def pyNCcomputeTestSet(cnstfile1,
saefile1,
nnfdir1,
dtdir,
dtdftpref,
dtpm6dir,
dtpm6pref,
N,
P=1.0):
# Construct pyNeuroChem classes
nc = pync.pyNeuroChem(cnstfile1, saefile1, nnfdir1, 0)
Eact = []
Ecmp = []
Eotr = []
Ndat = 0
Nmol = 0
t = 0.0
for i in range(0, N):
print('|----- File ' + str(i) + ' -----|')
print('Name: ' + dtdir + dtdftpref + str(i) + '_test.dat')
rv = bool(np.random.binomial(1, P))
if (os.path.isfile(dtdir + dtdftpref + str(i) + '_test.dat') and rv):
xyz, typ, Eact_t = gt.readncdat(dtdir + dtdftpref + str(i) +
'_test.dat')
xyz1, typ1, Eotr_t = gt.readncdat(dtpm6dir + dtpm6pref + str(i) +
'_test.dat')
if len(Eact_t) == len(Eotr_t):
Eact += shiftlsttomin(Eact_t)
Eotr += shiftlsttomin(Eotr_t)
#Eact += Eact_t
#Eotr += Eotr_t
Nmol += 1
Ndat += len(Eact_t)
# Set the conformers in NeuroChem
nc.setConformers(confs=xyz, types=typ)
# Compute Forces of Conformations
print('Computing energies 1...')
_t1b = tm.time()
Ecmp_t = nc.computeEnergies()
_t2b = (tm.time() - _t1b) * 1000.0
t += _t2b
print('Computation complete 1. Time: ' +
"{:.4f}".format(_t2b) + 'ms')
Ecmp += shiftlsttomin(Ecmp_t)
#Ecmp += Ecmp_t
Eact = np.array(Eact, dtype=float)
Eotr = np.array(Eotr, dtype=float)
Ecmp = np.array(Ecmp, dtype=float)
return Eact, Ecmp, Eotr, Ndat, Nmol, t
def pyNCcomputeTestSet(cnstfile1,saefile1,nnfdir1,dtdir,dtdftpref,dtpm6dir,dtpm6pref,N,P=1.0):
# Construct pyNeuroChem classes
nc = pync.pyNeuroChem(cnstfile1,saefile1,nnfdir1,0)
Eact = []
Ecmp = []
Eotr = []
Ndat = 0
Nmol = 0
t = 0.0
for i in range(0,N):
rv = bool(np.random.binomial(1,P))
if (os.path.isfile(dtdir + dtdftpref + str(i) + '_test.dat') and rv):
xyz,typ,Eact_t,chk = gt.readncdat(dtdir + dtdftpref + str(i) + '_test.dat')
xyz1,typ1,Eotr_t,chk = gt.readncdat(dtpm6dir + dtpm6pref + str(i) + '_test.dat')
if len(Eact_t) == len(Eotr_t):
#print ('|----- File ' + str(i) + ' -----|')
#print ('Name: ' + dtdir + dtdftpref + str(i) + '_test.dat')
Eact += shiftlsttomin( Eact_t )
Eotr += shiftlsttomin( Eotr_t )
#Eact += Eact_t
#Eotr += Eotr_t
Nmol += 1
Ndat += len( Eact_t )
# Set the conformers in NeuroChem
nc.setConformers(confs=xyz,types=typ)
# Compute Forces of Conformations
print(' ' + str(Nmol) + ') Computing ' + str(len( Eact_t )) + ' energies...')
_t1b = tm.time()
Ecmp_t = nc.computeEnergies()
_t2b = (tm.time() - _t1b) * 1000.0
t += _t2b
#print('Computation complete. Time: ' + "{:.4f}".format(_t2b) + 'ms')
Ecmp += shiftlsttomin( Ecmp_t )
#Ecmp += Ecmp_t99
else:
print (str(len(Eact_t)) + '!=' + str(len(Eotr_t)) + ' File: ' + dtdir + dtdftpref + str(i) + '_test.dat')
else:
print('File not found: ' + dtdir + dtdftpref + str(i) + '_test.dat')
Eact = np.array(Eact,dtype=float)
Eotr = np.array(Eotr,dtype=float)
Ecmp = np.array(Ecmp,dtype=float)
return Eact,Ecmp,Eotr,Ndat,Nmol,t
def read_data_files_convert_cm(file, N):
# Get training molecules
xyz_tr, typ_tr, Eact_tr, readf = gt.readncdat(file, np.float32)
# Compute energy of atoms at infinite separation
ise = cm.computerISE(typ_tr)
Eact_tr = (Eact_tr - ise)
Nm = Eact_tr.shape[0]
cmat = cm.GenCMatData2(xyz_tr, typ_tr, N)
return cmat, Eact_tr, ise
plt.rc('font', **font)
# Set required files for pyNeuroChem
wkdir = '/home/jujuman/Research/GDB-11-wB97X-6-31gd/train_08_5/'
# Network Files
cnstfile = wkdir + 'rHCNO-4.7A_32-3.2A_a8-8.params'
saefile = wkdir + 'sae_6-31gd.dat'
nnfdir = wkdir + 'networks/'
# Construct pyNeuroChem classes
nc = pync.pyNeuroChem(cnstfile, saefile, nnfdir, 0)
# Read nc DATA
xyz, typ, Eact = gt.readncdat(
'/home/jujuman/Dropbox/ChemSciencePaper.AER/TestCases/Retinol/data/retinolconformer_DFT.dat'
)
xyz1, typ1, Eact1 = gt.readncdat(
'/home/jujuman/Dropbox/ChemSciencePaper.AER/TestCases/Retinol/data/retinolconformer_DFTB.dat'
)
Eact = np.array(Eact)
Eact1 = np.array(Eact1)
# Set the conformers in NeuroChem
nc.setConformers(confs=xyz, types=typ)
# Print some data from the NeuroChem
print('Number of Atoms Loaded: ' + str(nc.getNumAtoms()))
print('Number of Confs Loaded: ' + str(nc.getNumConfs()))
def produce_scan(ax,title,xlabel,cnstfile,saefile,nnfdir,dtdir,dt1,dt2,dt3,smin,smax,iscale,ishift):
xyz, typ, Eact, chk = gt.readncdat(dtdir + dt1,np.float32)
xyz2, typ2, Eact2, chk = gt.readncdat(dtdir + dt2)
xyz3, typ3, Eact3, chk = gt.readncdat(dtdir + dt3)
#gt.writexyzfile("/home/jujuman/Dropbox/ChemSciencePaper.AER/TestCases/Dihedrals/4-Cyclohexyl-1-butanol/optimization/dihedral_"+dt1+".xyz",xyz,typ)
#Eact = np.array(Eact)
#Eact2 = np.array(Eact2)
#Eact3 = np.array(Eact3)
# Construct pyNeuroChem classes
nc1 = pync.pyNeuroChem(cnstfile, saefile, nnfdir, 1)
# Set the conformers in NeuroChem
nc1.setConformers(confs=xyz, types=typ)
# Print some data from the NeuroChem
print('1) Number of Atoms Loaded: ' + str(nc1.getNumAtoms()))
print('1) Number of Confs Loaded: ' + str(nc1.getNumConfs()))
# Compute Forces of Conformations
print('Computing energies 1...')
_t1b = tm.time()
Ecmp1 = nc1.energy()
print('Computation complete 1. Time: ' + "{:.4f}".format((tm.time() - _t1b) * 1000.0) + 'ms')
n = smin
m = smax
Ecmp1 = gt.hatokcal * Ecmp1
Eact = gt.hatokcal * Eact
Eact2 = gt.hatokcal * Eact2
Eact3 = gt.hatokcal * Eact3
IDX = np.arange(0, Eact.shape[0], 1, dtype=float) * iscale + ishift
IDX = IDX[n:m]
Eact = Eact[n:m]
Eact2 = Eact2[n:m]
Eact3 = Eact3[n:m]
Ecmp1 = Ecmp1[n:m]
Ecmp1 = Ecmp1 - Ecmp1.min()
Eact = Eact - Eact.min()
Eact2 = Eact2 - Eact2.min()
Eact3 = Eact3 - Eact3.min()
rmse1 = gt.calculaterootmeansqrerror(Eact, Ecmp1)
rmse3 = gt.calculaterootmeansqrerror(Eact, Eact2)
rmse4 = gt.calculaterootmeansqrerror(Eact, Eact3)
print("Spearman corr. 1: " + "{:.3f}".format(st.spearmanr(Ecmp1, Eact)[0]))
print("Spearman corr. 2: " + "{:.3f}".format(st.spearmanr(Eact2, Eact)[0]))
print("Spearman corr. 3: " + "{:.3f}".format(st.spearmanr(Eact3, Eact)[0]))
ax.plot(IDX, Eact, '-', marker=r'o', color='black', label='DFT',
linewidth=2, markersize=7)
ax.plot(IDX, Ecmp1, ':', marker=r'D', color='red', label='ANI-1 RMSE: ' + '%s' % float('%.3g' % rmse1) + ' kcal/mol',
linewidth=2, markersize=5)
ax.plot(IDX, Eact2, ':', marker=r'v', color='blue', label='DFTB RMSE: ' + '%s' % float('%.3g' % rmse3) + ' kcal/mol',
linewidth=2, markersize=5)
ax.plot(IDX, Eact3, ':', marker=r'*', color='orange', label='PM6 RMSE: ' + '%s' % float('%.3g' % rmse4) + ' kcal/mol',
linewidth=2, markersize=7)
#ax.plot(IDX, Eact, color='black', label='DFT', linewidth=3)
#ax.scatter(IDX, Eact, marker='o', color='black', linewidth=4)
th = ax.set_title(title,fontsize=16)
th.set_position([0.5,1.005])
# Set Limits
ax.set_xlim([ IDX.min(),IDX.max()])
ax.set_ylim([Eact.min()-1.0,Eact.max()+1.0])
ax.set_ylabel('$\Delta$E calculated (kcal/mol)')
ax.set_xlabel(xlabel)
ax.legend(bbox_to_anchor=(0.2, 0.98), loc=2, borderaxespad=0., fontsize=14)
import numpy as np
from scipy import stats as st
# Set required files for pyNeuroChem
wkdir = '/home/jujuman/Research/GDB-11-wB97X-6-31gd/train_07/'
#Network Parameter Files
cnstfile = wkdir + 'rHCNO-4.5A_32-3.1A_a8-8.params'
saefile = wkdir + 'sae_6-31gd.dat'
nnfdir = wkdir + 'networks/'
# Construct pyNeuroChem classes
nc = pync.pyNeuroChem(cnstfile,saefile,nnfdir,0)
xyz,typ,Eact = gt.readncdat('/home/jujuman/Dropbox/Research/ChemSciencePaper/TestCases/Atomoxetine/Atomoxetine_conformersC_test.dat')
Eact = np.array(Eact)
# Set the conformers in NeuroChem
nc.setConformers(confs=xyz,types=typ)
# Print some data from the NeuroChem
print( 'Number of Atoms Loaded: ' + str(nc.getNumAtoms()) )
print( 'Number of Confs Loaded: ' + str(nc.getNumConfs()) )
# Compute Forces of Conformations
print('Computing energies 1...')
Ecmp = np.array( nc.computeEnergies() )
print('Computation complete 1.')
from scipy import stats as st
# Set required files for pyNeuroChem
wkdir = '/home/jujuman/Research/GDB-11-wB97X-6-31gd/train_07/'
#Network Parameter Files
cnstfile = wkdir + 'rHCNO-4.5A_32-3.1A_a8-8.params'
saefile = wkdir + 'sae_6-31gd.dat'
nnfdir = wkdir + 'networks/'
# Construct pyNeuroChem classes
nc = pync.pyNeuroChem(cnstfile, saefile, nnfdir, 0)
xyz, typ, Eact = gt.readncdat(
'/home/jujuman/Dropbox/Research/ChemSciencePaper/TestCases/Atomoxetine/Atomoxetine_conformersC_test.dat'
)
Eact = np.array(Eact)
# Set the conformers in NeuroChem
nc.setConformers(confs=xyz, types=typ)
# Print some data from the NeuroChem
print('Number of Atoms Loaded: ' + str(nc.getNumAtoms()))
print('Number of Confs Loaded: ' + str(nc.getNumConfs()))
# Compute Forces of Conformations
print('Computing energies 1...')
Ecmp = np.array(nc.computeEnergies())
print('Computation complete 1.')
Eerr = []
time = 0.0
ld = [0,0.0]
sd = [0,100000.0]
err = []
sze = []
N = 0
for i in files:
#for i in range(rng[0],rng[1]):
#xyz,typ,Eact_t,readf = gt.readncdat(dtdir + fpref + str(i) + fpost)
xyz,typ,Eact_t,readf = gt.readncdat(dtdir + i)
if readf:
# Set the conformers in NeuroChem
nc.setConformers(confs=xyz,types=typ)
#print('FILE: ' + dtdir + fpref + str(i) + fpost)
print('FILE: ' + dtdir + i)
# Print some data from the NeuroChem
print( '1) Number of Atoms Loaded: ' + str(nc.getNumAtoms()) )
print( '1) Number of Confs Loaded: ' + str(nc.getNumConfs()) )
# Compute Energies of Conformations
print('Computing energies...')
_t1b = tm.time()
N = 30
print('Computing coulomb matrices...')
_t1b = tm.time()
cnt=0
for i in files:
cnt += 1
print('FILE: ' + str(cnt) + ' of ' + str(len(files)) + ' ' + i)
# Set file
file = dtdir + i
# Get training molecules
xyz_tr, typ_tr, Eact_tr, readf = gt.readncdat(file, np.float32)
# Compute energy of atoms at infinite separation
ise = cm.computerISE(typ_tr)
Eact_tr = (Eact_tr - ise)
Nm = Eact_tr.shape[0]
cmat = cm.GenCMatData(xyz_tr,typ_tr,N)
cmat[:, N * N ] = Eact_tr
output_file = open(cmdfile + '_' + str(cnt), 'wb')
cmat.tofile(output_file)
output_file.close()
plt.rc('font', **font)
# Set required files for pyNeuroChem
wkdir = '/home/jujuman/Research/GDB-11-wB97X-6-31gd/train_08_5/'
# Network Files
cnstfile = wkdir + 'rHCNO-4.7A_32-3.2A_a8-8.params'
saefile = wkdir + 'sae_6-31gd.dat'
nnfdir = wkdir + 'networks/'
# Construct pyNeuroChem classes
nc = pync.pyNeuroChem(cnstfile,saefile,nnfdir,0)
# Read nc DATA
xyz,typ,Eact = gt.readncdat('/home/jujuman/Dropbox/ChemSciencePaper.AER/TestCases/Retinol/data/retinolconformer_DFT.dat')
xyz1,typ1,Eact1 = gt.readncdat('/home/jujuman/Dropbox/ChemSciencePaper.AER/TestCases/Retinol/data/retinolconformer_DFTB.dat')
Eact = np.array(Eact)
Eact1 = np.array(Eact1)
# Set the conformers in NeuroChem
nc.setConformers(confs=xyz,types=typ)
# Print some data from the NeuroChem
print( 'Number of Atoms Loaded: ' + str(nc.getNumAtoms()) )
print( 'Number of Confs Loaded: ' + str(nc.getNumConfs()) )
# Compute Forces of Conformations
print('Computing energies...')
_t1b = tm.time()
sze = []
Herror = 0.0
Wfile = ''
Lerror = 100.0
Bfile = ''
Nf = len(files)
cnt = 0
for i in files:
cnt += 1
#for i in range(rng[0],rng[1]):
#xyz,typ,Eact_t,readf = gt.readncdat(dtdir + fpref + str(i) + fpost)
xyz,typ,Eact_W,readf = gt.readncdat(dtdir + i)
xyz = np.asarray(xyz,dtype=np.float32)
xyz = xyz.reshape((xyz.shape[0],len(typ),3))
if readf and xyz.shape[0] > 0:
#print('FILE: ' + dtdir + fpref + str(i) + fpost)
print('FILE: ' + str(cnt) + ' of ' + str(Nf) + ' ' + i)
Nm = xyz.shape[0]
Na = xyz.shape[1]
Nat = Na * Nm
Nit = int(np.ceil(Nat/65000.0))
for d in dtdirs:
tmp = listdir(d)
files = list({("_".join(f.split("_")[:-1])) for f in tmp})
allarr = []
Natoms = []
Nconfs = []
typarr = []
for i in files:
print(d+i)
nc = 0
for k in namelist:
try:
_timeloop = tm.time()
readarrays = gt.readncdat(d+i+k)
_timeloop2 = (tm.time() - _timeloop)
totaltime += _timeloop2
#print('Computation complete. Time: ' + "{:.4f}".format(_timeloop2) + 'ms')
shapesarr = [x.shape for x in readarrays]
typ = readarrays[1]
except FileNotFoundError:
readarrays = [np.zeros((0,*x[1:])) for x in shapesarr]
ncsub, nat, ndim = readarrays[0].shape
nc += ncsub
readarrays[0] = readarrays[0].reshape(ncsub*nat,ndim)
allarr.append(readarrays)
Natoms.append(nat)
#Network 1 Files
cnstfile1 = wkdir1 + 'rHCNO-4.6A_32-3.1A_a8-8.params'
saefile1 = wkdir1 + 'sae_6-31gd.dat'
nnfdir1 = wkdir1 + 'networks/'
# Network 2 Files
cnstfile2 = wkdir2 + 'rHCNO-4.6A_32-3.1A_a8-8.params'
saefile2 = wkdir2 + 'sae_6-31gd.dat'
nnfdir2 = wkdir2 + 'networks/'
# Construct pyNeuroChem classes
nc1 = pync.pyNeuroChem(cnstfile1, saefile1, nnfdir1, 0)
nc2 = pync.pyNeuroChem(cnstfile2, saefile2, nnfdir2, 0)
xyz, typ, Eact = gt.readncdat(
'/home/jujuman/Dropbox/Research/ChemSciencePaper/TestCases/C10H20Isomers/isomer_structures_DFT.dat'
)
Eact = np.array(Eact)
# Set the conformers in NeuroChem
nc1.setConformers(confs=xyz, types=typ)
nc2.setConformers(confs=xyz, types=typ)
# Print some data from the NeuroChem
print('Number of Atoms Loaded: ' + str(nc1.getNumAtoms()))
print('Number of Confs Loaded: ' + str(nc1.getNumConfs()))
# Compute Forces of Conformations
print('Computing energies 1...')
_t1b = tm.time()
xy = sorted(xy, key=lambda x: x[0])
X, Y = zip(*xy)
return np.array(Y)
# Set required files for pyNeuroChem
wkdir1 = '/home/jujuman/Research/GDB-11-wB97X-6-31gd/dataset_size_testing/train_08_1.00/'
#Network 1 Files
cnstfile1 = wkdir1 + 'rHCNO-4.5A_32-3.1A_a8-8.params'
saefile1 = wkdir1 + 'sae_6-31gd.dat'
nnfdir1 = wkdir1 + 'networks/'
dtdir = '/home/jujuman/Dropbox/ChemSciencePaper.AER/TestCases/C10H20Isomers/'
#xyz,typ,Eact = gt.readncdat('../data_irc.dat')
xyz,typ,Eact,tmp = gt.readncdat(dtdir + 'isomer_structures_DFT.dat')
#xyz2,typ2,Eact2,tmp = gt.readncdat(dtdir + 'isomer_structures_DFTB.dat')
#xyz3,typ3,Eact3,tmp = gt.readncdat(dtdir + 'isomer_structures_PM6.dat')
#xyz = [xyz[0],xyz[1]]
#xyz2 = [xyz2[0],xyz2[1]]
#xyz3 = [xyz3[0],xyz3[1]]
xyz = np.asarray(xyz,dtype=np.float32)
xyz = xyz.reshape((xyz.shape[0],len(typ),3))
Eact = np.array(Eact)
#Eact2 = np.array(Eact2)
#Eact3 = np.array(Eact3)
# Construct pyNeuroChem classes
print('Computing coulomb matrices...')
_t1b = tm.time()
cnt = 0
output_file = open(cmdfile + '.dat', 'wb')
for i in files:
cnt += 1
print('FILE: ' + str(cnt) + ' of ' + str(len(files)) + ' ' + i)
# Set file
file = dtdir + i
# Get training molecules
xyz_tr, typ_tr, Eact_tr, readf = gt.readncdat(file, np.float32)
xyz_tr = xyz_tr[0:int(P * xyz_tr.shape[0])]
Eact_tr = Eact_tr[0:int(P * Eact_tr.shape[0])]
# Compute energy of atoms at infinite separation
ise = cm.computerISE(typ_tr)
Eact_tr = (Eact_tr - ise)
Nm = Eact_tr.shape[0]
cmat = cm.GenCMatData(xyz_tr, typ_tr, N)
cmat[:, N * N] = Eact_tr
cmat.tofile(output_file)
