def mol2drs(self):
for raw_name, fname, input_fpath, drs_input_fpath, output_fpath in self.iter_files():
XMLWriter(input_fpath, drs_input_fpath, output_fpath)
# dragon6shell -s *.drs to get the result
cmd = "%s '%s'" % (CALCULATE_CMD_TYPE.DRAGON, drs_input_fpath)
chemistry_logger.info('mol2drs cmd %s' % cmd)
CalcoreCmd(cmd, output=output_fpath).run()
def logPL(self):
from .matrix.pl import plX
#PL 模型，有温度参数
#CAS-Number 1/T nHDon X1sol GATS1v μ nROH
abstract_value = self.dragon_model.extractparameter(
["nHDon", "X1sol", "GATS1v", "μ", "nROH"])
for smilenum in abstract_value.keys():
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
value = 13.33 - \
2571.0 * (1 / self.T) - \
0.5061 * abstract_value[smilenum]['nHDon'] - \
0.6896 * abstract_value[smilenum]['X1sol'] + \
0.8014 * abstract_value[smilenum]['GATS1v'] - \
0.1363 * abstract_value[smilenum]['μ'] - \
0.6094 * abstract_value[smilenum]['nROH']
self.predict_result[smilenum]['logPL']['value'] = self.round(value)
chemistry_logger.info('pl(%s) dragon: %s' %
(smilenum, abstract_value[smilenum]))
x = matrix([[(1 / self.T), abstract_value[smilenum]['nHDon'],
abstract_value[smilenum]['X1sol'],
abstract_value[smilenum]['GATS1v'],
abstract_value[smilenum]['μ'],
abstract_value[smilenum]['nROH']]])
williams = self.get_williams(plX, x)
self.predict_result[smilenum]['logPL'].update(williams)
def logPL(self):
from .matrix.pl import plX
#PL 模型，有温度参数
#CAS-Number 1/T nHDon X1sol GATS1v μ nROH
abstract_value = self.dragon_model.extractparameter([
"nHDon", "X1sol", "GATS1v", "μ", "nROH"])
for smilenum in abstract_value.keys():
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
value = 13.33 - \
2571.0 * (1 / self.T) - \
0.5061 * abstract_value[smilenum]['nHDon'] - \
0.6896 * abstract_value[smilenum]['X1sol'] + \
0.8014 * abstract_value[smilenum]['GATS1v'] - \
0.1363 * abstract_value[smilenum]['μ'] - \
0.6094 * abstract_value[smilenum]['nROH']
self.predict_result[smilenum]['logPL']['value'] = self.round(value)
chemistry_logger.info('pl(%s) dragon: %s' % (smilenum, abstract_value[smilenum]))
x = matrix([[(1 / self.T),
abstract_value[smilenum]['nHDon'],
abstract_value[smilenum]['X1sol'],
abstract_value[smilenum]['GATS1v'],
abstract_value[smilenum]['μ'],
abstract_value[smilenum]['nROH']]])
williams = self.get_williams(plX, x)
self.predict_result[smilenum]['logPL'].update(williams)
def run(self):
#FIXME: 运行时候加锁
if not self.check_output_exists():
chemistry_logger.info('[CalcoreCmd]%s(%s->%s)', self.cmd,
self.input, self.output)
s = check_call(self.cmd, shell=True)
chemistry_logger.info(s)
def mol2drs(self):
for raw_name, fname, input_fpath, drs_input_fpath, output_fpath in self.iter_files(
):
XMLWriter(input_fpath, drs_input_fpath, output_fpath)
# dragon6shell -s *.drs to get the result
cmd = "%s '%s'" % (CALCULATE_CMD_TYPE.DRAGON, drs_input_fpath)
chemistry_logger.info('mol2drs cmd %s' % cmd)
CalcoreCmd(cmd, output=output_fpath).run()
def logKOH_T(self):
from .matrix.koh import kohX
# KOH 温度依附性模型, 有温度参数
#CAS-Number X% EHOMO Mor29u NdsCH GATS1e X3A 1/T SdsCH nR=Cp F02[F-Br] RDF015m BIC1 SpMin8_Bh(p) NssssC
abstract_value = self.dragon_model.extractparameter([
"X%", "EHOMO", "Mor29u", "NdsCH", "GATS1e", "X3A", "SdsCH",
"nR=Cp", "F02[F-Br]", "RDF015m", "BIC1", "SpMin8_Bh(p)", "NssssC"
])
for smilenum in abstract_value:
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
abstract_value[smilenum]['EHOMO'] = fetch_ehomo(
smilenum, 'logKOH_T')
value = -8.613 - \
0.02100 * abstract_value[smilenum]['X%'] + \
14.38 * abstract_value[smilenum]['EHOMO'] - \
0.6430 * abstract_value[smilenum]['Mor29u'] + \
0.5870 * abstract_value[smilenum]['NdsCH'] + \
0.5870 * abstract_value[smilenum]['GATS1e'] + \
0.5770 * abstract_value[smilenum]['X3A'] - \
0.2450 * abstract_value[smilenum]['SdsCH'] - \
167.0 * (1 / self.T) + \
1.103 * abstract_value[smilenum]['BIC1'] + \
0.1170 * abstract_value[smilenum]['RDF015m'] - \
1.044 * abstract_value[smilenum]['SpMin8_Bh(p)'] + \
0.2390 * abstract_value[smilenum]['nR=Cp'] - \
0.1980 * abstract_value[smilenum]['NssssC'] - \
0.5080 * abstract_value[smilenum]['F02[F-Br]']
self.predict_result[smilenum]['logKOH_T']['value'] = self.round(
value)
chemistry_logger.info('koh_t(%s) dragon: %s' %
(smilenum, abstract_value[smilenum]))
x = matrix([[
abstract_value[smilenum]['X%'],
abstract_value[smilenum]['EHOMO'],
abstract_value[smilenum]['Mor29u'],
abstract_value[smilenum]['NdsCH'],
abstract_value[smilenum]['GATS1e'],
abstract_value[smilenum]['X3A'], 1.0 / self.T,
abstract_value[smilenum]['SdsCH'],
abstract_value[smilenum]['nR=Cp'],
abstract_value[smilenum]['F02[F-Br]'],
abstract_value[smilenum]['RDF015m'],
abstract_value[smilenum]['BIC1'],
abstract_value[smilenum]['SpMin8_Bh(p)'],
abstract_value[smilenum]['NssssC']
]])
williams = self.get_williams(kohX, x)
self.predict_result[smilenum]['logKOH_T'].update(williams)
def logBDG(self):
from .matrix.bdg import bdgX
# BDG 无温度参数
#CAS nN nHM O% MATS1e GATS1p GATS7p GGI1 GGI2 nCq nCrt C-040 H-048 H-051 O-059
abstract_value = self.dragon_model.extractparameter([
"nN", "nHM", "O%", "MATS1e", "GATS1p", "GATS7p", "GGI1", "GGI2",
"nCq", "nCrt", "C-040", "H-048", "H-051", "O-059"
])
for smilenum in abstract_value.keys():
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
x = 1.9025 + \
1.0457 * abstract_value[smilenum]['nN'] + \
0.6662 * abstract_value[smilenum]['nHM'] - \
0.1078 * abstract_value[smilenum]['O%'] + \
2.8362 * abstract_value[smilenum]['MATS1e'] - \
2.0019 * abstract_value[smilenum]['GATS1p'] - \
0.7015 * abstract_value[smilenum]['GATS7p'] + \
0.1131 * abstract_value[smilenum]['GGI1'] + \
0.7023 * abstract_value[smilenum]['GGI2'] + \
2.7793 * abstract_value[smilenum]['nCq'] + \
1.035 * abstract_value[smilenum]['nCrt'] - \
0.777 * abstract_value[smilenum]['C-040'] - \
0.7091 * abstract_value[smilenum]['H-048'] - \
0.1553 * abstract_value[smilenum]['H-051'] + \
0.955 * abstract_value[smilenum]['O-059']
self.predict_result[smilenum]['logBDG']['value'] = self.round(x)
self.predict_result[smilenum]['logBDG']['degrade'] = self.round(
1 / (1 + math.exp(-x)))
chemistry_logger.info('bdg(%s) dragon: %s' %
(smilenum, abstract_value[smilenum]))
x = matrix([[
abstract_value[smilenum]['nN'],
abstract_value[smilenum]['nHM'],
abstract_value[smilenum]['O%'],
abstract_value[smilenum]['MATS1e'],
abstract_value[smilenum]['GATS1p'],
abstract_value[smilenum]['GATS7p'],
abstract_value[smilenum]['GGI1'],
abstract_value[smilenum]['GGI2'],
abstract_value[smilenum]['nCq'],
abstract_value[smilenum]['nCrt'],
abstract_value[smilenum]['C-040'],
abstract_value[smilenum]['H-048'],
abstract_value[smilenum]['H-051'],
abstract_value[smilenum]['O-059']
]])
williams = self.get_williams(bdgX, x)
self.predict_result[smilenum]['logBDG'].update(williams)
def logKOH(self):
from .matrix.koh_298k import koh_298kX
# KOH 298K预测模型, 无温度参数
#CAS-Number EHOMO AMW NdsCH Mor14i nP nR=Cp X% nRCHO C-020 SpMaxA_AEA(dm) nCbH CATS2D_03_DL
abstract_value = self.dragon_model.extractparameter([
"EHOMO", "AMW", "NdsCH", "Mor14i", "nP", "nR=Cp", "X%", "nRCHO",
"C-020", "SpMaxA_AEA(dm)", "nCbH", "CATS2D_03_DL"
])
for smilenum in abstract_value:
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
abstract_value[smilenum]['EHOMO'] = fetch_ehomo(smilenum, 'logKOH')
value = -6.511 + \
15.85 * abstract_value[smilenum]['EHOMO'] - \
0.03800 * abstract_value[smilenum]['AMW'] + \
0.1300 * abstract_value[smilenum]['NdsCH'] + \
0.1630 * abstract_value[smilenum]['Mor14i'] + \
0.7790 * abstract_value[smilenum]['nP'] + \
0.3170 * abstract_value[smilenum]['nR=Cp'] - \
0.01900 * abstract_value[smilenum]['X%'] + \
0.3930 * abstract_value[smilenum]['nRCHO'] + \
0.5890 * abstract_value[smilenum]['C-020'] - \
0.4550 * abstract_value[smilenum]['SpMaxA_AEA(dm)'] - \
0.05600 * abstract_value[smilenum]['nCbH'] + \
0.1410 * abstract_value[smilenum]['CATS2D_03_DL']
self.predict_result[smilenum]['logKOH']['value'] = self.round(
value)
chemistry_logger.info('koh(%s) dragon: %s' %
(smilenum, abstract_value[smilenum]))
x = matrix([[
abstract_value[smilenum]['EHOMO'],
abstract_value[smilenum]['AMW'],
abstract_value[smilenum]['NdsCH'],
abstract_value[smilenum]['Mor14i'],
abstract_value[smilenum]['nP'],
abstract_value[smilenum]['nR=Cp'],
abstract_value[smilenum]['X%'],
abstract_value[smilenum]['nRCHO'],
abstract_value[smilenum]['C-020'],
abstract_value[smilenum]['SpMaxA_AEA(dm)'],
abstract_value[smilenum]['nCbH'],
abstract_value[smilenum]['CATS2D_03_DL']
]])
williams = self.get_williams(koh_298kX, x)
self.predict_result[smilenum]['logKOH'].update(williams)
def logKOH_T(self):
from .matrix.koh import kohX
# KOH 温度依附性模型, 有温度参数
#CAS-Number X% EHOMO Mor29u NdsCH GATS1e X3A 1/T SdsCH nR=Cp F02[F-Br] RDF015m BIC1 SpMin8_Bh(p) NssssC
abstract_value = self.dragon_model.extractparameter([
"X%", "EHOMO", "Mor29u", "NdsCH", "GATS1e", "X3A", "SdsCH",
"nR=Cp", "F02[F-Br]", "RDF015m", "BIC1", "SpMin8_Bh(p)", "NssssC"])
for smilenum in abstract_value:
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
abstract_value[smilenum]['EHOMO'] = fetch_ehomo(smilenum, 'logKOH_T')
value = -8.613 - \
0.02100 * abstract_value[smilenum]['X%'] + \
14.38 * abstract_value[smilenum]['EHOMO'] - \
0.6430 * abstract_value[smilenum]['Mor29u'] + \
0.5870 * abstract_value[smilenum]['NdsCH'] + \
0.5870 * abstract_value[smilenum]['GATS1e'] + \
0.5770 * abstract_value[smilenum]['X3A'] - \
0.2450 * abstract_value[smilenum]['SdsCH'] - \
167.0 * (1 / self.T) + \
1.103 * abstract_value[smilenum]['BIC1'] + \
0.1170 * abstract_value[smilenum]['RDF015m'] - \
1.044 * abstract_value[smilenum]['SpMin8_Bh(p)'] + \
0.2390 * abstract_value[smilenum]['nR=Cp'] - \
0.1980 * abstract_value[smilenum]['NssssC'] - \
0.5080 * abstract_value[smilenum]['F02[F-Br]']
self.predict_result[smilenum]['logKOH_T']['value'] = self.round(value)
chemistry_logger.info('koh_t(%s) dragon: %s' % (smilenum, abstract_value[smilenum]))
x = matrix([[abstract_value[smilenum]['X%'],
abstract_value[smilenum]['EHOMO'],
abstract_value[smilenum]['Mor29u'],
abstract_value[smilenum]['NdsCH'],
abstract_value[smilenum]['GATS1e'],
abstract_value[smilenum]['X3A'],
1.0 / self.T,
abstract_value[smilenum]['SdsCH'],
abstract_value[smilenum]['nR=Cp'],
abstract_value[smilenum]['F02[F-Br]'],
abstract_value[smilenum]['RDF015m'],
abstract_value[smilenum]['BIC1'],
abstract_value[smilenum]['SpMin8_Bh(p)'],
abstract_value[smilenum]['NssssC']
]])
williams = self.get_williams(kohX, x)
self.predict_result[smilenum]['logKOH_T'].update(williams)
def mol2dragon_folder(self):
mop_fname_set = set()
# STEP: 将smile码经过obabel转化，生成mop文件，并放在指定目录中
for element in self.iter_smiles_files(self.__smilenum_list, 'smile'):
smile, name, dragon_dpath, mopac_dpath, mop_fpath = element
# '-:smi' 可以直接对smile进行转化
cmd = 'obabel -:"%s" -o mop -O "%s" --gen3D' % (smile, mop_fpath)
chemistry_logger.info('mop2mopac, smi->mop: %s' % cmd)
CalcoreCmd(cmd, output=mop_fpath).run()
# 修改smi->mop文件的头部
lines = []
with open(mop_fpath, 'rb') as f:
lines = f.readlines()
if self.model_name in ('logKOA', ):
lines[0] = 'EF GNORM=0.0001 MMOK GEO-OK PM3\n'
elif self.model_name in ('logRP', ):
lines[
0] = 'eps=78.6 EF GNORM =0.0100 MMOK GEO-OK PM6 MULLIK GRAPH ESR HYPERFINE POLAR PRECISE BOND PI ENPART DEBUG\n'
elif self.model_name in ('logPL', ):
lines[0] = 'EF GNORM=0.01 MMOK GEO-OK PM6 MULLIK POLAR\n'
elif self.model_name in ('logO3', ):
lines[
0] = 'EF GNORM=0.001 PM6 MULLIK GRAPH ESR HYPERFINE POLAR\n'
elif self.model_name in ('logBDG', ):
lines[0] = 'EF GNORM=0.1 MMOK GEO-OK PM5\n'
else:
lines[0] = 'no keywords'
lines[1] = mopac_dpath + "\n"
with open(mop_fpath, 'wb') as f:
f.writelines(lines)
mop_fname_set.add('%s.mop' % name)
for element in self.iter_smiles_files(self.__molfile, 'file'):
mol_fpath, name, dragon_dpath, mopac_dpath, _ = element
mop_fpath = mol2mop(mol_fpath)
if not os.path.exists(mopac_dpath):
os.makedirs(mopac_dpath)
shutil.copy(mop_fpath, mopac_dpath)
mop_fname_set.add('%s.mop' % name)
# 使用mopac对dragon结果进行优化(输入转化生成的mop文件)
mop = MopacModel(mop_fname_set)
mop.opt4dragon(self.model_name)
def logO3(self):
from .matrix.o3 import o3X
#O3 有温度参数
#EHOMO, SaasC, SpPosA_B(m), nR=Cs, B01[C-C],
#X2Av, nCconj, nDB, N-076
abstract_value = self.dragon_model.extractparameter([
"EHOMO", "SaasC", "SpPosA_B(m)", "nR=Cs", "B01[C-C]", "X2Av",
"nCconj", "nDB", "N-076"
])
chemistry_logger.info(abstract_value)
for smilenum in abstract_value:
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
abstract_value[smilenum]['EHOMO'] = fetch_ehomo_by_mopac(
smilenum, 'logO3')
value = 2.549 + \
1.534 * abstract_value[smilenum]['EHOMO'] - \
1.284 * abstract_value[smilenum]['SaasC'] - \
2.092 * abstract_value[smilenum]['SpPosA_B(m)'] + \
0.3060 * abstract_value[smilenum]['nR=Cs'] + \
1.113 * abstract_value[smilenum]['B01[C-C]'] - \
2.289 * abstract_value[smilenum]['X2Av'] - \
0.3470 * abstract_value[smilenum]['nCconj'] + \
0.5960 * abstract_value[smilenum]['nDB'] - \
1.301 * abstract_value[smilenum]['N-076'] - \
878.7 / self.T
self.predict_result[smilenum]['logO3']['value'] = self.round(value)
chemistry_logger.info('O3(%s) dragon: %s' %
(smilenum, abstract_value[smilenum]))
x = matrix([[
1.0 / self.T,
abstract_value[smilenum]['EHOMO'],
abstract_value[smilenum]['nDB'],
abstract_value[smilenum]['SaasC'],
abstract_value[smilenum]['SpPosA_B(m)'],
abstract_value[smilenum]['nCconj'],
abstract_value[smilenum]['nR=Cs'],
abstract_value[smilenum]['B01[C-C]'],
abstract_value[smilenum]['X2Av'],
abstract_value[smilenum]['N-076'],
]])
williams = self.get_williams(o3X, x)
self.predict_result[smilenum]['logO3'].update(williams)
def logKOA(self):
from .matrix.koa import koaX
#KOA 有温度参数
#X1sol Mor13v H050 R5v TO..Cl HATS5v RDF035m RCl nRCOOR Mor15u RDF90m 1/T
abstract_value = self.dragon_model.extractparameter([
"X1sol", "Mor13v", "H-050", "R5v", "T(O..Cl)", "HATS5v", "RDF035m",
"RCI", "nRCOOR", "Mor15u", "RDF090m"
])
for smilenum in abstract_value:
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
value = -3.03 + \
313.0 * abstract_value[smilenum]['X1sol'] / self.T - \
85.7 * abstract_value[smilenum]['Mor13v'] / self.T + \
432.0 * abstract_value[smilenum]['H-050'] / self.T - \
1270.0 * abstract_value[smilenum]['R5v'] / self.T - \
5.54 * abstract_value[smilenum]['T(O..Cl)'] / self.T + \
125.0 * abstract_value[smilenum]['HATS5v'] / self.T - \
13.3 * abstract_value[smilenum]['RDF035m'] / self.T - \
61.1 * abstract_value[smilenum]['RCI'] / self.T - \
37.6 * abstract_value[smilenum]['nRCOOR'] / self.T + \
156.0 * abstract_value[smilenum]['Mor15u'] / self.T -\
5.49 * abstract_value[smilenum]['RDF090m'] / self.T + \
1040.0 / self.T
self.predict_result[smilenum]['logKOA']['value'] = self.round(
value)
chemistry_logger.info('koa(%s) dragon: %s' %
(smilenum, abstract_value[smilenum]))
x = matrix([[
abstract_value[smilenum]['X1sol'] / self.T,
abstract_value[smilenum]['Mor13v'] / self.T,
abstract_value[smilenum]['H-050'] / self.T,
abstract_value[smilenum]['R5v'] / self.T,
abstract_value[smilenum]['T(O..Cl)'] / self.T,
abstract_value[smilenum]['HATS5v'] / self.T,
abstract_value[smilenum]['RDF035m'] / self.T,
abstract_value[smilenum]['RCI'] / self.T,
abstract_value[smilenum]['nRCOOR'] / self.T,
abstract_value[smilenum]['Mor15u'] / self.T,
abstract_value[smilenum]['RDF090m'] / self.T,
1.0 / self.T,
]])
williams = self.get_williams(koaX, x)
self.predict_result[smilenum]['logKOA'].update(williams)
def mol2dragon_folder(self):
mop_fname_set = set()
# STEP: 将smile码经过obabel转化，生成mop文件，并放在指定目录中
for element in self.iter_smiles_files(self.__smilenum_list, 'smile'):
smile, name, dragon_dpath, mopac_dpath, mop_fpath = element
# '-:smi' 可以直接对smile进行转化
cmd = 'obabel -:"%s" -o mop -O "%s" --gen3D' % (smile,
mop_fpath)
chemistry_logger.info('mop2mopac, smi->mop: %s' % cmd)
CalcoreCmd(cmd, output=mop_fpath).run()
# 修改smi->mop文件的头部
lines = []
with open(mop_fpath, 'rb') as f:
lines = f.readlines()
if self.model_name in ('logKOA',):
lines[0] = 'EF GNORM=0.0001 MMOK GEO-OK PM3\n'
elif self.model_name in ('logRP',):
lines[0] = 'eps=78.6 EF GNORM =0.0100 MMOK GEO-OK PM6 MULLIK GRAPH ESR HYPERFINE POLAR PRECISE BOND PI ENPART DEBUG\n'
elif self.model_name in ('logPL',):
lines[0] = 'EF GNORM=0.01 MMOK GEO-OK PM6 MULLIK POLAR\n'
elif self.model_name in ('logO3', ):
lines[0] = 'EF GNORM=0.001 PM6 MULLIK GRAPH ESR HYPERFINE POLAR\n'
elif self.model_name in ('logBDG',):
lines[0] = 'EF GNORM=0.1 MMOK GEO-OK PM5\n'
else:
lines[0] = 'no keywords'
lines[1] = mopac_dpath + "\n"
with open(mop_fpath, 'wb') as f:
f.writelines(lines)
mop_fname_set.add('%s.mop' % name)
for element in self.iter_smiles_files(self.__molfile, 'file'):
mol_fpath, name, dragon_dpath, mopac_dpath, _ = element
mop_fpath = mol2mop(mol_fpath)
if not os.path.exists(mopac_dpath):
os.makedirs(mopac_dpath)
shutil.copy(mop_fpath, mopac_dpath)
mop_fname_set.add('%s.mop' % name)
# 使用mopac对dragon结果进行优化(输入转化生成的mop文件)
mop = MopacModel(mop_fname_set)
mop.opt4dragon(self.model_name)
def logBDG(self):
from .matrix.bdg import bdgX
# BDG 无温度参数
#CAS nN nHM O% MATS1e GATS1p GATS7p GGI1 GGI2 nCq nCrt C-040 H-048 H-051 O-059
abstract_value = self.dragon_model.extractparameter([
"nN", "nHM", "O%", "MATS1e", "GATS1p", "GATS7p", "GGI1", "GGI2",
"nCq", "nCrt", "C-040", "H-048", "H-051", "O-059"])
for smilenum in abstract_value.keys():
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
x = 1.9025 + \
1.0457 * abstract_value[smilenum]['nN'] + \
0.6662 * abstract_value[smilenum]['nHM'] - \
0.1078 * abstract_value[smilenum]['O%'] + \
2.8362 * abstract_value[smilenum]['MATS1e'] - \
2.0019 * abstract_value[smilenum]['GATS1p'] - \
0.7015 * abstract_value[smilenum]['GATS7p'] + \
0.1131 * abstract_value[smilenum]['GGI1'] + \
0.7023 * abstract_value[smilenum]['GGI2'] + \
2.7793 * abstract_value[smilenum]['nCq'] + \
1.035 * abstract_value[smilenum]['nCrt'] - \
0.777 * abstract_value[smilenum]['C-040'] - \
0.7091 * abstract_value[smilenum]['H-048'] - \
0.1553 * abstract_value[smilenum]['H-051'] + \
0.955 * abstract_value[smilenum]['O-059']
self.predict_result[smilenum]['logBDG']['value'] = self.round(x)
self.predict_result[smilenum]['logBDG']['degrade'] = self.round(1 / (1 + math.exp(-x)))
chemistry_logger.info('bdg(%s) dragon: %s' % (smilenum, abstract_value[smilenum]))
x = matrix([[abstract_value[smilenum]['nN'],
abstract_value[smilenum]['nHM'],
abstract_value[smilenum]['O%'],
abstract_value[smilenum]['MATS1e'],
abstract_value[smilenum]['GATS1p'],
abstract_value[smilenum]['GATS7p'],
abstract_value[smilenum]['GGI1'],
abstract_value[smilenum]['GGI2'],
abstract_value[smilenum]['nCq'],
abstract_value[smilenum]['nCrt'],
abstract_value[smilenum]['C-040'],
abstract_value[smilenum]['H-048'],
abstract_value[smilenum]['H-051'],
abstract_value[smilenum]['O-059']]])
williams = self.get_williams(bdgX, x)
self.predict_result[smilenum]['logBDG'].update(williams)
def logBCF(self):
from .matrix.bcf import bcfX
#BCF 无温度参数
#CAS NO. MLOGP2 F02[C-Cl] nROH P-117 Mor25m N% X4v O-058 LLS_01 H4v SM12_AEA(dm) O-057
abstract_value = self.dragon_model.extractparameter([
"MLOGP2", "F02[C-Cl]", "nROH", "P-117", "Mor25m", "N%", "X4v",
"O-058", "LLS_01", "H4v", "SM12_AEA(dm)", "O-057"
])
for smilenum in abstract_value:
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
value = 2.137 + \
0.061 * abstract_value[smilenum]['MLOGP2'] + \
0.034 * abstract_value[smilenum]['F02[C-Cl]'] - \
0.312 * abstract_value[smilenum]['nROH'] - \
1.282 * abstract_value[smilenum]['P-117'] + \
0.323 * abstract_value[smilenum]['Mor25m'] - \
0.052 * abstract_value[smilenum]['N%'] + \
0.080 * abstract_value[smilenum]['X4v'] - \
0.289 * abstract_value[smilenum]['O-058'] - \
1.137 * abstract_value[smilenum]['LLS_01'] - \
1.387 * abstract_value[smilenum]['H4v'] + \
0.071 * abstract_value[smilenum]['SM12_AEA(dm)'] - \
0.269 * abstract_value[smilenum]['O-057']
self.predict_result[smilenum]['logBCF']['value'] = self.round(
value)
chemistry_logger.info('bcf(%s) dragon: %s' %
(smilenum, abstract_value[smilenum]))
x = matrix([[
abstract_value[smilenum]['MLOGP2'],
abstract_value[smilenum]['F02[C-Cl]'],
abstract_value[smilenum]['nROH'],
abstract_value[smilenum]['P-117'],
abstract_value[smilenum]['Mor25m'],
abstract_value[smilenum]['N%'],
abstract_value[smilenum]['X4v'],
abstract_value[smilenum]['O-058'],
abstract_value[smilenum]['LLS_01'],
abstract_value[smilenum]['H4v'],
abstract_value[smilenum]['SM12_AEA(dm)'],
abstract_value[smilenum]['O-057']
]])
williams = self.get_williams(bcfX, x)
self.predict_result[smilenum]['logBCF'].update(williams)
def hide(request, id):
category = request.GET.get('category')
if category == 'suite':
rs = SuiteTask.objects.filter(sid=id)
elif category == 'single':
rs = SingleTask.objects.filter(pid=id)
else:
return HttpResponseRedirect('/history')
chemistry_logger.info('hide %s %s' % (category, id))
for r in rs:
r.is_hide = True
r.save()
return HttpResponseRedirect('/history')
def hide(request, id):
category = request.GET.get("category")
if category == "suite":
rs = SuiteTask.objects.filter(sid=id)
elif category == "single":
rs = SingleTask.objects.filter(pid=id)
else:
return HttpResponseRedirect("/history")
chemistry_logger.info("hide %s %s" % (category, id))
for r in rs:
r.is_hide = True
r.save()
return HttpResponseRedirect("/history")
def logO3(self):
from .matrix.o3 import o3X
#O3 有温度参数
#EHOMO, SaasC, SpPosA_B(m), nR=Cs, B01[C-C],
#X2Av, nCconj, nDB, N-076
abstract_value = self.dragon_model.extractparameter([
"EHOMO", "SaasC", "SpPosA_B(m)", "nR=Cs",
"B01[C-C]", "X2Av", "nCconj", "nDB", "N-076"
])
chemistry_logger.info(abstract_value)
for smilenum in abstract_value:
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
abstract_value[smilenum]['EHOMO'] = fetch_ehomo_by_mopac(smilenum, 'logO3')
value = 2.549 + \
1.534 * abstract_value[smilenum]['EHOMO'] - \
1.284 * abstract_value[smilenum]['SaasC'] - \
2.092 * abstract_value[smilenum]['SpPosA_B(m)'] + \
0.3060 * abstract_value[smilenum]['nR=Cs'] + \
1.113 * abstract_value[smilenum]['B01[C-C]'] - \
2.289 * abstract_value[smilenum]['X2Av'] - \
0.3470 * abstract_value[smilenum]['nCconj'] + \
0.5960 * abstract_value[smilenum]['nDB'] - \
1.301 * abstract_value[smilenum]['N-076'] - \
878.7 / self.T
self.predict_result[smilenum]['logO3']['value'] = self.round(value)
chemistry_logger.info('O3(%s) dragon: %s' % (smilenum, abstract_value[smilenum]))
x = matrix([[1.0 / self.T,
abstract_value[smilenum]['EHOMO'],
abstract_value[smilenum]['nDB'],
abstract_value[smilenum]['SaasC'],
abstract_value[smilenum]['SpPosA_B(m)'],
abstract_value[smilenum]['nCconj'],
abstract_value[smilenum]['nR=Cs'],
abstract_value[smilenum]['B01[C-C]'],
abstract_value[smilenum]['X2Av'],
abstract_value[smilenum]['N-076'],
]])
williams = self.get_williams(o3X, x)
self.predict_result[smilenum]['logO3'].update(williams)
def logKOA(self):
from .matrix.koa import koaX
#KOA 有温度参数
#X1sol Mor13v H050 R5v TO..Cl HATS5v RDF035m RCl nRCOOR Mor15u RDF90m 1/T
abstract_value = self.dragon_model.extractparameter([
"X1sol", "Mor13v", "H-050", "R5v",
"T(O..Cl)", "HATS5v", "RDF035m",
"RCI", "nRCOOR", "Mor15u",
"RDF090m"])
for smilenum in abstract_value:
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
value = -3.03 + \
313.0 * abstract_value[smilenum]['X1sol'] / self.T - \
85.7 * abstract_value[smilenum]['Mor13v'] / self.T + \
432.0 * abstract_value[smilenum]['H-050'] / self.T - \
1270.0 * abstract_value[smilenum]['R5v'] / self.T - \
5.54 * abstract_value[smilenum]['T(O..Cl)'] / self.T + \
125.0 * abstract_value[smilenum]['HATS5v'] / self.T - \
13.3 * abstract_value[smilenum]['RDF035m'] / self.T - \
61.1 * abstract_value[smilenum]['RCI'] / self.T - \
37.6 * abstract_value[smilenum]['nRCOOR'] / self.T + \
156.0 * abstract_value[smilenum]['Mor15u'] / self.T -\
5.49 * abstract_value[smilenum]['RDF090m'] / self.T + \
1040.0 / self.T
self.predict_result[smilenum]['logKOA']['value'] = self.round(value)
chemistry_logger.info('koa(%s) dragon: %s' % (smilenum, abstract_value[smilenum]))
x = matrix([[abstract_value[smilenum]['X1sol'] / self.T,
abstract_value[smilenum]['Mor13v'] / self.T,
abstract_value[smilenum]['H-050'] / self.T,
abstract_value[smilenum]['R5v'] / self.T,
abstract_value[smilenum]['T(O..Cl)'] / self.T,
abstract_value[smilenum]['HATS5v'] / self.T,
abstract_value[smilenum]['RDF035m'] / self.T,
abstract_value[smilenum]['RCI'] / self.T,
abstract_value[smilenum]['nRCOOR'] / self.T,
abstract_value[smilenum]['Mor15u'] / self.T,
abstract_value[smilenum]['RDF090m'] / self.T,
1.0 / self.T,
]])
williams = self.get_williams(koaX, x)
self.predict_result[smilenum]['logKOA'].update(williams)
def submit_calculate_task(user, smile=None, draw_mol_data=None,
task_notes=None, task_name=None,
files_id_list=None, models=None,
local_search_id=None):
chemistry_logger.info("smile: %s" % smile)
chemistry_logger.info("draw_mol_data: %s" % draw_mol_data)
chemistry_logger.info("files_id_list: %s" % files_id_list)
chemistry_logger.info("models: %s" % models)
tasks_num = calculate_tasks(files_id_list, smile, draw_mol_data, models)
if tasks_num == 0:
status = False
info = "请至少选择一种输入方式和计算模型!"
id = None
return (status, info, id)
try:
s = SuiteTask()
s.sid = id = str(uuid.uuid4())
s.user = UserProfile.objects.get(user=user)
s.total_tasks = tasks_num
s.has_finished_tasks = 0
s.start_time = utils.get_real_now()
s.name = task_name
s.notes = task_notes
s.models_str, s.models_category_str = parse_models(models)
s.status = StatusCategory.objects.get(category=STATUS_WORKING)
s.email = user.email
s.save()
chemistry_logger.info('~~~~~~~~ s:%s' % s.start_time)
generate_calculate_task.delay(models, smile, draw_mol_data,
files_id_list, id, local_search_id)
except:
chemistry_logger.exception('failed to generate suite_task')
s.delete()
status = False
info = "计算任务添加不成功，将重试或联系网站管理员!"
id = None
else:
status = True
info = "恭喜,计算任务已经提交!"
return (status, info, id)
def logKOH(self):
from .matrix.koh_298k import koh_298kX
# KOH 298K预测模型, 无温度参数
#CAS-Number EHOMO AMW NdsCH Mor14i nP nR=Cp X% nRCHO C-020 SpMaxA_AEA(dm) nCbH CATS2D_03_DL
abstract_value = self.dragon_model.extractparameter([
"EHOMO", "AMW", "NdsCH", "Mor14i", "nP", "nR=Cp",
"X%", "nRCHO", "C-020", "SpMaxA_AEA(dm)",
"nCbH", "CATS2D_03_DL"])
for smilenum in abstract_value:
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
abstract_value[smilenum]['EHOMO'] = fetch_ehomo(smilenum, 'logKOH')
value = -6.511 + \
15.85 * abstract_value[smilenum]['EHOMO'] - \
0.03800 * abstract_value[smilenum]['AMW'] + \
0.1300 * abstract_value[smilenum]['NdsCH'] + \
0.1630 * abstract_value[smilenum]['Mor14i'] + \
0.7790 * abstract_value[smilenum]['nP'] + \
0.3170 * abstract_value[smilenum]['nR=Cp'] - \
0.01900 * abstract_value[smilenum]['X%'] + \
0.3930 * abstract_value[smilenum]['nRCHO'] + \
0.5890 * abstract_value[smilenum]['C-020'] - \
0.4550 * abstract_value[smilenum]['SpMaxA_AEA(dm)'] - \
0.05600 * abstract_value[smilenum]['nCbH'] + \
0.1410 * abstract_value[smilenum]['CATS2D_03_DL']
self.predict_result[smilenum]['logKOH']['value'] = self.round(value)
chemistry_logger.info('koh(%s) dragon: %s' % (smilenum, abstract_value[smilenum]))
x = matrix([[abstract_value[smilenum]['EHOMO'],
abstract_value[smilenum]['AMW'],
abstract_value[smilenum]['NdsCH'],
abstract_value[smilenum]['Mor14i'],
abstract_value[smilenum]['nP'],
abstract_value[smilenum]['nR=Cp'],
abstract_value[smilenum]['X%'],
abstract_value[smilenum]['nRCHO'],
abstract_value[smilenum]['C-020'],
abstract_value[smilenum]['SpMaxA_AEA(dm)'],
abstract_value[smilenum]['nCbH'],
abstract_value[smilenum]['CATS2D_03_DL']]])
williams = self.get_williams(koh_298kX, x)
self.predict_result[smilenum]['logKOH'].update(williams)
def gjf4dragon(self, model_name):
for name in self.gjf_fname_list_no_ext:
mol_path = join(CALCULATE_DATA_PATH.DRAGON, model_name, name,
'%s.mol' % name)
gjf_path = join(CALCULATE_DATA_PATH.GAUSSIAN, model_name, name,
'%s.gjf' % name)
log_path = join(CALCULATE_DATA_PATH.GAUSSIAN, model_name, name,
'%s.log' % name)
cmd = '%s "%s"' % (CALCULATE_CMD_TYPE.GAUSSIAN, gjf_path)
chemistry_logger.info('gif4dragon part1 cmd: %s' % cmd)
# 此处需要注意g09输出是.log文件
CalcoreCmd(cmd, output=log_path).run()
cmd = 'obabel -ig09 "%s" -omol -O "%s"' % (log_path, mol_path)
chemistry_logger.info('gif4dragon part2 cmd: %s' % cmd)
CalcoreCmd(cmd, output=mol_path).run()
def logKOC(self):
from .matrix.koc import kocX
# KOC 使用g09，无温度参数
#CAS No. nN ATSC8v SpMaxA_G/D Mor16u nROH O-058 P-117 MLOGP2 Molecular Polarizability
abstract_value = self.dragon_model.extractparameter([
"nN", "ATSC8v", "SpMaxA_G/D", "Mor16u", "nROH", "O-058", "P-117",
"MLOGP2", "α"
])
for smilenum in abstract_value:
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
abstract_value[smilenum]['α'] = fetch_polarizability(
smilenum, 'logKOC')
value = -1.612 + \
0.039 * abstract_value[smilenum]['MLOGP2'] +\
0.010 * abstract_value[smilenum]["α"] -\
0.342 * abstract_value[smilenum]['O-058'] -\
0.069 * abstract_value[smilenum]['ATSC8v'] -\
0.123 * abstract_value[smilenum]['nN'] -\
0.368 * abstract_value[smilenum]['nROH'] -\
0.473 * abstract_value[smilenum]['P-117'] +\
2.335 * abstract_value[smilenum]['SpMaxA_G/D'] +\
0.302 * abstract_value[smilenum]['Mor16u']
self.predict_result[smilenum]['logKOC']['value'] = self.round(
value)
chemistry_logger.info('koc(%s) dragon: %s' %
(smilenum, abstract_value[smilenum]))
x = matrix([[
abstract_value[smilenum]['nN'],
abstract_value[smilenum]['ATSC8v'],
abstract_value[smilenum]['SpMaxA_G/D'],
abstract_value[smilenum]['Mor16u'],
abstract_value[smilenum]['nROH'],
abstract_value[smilenum]['O-058'],
abstract_value[smilenum]['P-117'],
abstract_value[smilenum]['MLOGP2'],
abstract_value[smilenum]["α"]
]])
williams = self.get_williams(kocX, x)
self.predict_result[smilenum]['logKOC'].update(williams)
def opt4dragon(self, model_name):
for name in self.mop_fname_list_no_ext:
mol_path = join(CALCULATE_DATA_PATH.DRAGON, model_name, name,
'%s.mol' % name)
mop_path = join(CALCULATE_DATA_PATH.MOPAC, model_name, name,
'%s.mop' % name)
out_path = join(CALCULATE_DATA_PATH.MOPAC, model_name, name,
'%s.out' % name)
cmd = '%s "%s"' % (CALCULATE_CMD_TYPE.MOPAC, mop_path)
chemistry_logger.info('opt4dragon part1 cmd: %s' % cmd)
# 此处输出是.out文件
CalcoreCmd(cmd, output=out_path).run()
cmd = 'obabel -imoo "%s" -omol -O "%s" --gen3D' % (out_path,
mol_path)
chemistry_logger.info('opt4dragon part2 cmd: %s' % cmd)
CalcoreCmd(cmd, output=mol_path).run()
def logBCF(self):
from .matrix.bcf import bcfX
#BCF 无温度参数
#CAS NO. MLOGP2 F02[C-Cl] nROH P-117 Mor25m N% X4v O-058 LLS_01 H4v SM12_AEA(dm) O-057
abstract_value = self.dragon_model.extractparameter([
"MLOGP2", "F02[C-Cl]", "nROH",
"P-117", "Mor25m", "N%", "X4v",
"O-058", "LLS_01", "H4v", "SM12_AEA(dm)", "O-057"])
for smilenum in abstract_value:
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
value = 2.137 + \
0.061 * abstract_value[smilenum]['MLOGP2'] + \
0.034 * abstract_value[smilenum]['F02[C-Cl]'] - \
0.312 * abstract_value[smilenum]['nROH'] - \
1.282 * abstract_value[smilenum]['P-117'] + \
0.323 * abstract_value[smilenum]['Mor25m'] - \
0.052 * abstract_value[smilenum]['N%'] + \
0.080 * abstract_value[smilenum]['X4v'] - \
0.289 * abstract_value[smilenum]['O-058'] - \
1.137 * abstract_value[smilenum]['LLS_01'] - \
1.387 * abstract_value[smilenum]['H4v'] + \
0.071 * abstract_value[smilenum]['SM12_AEA(dm)'] - \
0.269 * abstract_value[smilenum]['O-057']
self.predict_result[smilenum]['logBCF']['value'] = self.round(value)
chemistry_logger.info('bcf(%s) dragon: %s' % (smilenum, abstract_value[smilenum]))
x = matrix([[abstract_value[smilenum]['MLOGP2'],
abstract_value[smilenum]['F02[C-Cl]'],
abstract_value[smilenum]['nROH'],
abstract_value[smilenum]['P-117'],
abstract_value[smilenum]['Mor25m'],
abstract_value[smilenum]['N%'],
abstract_value[smilenum]['X4v'],
abstract_value[smilenum]['O-058'],
abstract_value[smilenum]['LLS_01'],
abstract_value[smilenum]['H4v'],
abstract_value[smilenum]['SM12_AEA(dm)'],
abstract_value[smilenum]['O-057']]])
williams = self.get_williams(bcfX, x)
self.predict_result[smilenum]['logBCF'].update(williams)
def generate_mol_image(singletask):
"""generate smile and image for task"""
chemistry_logger.info('generate smile image %s' % singletask.pid)
fpath = settings.SETTINGS_ROOT + singletask.file_obj.file_obj.url
mol = pybel.readfile("mol", fpath).next()
singletask.file_obj.smiles = ("%s" % mol).split("\t")[0]
pname = str(uuid.uuid4()) + ".png"
if not os.path.exists(settings.SEARCH_IMAGE_PATH):
os.makedirs(settings.SEARCH_IMAGE_PATH)
ppath = os.path.join(settings.SEARCH_IMAGE_PATH, pname)
mol.draw(show=False, filename=ppath)
f = File(open(ppath, "r"))
singletask.file_obj.image = f
singletask.file_obj.save()
singletask.save()
f.close()
def save_record(f, model, sid, source_type, smile=None, local_search_id=None):
from chemistry.tasks import calculateTask, DEFAULT_TEMPERATURE_ARGS
task = SingleTask()
task.start_time = utils.get_real_now()
task.sid = SuiteTask.objects.get(sid=sid)
task.pid = str(uuid.uuid4())
task.model = ModelCategory.objects.get(category=model['model'])
temperature = model.get('temperature')
task.temperature = float(temperature) if temperature else DEFAULT_TEMPERATURE_ARGS
if source_type == ORIGIN_UPLOAD:
# here, f is ProcessedFile record instance
f.file_source = FileSourceCategory.objects.get(category=source_type)
f.file_type = "mol"
task.file_obj = f
f.save()
elif source_type in (ORIGIN_SMILE, ORIGIN_DRAW):
# here, f is a file path
processed_f = ProcessedFile()
obj = File(open(f, "r"))
processed_f.title = os.path.basename(obj.name)
processed_f.file_type = source_type
processed_f.file_source = FileSourceCategory.objects.get(category=source_type)
processed_f.file_obj = obj
if smile:
processed_f.smiles = smile
if source_type == ORIGIN_SMILE and local_search_id is not None:
processed_f.local_search_id = int(local_search_id)
processed_f.save()
task.file_obj = processed_f
obj.close()
else:
raise ErrorCalculateType('Cannot recongize this source type')
task.status = StatusCategory.objects.get(category=STATUS_WORKING)
task.save()
chemistry_logger.info('~~~~~~~~ t:%s' % task.start_time)
calculateTask.delay(task, sid, model)
def fetch_ehomo(name, model):
path = os.path.join(config.GAUSSIAN_PATH, model, name, '%s.log' % name)
if not os.path.exists(path):
chemistry_logger.error('Cannot fetch H**O %s' % path)
return 0.0
ret = []
with open(path, 'r') as f:
for line in f.readlines():
if EHOMO_RE.search(line):
chemistry_logger.info(line)
ret.append(line)
try:
v = float(ret[-1].split()[-1])
except Exception as e:
chemistry_logger.exception('Cannot get EHOMO value %s ' % line)
v = 0
return v
def fetch_ehomo(name, model):
path = os.path.join(config.GAUSSIAN_PATH, model, name, '%s.log' % name)
if not os.path.exists(path):
chemistry_logger.error('Cannot fetch H**O %s' % path)
return 0.0
ret = []
with open(path, 'r') as f:
for line in f.readlines():
if EHOMO_RE.search(line):
chemistry_logger.info(line)
ret.append(line)
try:
v = float(ret[-1].split()[-1])
except Exception as e:
chemistry_logger.exception('Cannot get EHOMO value %s ' % line)
v = 0
return v
def add_counter(suite_id):
"""
use filter to get task numbers
"""
finished_count = SingleTask.objects.filter(sid=suite_id)\
.exclude(status=StatusCategory.objects.get(category=STATUS_WORKING))\
.count()
suite = SuiteTask.objects.get(sid=suite_id)
if finished_count == suite.total_tasks:
suite.has_finished_tasks = suite.total_tasks
suite.status_id = StatusCategory.objects.get(category=STATUS_SUCCESS)
suite.end_time = utils.get_real_now()
suite.save()
send_email_task(suite.email, suite.sid)
else:
suite.has_finished_tasks = finished_count
suite.save()
chemistry_logger.info('~~~~~~~~ s-e:%s' % suite.end_time)
def add_counter(suite_id):
"""
use filter to get task numbers
"""
finished_count = SingleTask.objects.filter(sid=suite_id)\
.exclude(status=StatusCategory.objects.get(category=STATUS_WORKING))\
.count()
suite = SuiteTask.objects.get(sid=suite_id)
if finished_count == suite.total_tasks:
suite.has_finished_tasks = suite.total_tasks
suite.status_id = StatusCategory.objects.get(category=STATUS_SUCCESS)
suite.end_time = utils.get_real_now()
suite.save()
send_email_task(suite.email, suite.sid)
else:
suite.has_finished_tasks = finished_count
suite.save()
chemistry_logger.info('~~~~~~~~ s-e:%s' % suite.end_time)
def calculateTask(task, sid, model):
try:
generate_mol_image(task)
suite = SuiteTask.objects.get(sid=sid)
map_model_name = get_model_name(model['model'])
smile = task.file_obj.smiles.encode(
'utf-8') if task.file_obj.file_type != 'mol' else ''
# smile, mol_fpath 输入只选择一种方式(优先smile)
mol_fpath = os.path.join(
settings.SETTINGS_ROOT,
task.file_obj.file_obj.path) if not smile else None
try:
temperature = float(model.get('temperature'))
except:
temperature = DEFAULT_TEMPERATURE_ARGS
chemistry_logger.info(
'PredictionModel calculating: model name(%s),'
'smile(%s) mol path(%s) temperature(%s)', map_model_name, smile,
mol_fpath, temperature)
# 后台模型计算入口
predict_results = prediction_model_calculate(map_model_name, smile,
mol_fpath, temperature)
if task.file_obj.file_type == 'mol':
name = os.path.basename(mol_fpath).split('.')[0]
else:
name = smile
result = predict_results[name][map_model_name]
chemistry_logger.info('[task]result %s' % result)
except KeyError:
chemistry_logger.exception('still cannot support this model')
result = None
task.result_state = "We don't support this model now"
task.status = StatusCategory.objects.get(category=STATUS_FAILED)
suite.status_id = StatusCategory.objects.get(category=STATUS_FAILED)
except Exception as e:
chemistry_logger.exception('failed to submit task to prediction model')
result = None
task.result_state = str(e)
task.status = StatusCategory.objects.get(category=STATUS_FAILED)
suite.status_id = StatusCategory.objects.get(category=STATUS_FAILED)
else:
chemistry_logger.info("calculate Successfully in celery queue!")
task.result_state = "Calculate Successfully!"
task.status = StatusCategory.objects.get(category=STATUS_SUCCESS)
suite.status_id = StatusCategory.objects.get(category=STATUS_WORKING)
task.end_time = utils.get_real_now()
task.results = json.dumps(result)
suite.save()
task.save()
add_counter(suite.sid)
return result
def logKOC(self):
from .matrix.koc import kocX
# KOC 使用g09，无温度参数
#CAS No. nN ATSC8v SpMaxA_G/D Mor16u nROH O-058 P-117 MLOGP2 Molecular Polarizability
abstract_value = self.dragon_model.extractparameter([
"nN", "ATSC8v", "SpMaxA_G/D", "Mor16u", "nROH",
"O-058", "P-117", "MLOGP2", "α"])
for smilenum in abstract_value:
if smilenum not in self.predict_result:
self.predict_result[smilenum] = defaultdict(dict)
abstract_value[smilenum]['α'] = fetch_polarizability(smilenum, 'logKOC')
value = -1.612 + \
0.039 * abstract_value[smilenum]['MLOGP2'] +\
0.010 * abstract_value[smilenum]["α"] -\
0.342 * abstract_value[smilenum]['O-058'] -\
0.069 * abstract_value[smilenum]['ATSC8v'] -\
0.123 * abstract_value[smilenum]['nN'] -\
0.368 * abstract_value[smilenum]['nROH'] -\
0.473 * abstract_value[smilenum]['P-117'] +\
2.335 * abstract_value[smilenum]['SpMaxA_G/D'] +\
0.302 * abstract_value[smilenum]['Mor16u']
self.predict_result[smilenum]['logKOC']['value'] = self.round(value)
chemistry_logger.info('koc(%s) dragon: %s' % (smilenum, abstract_value[smilenum]))
x = matrix([[abstract_value[smilenum]['nN'],
abstract_value[smilenum]['ATSC8v'],
abstract_value[smilenum]['SpMaxA_G/D'],
abstract_value[smilenum]['Mor16u'],
abstract_value[smilenum]['nROH'],
abstract_value[smilenum]['O-058'],
abstract_value[smilenum]['P-117'],
abstract_value[smilenum]['MLOGP2'],
abstract_value[smilenum]["α"]]])
williams = self.get_williams(kocX, x)
self.predict_result[smilenum]['logKOC'].update(williams)
def mol2gjf2dragon_folder(self):
gaussian_files_set = set()
for element in self.iter_smiles_files(self.__smilenum_list, 'smile'):
smile, name, dragon_dpath, mopac_dpath, mop_fpath = element
gaussian_dpath = join(CALCULATE_DATA_PATH.GAUSSIAN,
self.model_name, name)
# smile-> mol
mol_fpath = self.smile2_3d(smile)
# mol -> gjf file
gjf_fpath = mol2gjf(mol_fpath, self.model_name)
if not os.path.exists(dragon_dpath):
os.makedirs(dragon_dpath)
if not os.path.exists(gaussian_dpath):
os.makedirs(gaussian_dpath)
shutil.copy(mol_fpath, dragon_dpath)
shutil.copy(gjf_fpath, gaussian_dpath)
gaussian_files_set.add('%s.gjf' % name)
for element in self.iter_smiles_files(self.__molfile, 'file'):
mol_fpath, name, dragon_dpath, mopac_dpath, mop_fpath = element
if not os.path.exists(dragon_dpath):
os.makedirs(dragon_dpath)
shutil.copy(mol_fpath, dragon_dpath)
gaussian_files_set.add('%s.gjf' % name)
#FIXME: gaussian 计算很慢吗?
chemistry_logger.info('GaussianOptimizeModel gjf4dragon')
gjf = GaussianOptimizeModel(gaussian_files_set)
gjf.gjf4dragon(self.model_name)
def mol2gjf2dragon_folder(self):
gaussian_files_set = set()
for element in self.iter_smiles_files(self.__smilenum_list, 'smile'):
smile, name, dragon_dpath, mopac_dpath, mop_fpath = element
gaussian_dpath = join(CALCULATE_DATA_PATH.GAUSSIAN, self.model_name, name)
# smile-> mol
mol_fpath = self.smile2_3d(smile)
# mol -> gjf file
gjf_fpath = mol2gjf(mol_fpath, self.model_name)
if not os.path.exists(dragon_dpath):
os.makedirs(dragon_dpath)
if not os.path.exists(gaussian_dpath):
os.makedirs(gaussian_dpath)
shutil.copy(mol_fpath, dragon_dpath)
shutil.copy(gjf_fpath, gaussian_dpath)
gaussian_files_set.add('%s.gjf' % name)
for element in self.iter_smiles_files(self.__molfile, 'file'):
mol_fpath, name, dragon_dpath, mopac_dpath, mop_fpath = element
if not os.path.exists(dragon_dpath):
os.makedirs(dragon_dpath)
shutil.copy(mol_fpath, dragon_dpath)
gaussian_files_set.add('%s.gjf' % name)
#FIXME: gaussian 计算很慢吗?
chemistry_logger.info('GaussianOptimizeModel gjf4dragon')
gjf = GaussianOptimizeModel(gaussian_files_set)
gjf.gjf4dragon(self.model_name)
def calculateTask(task, sid, model):
try:
generate_mol_image(task)
suite = SuiteTask.objects.get(sid=sid)
map_model_name = get_model_name(model['model'])
smile = task.file_obj.smiles.encode('utf-8') if task.file_obj.file_type != 'mol' else ''
# smile, mol_fpath 输入只选择一种方式(优先smile)
mol_fpath = os.path.join(settings.SETTINGS_ROOT, task.file_obj.file_obj.path) if not smile else None
try:
temperature = float(model.get('temperature'))
except:
temperature = DEFAULT_TEMPERATURE_ARGS
chemistry_logger.info('PredictionModel calculating: model name(%s),'
'smile(%s) mol path(%s) temperature(%s)',
map_model_name, smile, mol_fpath, temperature)
# 后台模型计算入口
predict_results = prediction_model_calculate(map_model_name, smile,
mol_fpath, temperature)
if task.file_obj.file_type == 'mol':
name = os.path.basename(mol_fpath).split('.')[0]
else:
name = smile
result = predict_results[name][map_model_name]
chemistry_logger.info('[task]result %s' % result)
except KeyError:
chemistry_logger.exception('still cannot support this model')
result = None
task.result_state = "We don't support this model now"
task.status = StatusCategory.objects.get(category=STATUS_FAILED)
suite.status_id = StatusCategory.objects.get(category=STATUS_FAILED)
except Exception as e:
chemistry_logger.exception('failed to submit task to prediction model')
result = None
task.result_state = str(e)
task.status = StatusCategory.objects.get(category=STATUS_FAILED)
suite.status_id = StatusCategory.objects.get(category=STATUS_FAILED)
else:
chemistry_logger.info("calculate Successfully in celery queue!")
task.result_state = "Calculate Successfully!"
task.status = StatusCategory.objects.get(category=STATUS_SUCCESS)
suite.status_id = StatusCategory.objects.get(category=STATUS_WORKING)
task.end_time = utils.get_real_now()
task.results = json.dumps(result)
suite.save()
task.save()
add_counter(suite.sid)
return result
def run(self):
#FIXME: 运行时候加锁
if not self.check_output_exists():
chemistry_logger.info('[CalcoreCmd]%s(%s->%s)', self.cmd, self.input, self.output)
s= check_call(self.cmd, shell=True)
chemistry_logger.info(s)
def mol2gjf(fpath, model_name):
if model_name in ('logKOH', 'logKOH_T'):
element = dict.fromkeys([
'H', 'C', 'N', 'O', 'F', 'P', 'S', 'Cl', 'Se', 'Br', 'I', 'Si',
'Hg', 'Pb'
], 0)
fname = os.path.basename(fpath)
fname_no_ext = fname.split('.')[0]
content = []
with open(fpath, 'r') as f:
for line in f.readlines():
try:
values = line.split()
if 'A' < values[3] < 'Z':
content.append(
' %s %s %s %s\n' %
(values[3], values[0], values[1], values[2]))
if model_name in ('logKOH', 'logKOH_T'):
element[values[3]] = 1
except Exception:
pass
gjf_list = []
gjf_list.append('%%chk=%s.chk\n' % fname_no_ext)
gjf_list.append('%nproc=2\n')
gjf_list.append('%mem=2GB\n')
if model_name in ('logKOH', 'logKOH_T'):
element_op = (element['I'] | element['Si'] | element['Hg']
| element['Pb']) == 1
if element_op:
gjf_list.append(
'#p opt freq b3lyp/genecp scf=tight int=ultrafine\n')
else:
gjf_list.append(
'#p opt freq b3lyp/6-311+G(d,p) scf=tight int=ultrafine\n')
elif model_name in ("logKOC", "logBCF"):
gjf_list.append(
'#p opt freq b3lyp/6-31+g(d,p) SCRF=(IEFPCM,SOLVENT=WATER)\n')
gjf_list.append('\n')
gjf_list.append('Title Card Required\n')
gjf_list.append('\n')
gjf_list.append('0 1\n')
gjf_list.extend(content)
gjf_list.append('\n')
if model_name in ('logKOH', 'logKOH_T') and element_op:
tempC = ""
tempHg = ""
for t in ('I', 'Si', 'Hg', 'Pb'):
if element[t] == 1:
element[t] = 0
tempHg += "%s " % t
for key in element:
if element[key] == 1:
tempC += "%s " % key
gjf_list.append('%s 0\n' % tempC)
gjf_list.append('6-31+g(d,p)\n')
gjf_list.append('****\n')
gjf_list.append('%s 0\n' % tempHg)
gjf_list.append('LANL2DZ\n')
gjf_list.append('****\n\n')
gjf_list.append('%s 0\n' % tempHg)
gjf_list.append('LANL2DZ\n\n')
gjf_fpath = join(settings.MOL_ABSTRACT_FILE_PATH, '%s.gjf' % fname_no_ext)
chemistry_logger.info('mol->gjf gjf path: %s' % gjf_fpath)
chemistry_logger.info('mol->gjf, content: %s' % gjf_list)
with open(gjf_fpath, 'w') as f:
f.writelines(tuple(gjf_list))
return gjf_fpath
def prediction_model_calculate(model_name, smile, mol_fpath, temperature):
pm = PredictionModel(model_name, smile, mol_fpath, T=temperature)
pm.calculate(model_name)
chemistry_logger.info("pm result:%s", pm.predict_result)
return pm.predict_result
def extractparameter(self, parameters=None):
'''从drs文件中将对应参数名列表中对应的描述符名称的值提取出来，返回的是一个字典'''
firsttraverse = True
para_dic = {}
# record para position in drs file
temp_dic = {}
for raw_name, fname, input_fpath, drs_input_fpath, output_fpath in self.iter_files(
):
para_dic[raw_name] = {p: 0 for p in parameters}
with open(output_fpath, 'r') as fp:
lines = fp.readlines()
paraline = lines[0].split()
valueline = lines[1].split()
if firsttraverse:
firsttraverse = False
for i in range(len(paraline)):
if paraline[i] in para_dic[raw_name]:
temp_dic[paraline[i]] = i
para_dic[raw_name][paraline[i]] = float(valueline[i])
else:
for key in temp_dic.keys():
try:
para_dic[raw_name][key] = float(
valueline[temp_dic[key]])
except:
print key, temp_dic[i], valueline[temp_dic[key]]
if self.model_name in ('logKOH', 'logKOH_T'):
f_log = join(CALCULATE_DATA_PATH.GAUSSIAN, self.model_name,
raw_name, '%s.log' % raw_name)
f = open(f_log, 'r')
lines = f.readlines()
f.close()
regex = '.*Alpha occ. eigenvalues.*'
for lineNum in range(len(lines)):
if re.match(regex, lines[lineNum]):
while (re.match(regex, lines[lineNum])):
lineNum = lineNum + 1
List = list(lines[lineNum].split(' '))
while (1):
try:
List.remove('')
except:
break
EHOMO = lines[lineNum - 1].split(' ')[-1]
para_dic[raw_name]["EHOMO"] = float(EHOMO)
break
elif self.model_name in ('logPL', ):
f_out = join(CALCULATE_DATA_PATH.MOPAC, self.model_name,
raw_name, '%s.out' % raw_name)
f = open(f_out, 'r')
lines = f.readlines()
f.close()
regex = '.*ATOM NO\..*TYPE.*CHARGE.*No\.'
j = 0
for lineNum in range(len(lines)):
if re.match(regex, lines[lineNum]):
chemistry_logger.info('PL out %s' % lines[lineNum])
while (not re.match('.*DIPOLE.*', lines[lineNum + j])):
j = j + 1
chemistry_logger.info('PL out2 %s' %
lines[lineNum + j + 3])
para_dic[raw_name]["u"] = float(lines[lineNum + j +
3].split()[-1])
break
return para_dic
def extractparameter(self, parameters=None):
'''从drs文件中将对应参数名列表中对应的描述符名称的值提取出来，返回的是一个字典'''
firsttraverse = True
para_dic = {}
# record para position in drs file
temp_dic = {}
for raw_name, fname, input_fpath, drs_input_fpath, output_fpath in self.iter_files():
para_dic[raw_name] = {p: 0 for p in parameters}
with open(output_fpath, 'r') as fp:
lines = fp.readlines()
paraline = lines[0].split()
valueline = lines[1].split()
if firsttraverse:
firsttraverse = False
for i in range(len(paraline)):
if paraline[i] in para_dic[raw_name]:
temp_dic[paraline[i]] = i
para_dic[raw_name][paraline[i]] = float(valueline[i])
else:
for key in temp_dic.keys():
try:
para_dic[raw_name][key] = float(valueline[temp_dic[key]])
except:
print key, temp_dic[i], valueline[temp_dic[key]]
if self.model_name in ('logKOH', 'logKOH_T'):
f_log = join(CALCULATE_DATA_PATH.GAUSSIAN, self.model_name, raw_name,
'%s.log' % raw_name)
f = open(f_log, 'r')
lines = f.readlines()
f.close()
regex = '.*Alpha occ. eigenvalues.*'
for lineNum in range(len(lines)):
if re.match(regex, lines[lineNum]):
while(re.match(regex, lines[lineNum])):
lineNum = lineNum + 1
List = list(lines[lineNum].split(' '))
while(1):
try:
List.remove('')
except:
break
EHOMO = lines[lineNum - 1].split(' ')[-1]
para_dic[raw_name]["EHOMO"] = float(EHOMO)
break
elif self.model_name in ('logPL',):
f_out = join(CALCULATE_DATA_PATH.MOPAC, self.model_name, raw_name,
'%s.out' % raw_name)
f = open(f_out, 'r')
lines = f.readlines()
f.close()
regex = '.*ATOM NO\..*TYPE.*CHARGE.*No\.'
j = 0
for lineNum in range(len(lines)):
if re.match(regex, lines[lineNum]):
chemistry_logger.info('PL out %s' % lines[lineNum])
while(not re.match('.*DIPOLE.*', lines[lineNum + j])):
j = j + 1
chemistry_logger.info('PL out2 %s' % lines[lineNum + j + 3])
para_dic[raw_name]["u"] = float(lines[lineNum + j + 3].split()[-1])
break
return para_dic
def prediction_model_calculate(model_name, smile, mol_fpath, temperature):
pm = PredictionModel(model_name, smile, mol_fpath, T=temperature)
pm.calculate(model_name)
chemistry_logger.info("pm result:%s", pm.predict_result)
return pm.predict_result
def mol2gjf(fpath, model_name):
if model_name in ('logKOH', 'logKOH_T'):
element = dict.fromkeys([
'H', 'C', 'N', 'O', 'F', 'P', 'S', 'Cl',
'Se', 'Br', 'I', 'Si', 'Hg', 'Pb'], 0)
fname = os.path.basename(fpath)
fname_no_ext = fname.split('.')[0]
content = []
with open(fpath, 'r') as f:
for line in f.readlines():
try:
values = line.split()
if 'A' < values[3] < 'Z':
content.append(' %s %s %s %s\n' % (values[3], values[0],
values[1], values[2]))
if model_name in ('logKOH', 'logKOH_T'):
element[values[3]] = 1
except Exception:
pass
gjf_list = []
gjf_list.append('%%chk=%s.chk\n' % fname_no_ext)
gjf_list.append('%nproc=2\n')
gjf_list.append('%mem=2GB\n')
if model_name in ('logKOH', 'logKOH_T'):
element_op = (element['I'] | element['Si'] | element['Hg'] |
element['Pb']) == 1
if element_op:
gjf_list.append('#p opt freq b3lyp/genecp scf=tight int=ultrafine\n')
else:
gjf_list.append('#p opt freq b3lyp/6-311+G(d,p) scf=tight int=ultrafine\n')
elif model_name in ("logKOC", "logBCF"):
gjf_list.append('#p opt freq b3lyp/6-31+g(d,p) SCRF=(IEFPCM,SOLVENT=WATER)\n')
gjf_list.append('\n')
gjf_list.append('Title Card Required\n')
gjf_list.append('\n')
gjf_list.append('0 1\n')
gjf_list.extend(content)
gjf_list.append('\n')
if model_name in ('logKOH', 'logKOH_T') and element_op:
tempC = ""
tempHg = ""
for t in ('I', 'Si', 'Hg', 'Pb'):
if element[t] == 1:
element[t] = 0
tempHg += "%s " % t
for key in element:
if element[key] == 1:
tempC += "%s " % key
gjf_list.append('%s 0\n' % tempC)
gjf_list.append('6-31+g(d,p)\n')
gjf_list.append('****\n')
gjf_list.append('%s 0\n' % tempHg)
gjf_list.append('LANL2DZ\n')
gjf_list.append('****\n\n')
gjf_list.append('%s 0\n' % tempHg)
gjf_list.append('LANL2DZ\n\n')
gjf_fpath = join(settings.MOL_ABSTRACT_FILE_PATH, '%s.gjf' % fname_no_ext)
chemistry_logger.info('mol->gjf gjf path: %s' % gjf_fpath)
chemistry_logger.info('mol->gjf, content: %s' % gjf_list)
with open(gjf_fpath, 'w') as f:
f.writelines(tuple(gjf_list))
return gjf_fpath
