(timestep)
boxsize_whole_dict[key] = [
boxsizev[0], boxsizev[1], boxsizev[2]
]
index = index + natoms
timestep += 1
else:
index += 1
except IndexError:
index += 1
return coord, boxsize_whole_dict
return coord