def importNomalized(pathName, exp):
labelsList = list()
ch_names = []
create_ch_name = False
# carregando pasta "large_train"
path = gop('ls {}'.format(pathName)).split('\n')
# 1ª dimensão dos dados contendo os sujeitos. Ex.: C_1, a_m, etc
subjects = list()
for types in path:
if("co2c" in types):
for i in range (0,10):
labelsList.append([[0]]*61)
else:
for i in range (0,10):
labelsList.append([[1]]*61)
files = gop('ls {}/{}'.format(pathName, types)).split('\n')
# 2ª dimensão dos dados contendo as sessões (trials)
trials = list()
for f in files:
arquivo = open('{}/{}/{}'.format(pathName, types, f))
text = arquivo.readlines()
# 3ª dimensão dos dados contendo os canais (eletrodos)
chs = list()
# 4ª dimensão dos dados contendo os valores em milivolts
values = list()
for line in text:
# ex: "# FP1 chan 0"
t = search('(?P<ch_name>\w{1,3}) chan \d{1,2}', line)
# ex: "0 FP1 0 -8.921"
p = search('^\d{1,2}\ \w{1,3}\ \d{1,3}\ (?P<value>.+$)', line)
if p:
values.append(float(p.group('value')))
# mudou para outro eletrodo
elif t:
if values:
chs.append(values)
values = list()
if not create_ch_name:
ch_names.append(t.group('ch_name').lower())
create_ch_name = True
chs.append(values)
arquivo.seek(32*3)
line = arquivo.readline()
arquivo.close()
if exp in line:
trials.append(chs)
subjects.append(trials)
data = np.array(subjects)
return [filterData(data),np.asarray(labelsList)]
def calculate_boxsize(pdbfile, tmp_prefix=".tmp"):
tmp_prefix=f"{tmpdir}/{tmp_prefix}"
with open(f"{tmp_prefix}.in", "w") as fout:
fout.write(tmp_leap.format(pdbfile=pdbfile, tmp_prefix=tmp_prefix))
print(gop(f"tleap -f {tmp_prefix}.in | tee {tmp_prefix}.in.result"))
center_str = gop(f"cat {tmp_prefix}.in.result | grep The\ center | cut -d: -f2 | sed s/,//g")
center = [float(s) for s in center_str.split()]
box_size_str = gop(f"tail -n 1 {tmp_prefix}.rst7 | cut -c -36")
box_size = [float(s) for s in box_size_str.split()]
box_size = max(box_size)
return box_size
def load_bases2(path_files):
diretory = gop('ls {}'.format(folders[path_files])).split('\n')
subA = {"S1_obj": [], "S2_nomatch": [], "S2_match": []}
subC = {"S1_obj": [], "S2_nomatch": [], "S2_match": []}
for types in diretory:
files = gop('ls {}/{}'.format(folders[path_files], types)).split('\n')
if 'Co2A' in types.title():
get_all_datas(files, subA, path_files, types)
else:
get_all_datas(files, subC, path_files, types)
return [subA, subC]
def gen_packmol_input(protein_pdb, cosolv_pdbs, box_pdb, inp, box_size, molar):
# shorten path length to pdb file
# too long path cannot be treated by packmol
temp_protein_pdb = f"{tmpdir}/.temp_protein.pdb"
print(gop(f"cp {protein_pdb} {temp_protein_pdb}"))
temp_pdbs = [f"{tmpdir}/.temp_{i}.pdb" for i in range(len(cosolv_pdbs))]
[gop(f"cp {src} {dst}") for src, dst in zip(cosolv_pdbs, temp_pdbs)]
num = int(constants.N_A * molar * (box_size**3) * (10**-27))
with open(inp, "w") as fout:
fout.write(TEMPLATE_PACKMOL_HEADER.format(output=box_pdb, prot=temp_protein_pdb))
fout.write("\n")
for pdb in temp_pdbs:
fout.write(TEMPLATE_PACKMOL_STRUCT.format(cosolv=pdb, num=num, size=box_size/2))
fout.write("\n")
def run_tleap(tleap_path, inp):
output = gop("%s -f %s" % (tleap_path, inp))
print(output)
final_charge_info = [s.strip() for s in output.split("\n")
if s.strip().startswith("Total unperturbed charge")][0]
final_charge_value = float(final_charge_info.split()[-1])
return final_charge_value
def do_cpptraj(top,
resi_st,
resi_ed,
traj,
ref,
ref_resi_st,
ref_resi_ed,
cid,
prefix,
n=100,
d=1,
traj_start=1,
traj_stop="last",
traj_offset=1,
cpptraj_exe="cpptraj"):
log = gop(f"""{cpptraj_exe} -p {top} << EOF
trajin {traj} {traj_start} {traj_stop} {traj_offset}
parm {ref}
reference {ref} parm {ref} [REF]
unwrap :{resi_st}-{resi_ed}@CA
center :{resi_st}-{resi_ed}@CA
fiximagedbonds
autoimage
rms ToREF ref [REF] :{resi_st}-{resi_ed}@CA :{ref_resi_st}-{ref_resi_ed}@CA out rmsd.dat
# "center" maybe changed AFTER grid generation. be careful...
# file name must not start with a number (i.e. 1L2S_A11_nV.dx)
# grid map_{prefix}_nV.dx {n} {d} {n} {d} {n} {d} center :{resi_st}-{resi_ed}@CA :{cid}&(!@VIS)
# grid map_{prefix}_nVH.dx {n} {d} {n} {d} {n} {d} center :{resi_st}-{resi_ed}@CA :{cid}&(!@VIS)&(!@H*)
# grid map_{prefix}_V.dx {n} {d} {n} {d} {n} {d} center :{resi_st}-{resi_ed}@CA :{cid}@VIS
# grid map_{prefix}_O.dx {n} {d} {n} {d} {n} {d} center :{resi_st}-{resi_ed}@CA :{cid}@O*
grid map_{prefix}_nV.dx {n} {d} {n} {d} {n} {d} :{cid}&(!@VIS)
grid map_{prefix}_nVH.dx {n} {d} {n} {d} {n} {d} :{cid}&(!@VIS)&(!@H*)
grid map_{prefix}_V.dx {n} {d} {n} {d} {n} {d} :{cid}@VIS
grid map_{prefix}_O.dx {n} {d} {n} {d} {n} {d} :{cid}@O*
trajout {prefix}_position_check.pdb start 1 stop 1 offset 1
trajout {prefix}_position_check2.pdb offset 100
go
exit
EOF""")
print(log)
def main():
# arquivos
arquivos = gop('ls data_10hz').split('\n')
# Tempos selecionados a mão
times = [[95, 96, 234, 367, 597, 786],
[39, 46, 110, 130, 280, 363],
[103, 114, 136, 154, 171, 185]]
window = 3
# Para cada arquivo
for arqPos, arq in enumerate(arquivos):
data, fileLimitTime = loadData(arq)
print(data.shape)
raw = prepareRaw(data)
# acha os tempos do ritmo alpha
for i in range(fileLimitTime - window):
findAlphaTime(raw, i, window)
# plota os tempos selecionados
for i in times[arqPos]:
plotWindowTime(raw, i, window)
def load_data_with_label(self):
create_ch_name = False
subjects = []
labels = []
# get all dirs inside folder
folder = gop('ls {}'.format(self.folder_path)).split('\n')
# for all dir insider folder
for types in folder:
if (not os.path.isdir('{}/{}'.format(self.folder_path, types))):
continue
label_search = search('^co2(?P<label>\w{1})', types)
files = gop('ls {}/{}'.format(self.folder_path, types)).split('\n')
# 2ª dimensão dos dados contendo as sessões (trials)
trials = list()
# for all file in files
for f in files:
if (label_search.group("label") == "a"):
labels.append(1)
else:
labels.append(0)
arquivo = open('{}/{}/{}'.format(self.folder_path, types, f))
text = arquivo.readlines()
# 3ª dimensão dos dados contendo os canais (eletrodos)
chs = list()
# 4ª dimensão dos dados contendo os valores em milivolts
values = list()
# for each line inside a file
for line in text:
# ex: "# FP1 chan 0"
# look if this line is a new eletrodo info
t = search('(?P<ch_name>\w{1,3}) chan \d{1,2}', line)
# ex: "0 FP1 0 -8.921"
# or if is a data line
p = search('^\d{1,3}\ \w{1,3}\ \d{1,3}\ (?P<value>.+$)',
line)
# if has a eeg data
if p:
values.append(float(p.group('value')))
# mudou para outro eletrodo
elif t:
if values:
chs.append(values)
values = list()
if not create_ch_name:
self.ch_names.append(t.group('ch_name').lower())
# end for line
# append last channel
chs.append(values)
create_ch_name = True
# append all channels to one trial
trials.append(chs)
arquivo.close()
# append all trials to one subject
subjects.append(trials)
self.subjects = np.array(subjects)
self.labels = np.array(labels)
def run_packmol(packmol_path, inp):
print(gop("%s < %s" % (packmol_path, inp)))
def run_parmchk(mol2, frcmod, at):
at_id = {"gaff": 1, "gaff2": 2}[at]
print(gop(f"parmchk2 -i {mol2} -f mol2 -o {frcmod} -s {at_id}"))
cosolv_box_size = ((constants.N_A * float(params["Cosolvent"]["molar"]))
* 1e-27)**(-1/3.0) # /L -> /A^3 : 1e-27
# 1. generate cosolvent box with packmol
packmol_input = f"{tmpdir}/.temp_packmol.input"
packmol_box_pdb = f"{tmpdir}/.temp_box.pdb"
cfrcmods = [f"{tmpdir}/.temp_cosolvent_{cid}.frcmod" for cid in cids]
gen_packmol_input(params["Protein"]["pdb"], cpdbs,
packmol_box_pdb,
packmol_input,
boxsize,
float(params["Cosolvent"]["molar"]))
while True:
run_packmol(args.packmol, packmol_input) # -> output: packmol_box_pdb
temp_box = f"{tmpdir}/.temp_packmol_box_2.pdb"
gop("grep -v OXT {} > {}".format(packmol_box_pdb, temp_box))
for mol2, frcmod in zip(cmols, cfrcmods):
run_parmchk(mol2, frcmod, params["Cosolvent"]["atomtype"])
# 2. amber tleap
tleap_input = f"{tmpdir}/.temp_tleap.in"
gen_tleap_input(expandpath(args.tleap_input), tleap_input,
cids, cmols,
cfrcmods,
temp_box,
boxsize,
args.output_prefix,
ssbonds,
params["Cosolvent"]["atomtype"])
print('Downlod dos datasets concluído!')
else:
print('Dataset já baixado!')
from subprocess import getoutput as gop
import glob
# único arquivo somente empacotado (tar)
# os.mkdir('dataset/eeg_full/')
# gop('tar -xvf dataset/eeg_full.tar -C dataset/eeg_full')
# os.remove('dataset/eeg_full.tar')
while glob.glob('dataset/**/*.gz', recursive=True):
# quando o arquivo está empacotado (tar) e compactado (gz)
for f in glob.iglob('dataset/**/*.tar.gz', recursive=True):
gop('tar -zxvf {} -C {}'.format(f, f[:f.rindex('/')]))
os.remove(f)
# quando o arquivo está somente compactado (gz)
for f in glob.iglob('dataset/**/*.gz', recursive=True):
gop('gzip -d {}'.format(f))
print('Descompactações finalizadas!')
# organizando melhor as pastas
os.rename('dataset/smni_eeg_data', 'dataset/small')
# os.rename('dataset/eeg_full', 'dataset/full')
os.rename('dataset/SMNI_CMI_TRAIN/', 'dataset/large_train/')
os.rename('dataset/SMNI_CMI_TEST/', 'dataset/large_test/')
print(gop('ls -l dataset/'))
from re import search
import numpy as np