def get_atomization_energies(self, prog, meth):
""" get AE """
udct = {'h': io2.Units().h2kc}
const = 1.0
if meth in ['a', 'all']:
meths = list(self.props.keys())
else:
meths = [meth]
aes = {}
for mt in meths:
_es0 = np.zeros(1 + self.zsu[-1])
for zi in self.zsu:
si = chemical_symbols[zi]
#try:
ei0 = cda.dct[prog][si][mt]
#print('ei0=',ei0)
#except:
# ei0 = cda.dct[prog][mt][si]
_es0[zi] = ei0 * udct['h']
_aes = []
for i, na in enumerate(self.nas):
ib, ie = self.ias1[i], self.ias2[i]
zsi = self.zs[ib:ie]
ei0 = np.sum([_es0[zsi]])
#print('props=', self.props)
_aes.append(self.props[mt][i] * const - ei0)
aes[mt] = np.array(_aes)
#self.props[mt] = np.array(aes)
self.aes = aes
def __init__(self, fs, pns=None, unit='h'):
nas = []
zs = []
coords = []
nsheav = []
props = {}
self.nm = len(fs)
const = 1.
if unit in ['h']: const = io2.Units().h2kc
for f in fs:
_nas, _zs, _coords, _nsheav, _props = read_xyz_simple(
f, opt='z', property_names=pns)
#print('f,props=',f, _props)
nas += list(_nas)
zs += list(_zs)
coords += list(_coords)
nsheav += list(_nsheav)
for k in _props:
if k in props:
props[k] += [_props[k] * const]
else:
props[k] = [_props[k] * const]
self.nas = nas
self.zs = zs
self.coords = coords
self.nsheav = nsheav
self.props = props
self.ias2 = np.cumsum(nas)
self.ias1 = np.concatenate(([0], self.ias2[:-1]))
def __init__(self, es, basis, unit='h'): #, etype='corr'):
"""
Note: input energy (`es) has to be either HF energies or other type of energies
"""
self.es = es
self.basis = basis
uc = io2.Units()
self.const = {'kcal':uc.h2kc, 'h':1.0}[unit]
def calc_ae_dressed(self, idx1, idx2=None, meth=None, ref=None):
"""
calc dressed atom or/and bond energies as baseline
vars
====================
ref: if set to None, then E_atom will be regressed by input mols
otherwise, E_atom will be regressed from a set of small mols
located in folder `ref/ (i.e., h2,ch4,n2,o2,f2,s8,p4,cl2,br2,i2)
"""
ims = np.arange(self.nm)
if isinstance(idx1, int):
idx1 = ims[:idx1]
if idx2:
if isinstance(idx2, int):
idx2 = ims[idx2:]
else:
idx2 = np.setdiff1d(ims, idx1)
if meth is None:
meths = list(self.props.keys())
if len(meths) > 1:
raise Exception(
'#ERROR: multiple property avail. Please specify one!')
meth = meths[0]
ns1, ns2 = self.ngs[idx1], self.ngs[idx2]
nel = len(ns1[0])
uc = io2.Units()
const = 1.0 #{'h': uc.h2kc, 'kcal':1.0 }[iu]
if hasattr(self, 'ys'):
ys = self.ys
else:
ys = self.props[meth]
ys *= const
ys1, ys2 = ys[idx1], ys[idx2]
istat = T
if ref is None: # default case
esb, _, rank, _ = np.linalg.lstsq(ns1, ys1, rcond=None)
if rank < nel:
print(' ** number of molecules .le. number of elements')
esb = np.zeros(nel)
istat = F
else:
esb = get_reference_atomic_energy(ref, meth, self.zsu)
ys1p = np.dot(ns1, esb)
#print ' +++++ ys1.shape, ys1p.shape = ', ys1.shape, ys1p.shape
dys1 = ys1 - ys1p
ys2_base = np.dot(ns2, esb)
dys2 = ys2 - ys2_base
return istat, dys1, dys2, ys2_base
def get_force(self):
iou = io2.Units()
const = iou.h2e / iou.b2a # from hartree/bohr to eV/A
#cmd = "grep '^\s*[XYZ][0-9]* ' %s | awk '{print $3}'"%self.f #
cmd1 = "awk '/^ Variable Old X -DE/{print NR}' %s | tail -1" % self.f
Ln1 = int(io2.cmdout2(cmd1)) + 2
cmd2 = "awk '/^ Item Value Threshold Converged/{print NR}' %s | tail -1" % self.f
Ln2 = int(io2.cmdout2(cmd2)) - 1
cmd = "sed -n '%d,%dp' %s" % (Ln1, Ln2, self.f)
#print cmd
cs = io2.cmdout2(cmd).split('\n')
vals = [eval(ci.split()[2]) for ci in cs]
#print vals
#print len(vals)
forces = np.array(vals).reshape((self.na, 3)) * const
#abs_forces = np.linalg.norm( forces, axis=1 )
#self.forces = abs_forces[:, np.newaxis]
return forces
def obj2m(obj, property_names=None, isimple=F, idx=None, unit=None):
assert unit is not None
if property_names:
if ('nmr' in property_names) or ('chgs' in property_names):
isimple = T
if isinstance(obj, str):
if (not os.path.isfile(obj)):
raise Exception('#ERROR: file does not exist')
if (obj[-3:] == 'xyz'):
func = read_xyz_simple if (isimple and (idx is None)) else read_xyz
nas, zs, coords, nsheav, props = func(
obj, property_names=property_names, idx=idx)
elif (obj[-3:] == 'sdf'):
property_names = []
_mi = read_sdf(obj, iconn=F)
nas = [len(_mi)]
zs = _mi.numbers
coords = _mi.positions
nsheav = [(zs > 1).sum()]
props = {}
else:
raise Exception('#ERROR: non-xyz file')
elif isinstance(obj, list):
nas = []
zs = []
coords = []
nsheav = []
props = {}
for _obj in obj:
#print('type(obj)=',type(obj))
if (_obj[-3:] == 'xyz'):
func = read_xyz_simple if (isimple and
(idx is None)) else read_xyz
_nas, _zs, _coords, _nsheav, _props = func(
_obj, property_names=property_names, idx=idx)
elif (_obj[-3:] == 'sdf'):
property_names = []
_mi = read_sdf(_obj, iconn=F)
_nas = [len(_mi)]
_zs = _mi.numbers
_coords = _mi.positions
_nsheav = [(_zs > 1).sum()]
_props = {}
else:
raise Exception('#ERROR: not sdf/xyz file')
zs += list(_zs)
coords += list(_coords)
nsheav += list(_nsheav)
nas += list(_nas)
#print('props=', _props)
property_names = list(_props.keys())
for p in property_names:
yi = _props[p]
if p in props.keys():
props[p] += [yi]
else:
props[p] = [yi]
#elif 'molecules' in obj.__str__():
# nas, zs, coords, nsheav, props = obj._nas, obj._zs, obj._coords, obj._nsheav, obj.props
elif 'catoms' in obj.__str__():
nas, zs, coords, nsheav, props = obj.nas, obj.zs, obj.coords, obj.nsheav, obj.props
else:
print('obj=', obj)
raise Exception('#ERROR: unknown input `obj!')
const = 1.0
for p in props:
if unit in ['h', 'ha']:
const = io2.Units().h2kc
v = props[p]
if p in [
'chg', 'chgs', 'nmr', 'cls', 'force', 'forces', 'grad',
'grads', 'alpha'
]:
const = 1.0
props[p] = np.array(v) * const
return nas, zs, coords, nsheav, props
#!/usr/bin/env python
import re, os, sys
import aqml.io2 as io2
import aqml.cheminfo as co
import aqml.cheminfo.core as cc
import numpy as np
T, F = True, False
uc = io2.Units()
cardinal = {'vdz': 2, 'vtz': 3, 'vqz': 4}
bsts = ['aug-cc-pvdz', 'aug-cc-pvtz', 'aug-cc-pvqz', \
'cc-pvdz', 'cc-pvtz', 'cc-pvqz', \
'def2-sv(p)', 'def2-svp', 'def2-tzvp', 'def2-qzvp']
bsts_short = ['avdz', 'avtz', 'avqz', \
'vdz', 'vtz','vqz', \
'def2sv-p', 'def2svp', 'def2tzvp', 'def2qzvp']
dctbo = dict(zip(bsts, bsts_short))
dctbi = dict(zip(bsts_short, bsts))
hs_short = ['mp2', 'lmp2', 'lcc', 'lcc2', 'cc', 'cc2'] # Hamitonian
hs = ['RI-MP2', 'DLPNO-MP2', 'DLPNO-CCSD', 'CCSD', 'CCSD(T)']
dcthi = dict(zip(hs_short, hs))
dctho = dict(zip(hs, hs_short))
cmdout = lambda cmd: os.popen(cmd).read().strip()
cmdout1 = lambda cmd: os.popen(cmd).read().strip().split('\n')
