3
0
mirror of https://github.com/triqs/dft_tools synced 2024-11-09 07:33:47 +01:00
dft_tools/python/converters/vasp_converter.py
2015-10-13 11:36:43 +02:00

787 lines
37 KiB
Python

################################################################################
#
# TRIQS: a Toolbox for Research in Interacting Quantum Systems
#
# Copyright (C) 2011 by M. Aichhorn, L. Pourovskii, V. Vildosola
#
# TRIQS is free software: you can redistribute it and/or modify it under the
# terms of the GNU General Public License as published by the Free Software
# Foundation, either version 3 of the License, or (at your option) any later
# version.
#
# TRIQS is distributed in the hope that it will be useful, but WITHOUT ANY
# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
# details.
#
# You should have received a copy of the GNU General Public License along with
# TRIQS. If not, see <http://www.gnu.org/licenses/>.
#
################################################################################
from types import *
import numpy
from pytriqs.archive import *
from pytriqs.applications.dft.converters.converter_tools import *
import os.path
try:
import simplejson as json
except ImportError:
import json
#from plotools import ProjectorGroup, ProjectorShell
class VaspConverter(ConverterTools):
"""
Conversion from VASP output to an hdf5 file that can be used as input for the SumkDFT class.
"""
def __init__(self, filename, hdf_filename = None,
dft_subgrp = 'dft_input', symmcorr_subgrp = 'dft_symmcorr_input',
parproj_subgrp='dft_parproj_input', symmpar_subgrp='dft_symmpar_input',
bands_subgrp = 'dft_bands_input', misc_subgrp = 'dft_misc_input',
transp_subgrp = 'dft_transp_input', repacking = False):
"""
Init of the class. Variable filename gives the root of all filenames, e.g. case.ctqmcout, case.h5, and so on.
"""
assert type(filename)==StringType, "Please provide the DFT files' base name as a string."
if hdf_filename is None: hdf_filename = filename+'.h5'
self.hdf_file = hdf_filename
self.basename = filename
self.ctrl_file = filename+'.ctrl'
# self.pmat_file = filename+'.pmat'
self.dft_subgrp = dft_subgrp
self.symmcorr_subgrp = symmcorr_subgrp
self.parproj_subgrp = parproj_subgrp
self.symmpar_subgrp = symmpar_subgrp
self.bands_subgrp = bands_subgrp
self.misc_subgrp = misc_subgrp
self.transp_subgrp = transp_subgrp
# Checks if h5 file is there and repacks it if wanted:
if (os.path.exists(self.hdf_file) and repacking):
ConverterTools.repack(self)
def read_data(self, fh):
"""
Generator for reading plain data.
"""
for line in fh:
line_ = line.strip()
if not line or (line_ == '' or line_[0] == '#'):
continue
for val in map(float, line.split()):
yield val
def read_header_and_data(self, filename):
"""
Opens a file and returns a JSON-header and the generator for the plain data.
"""
fh = open(filename, 'rt')
header = ""
for line in fh:
if not "#END" in line:
header += line
else:
break
f_gen = self.read_data(fh)
return header, f_gen
def convert_dft_input(self):
"""
Reads the input files, and stores the data in the HDFfile
"""
energy_unit = 1.0 # VASP interface always uses eV
k_dep_projection = 1
# Symmetries are switched off for the moment
# TODO: implement symmetries
symm_op = 0 # Use symmetry groups for the k-sum
# Read and write only on the master node
if not (mpi.is_master_node()): return
mpi.report("Reading input from %s..."%self.ctrl_file)
# R is a generator : each R.Next() will return the next number in the file
jheader, rf = self.read_header_and_data(self.ctrl_file)
print jheader
ctrl_head = json.loads(jheader)
ng = ctrl_head['ngroups']
n_k = ctrl_head['nk']
# Note the difference in name conventions!
SP = ctrl_head['ns'] - 1
SO = ctrl_head['nc_flag']
kpts = numpy.zeros((n_k, 3))
bz_weights = numpy.zeros(n_k)
try:
for ik in xrange(n_k):
kx, ky, kz = rf.next(), rf.next(), rf.next()
kpts[ik, :] = kx, ky, kz
bz_weights[ik] = rf.next()
except StopIteration:
raise "VaspConverter: error reading %s"%self.ctrl_file
# if nc_flag:
## TODO: check this
# n_spin_blocs = 1
# else:
# n_spin_blocs = ns
n_spin_blocs = SP + 1 - SO
# Read PLO groups
# First, we read everything into a temporary data structure
# TODO: think about multiple shell groups and how to map them on h5 structures
assert ng == 1, "Only one group is allowed at the moment"
try:
for ig in xrange(ng):
gr_file = self.basename + '.pg%i'%(ig + 1)
jheader, rf = self.read_header_and_data(gr_file)
gr_head = json.loads(jheader)
e_win = gr_head['ewindow']
nb_max = gr_head['nb_max']
p_shells = gr_head['shells']
density_required = gr_head['nelect']
charge_below = 0.0 # This is not defined in VASP interface
# Note that in the DftTools convention each site gives a separate correlated shell!
n_corr_shells = sum([len(sh['ion_list']) for sh in p_shells])
corr_shells = []
shion_to_corr_shell = [[] for ish in xrange(len(p_shells))]
icsh = 0
for ish, sh in enumerate(p_shells):
ion_list = sh['ion_list']
for i, ion in enumerate(ion_list):
pars = {}
pars['atom'] = ion
pars['sort'] = sh['ion_sort']
pars['l'] = sh['lorb']
pars['dim'] = sh['ndim']
pars['SO'] = SO
# TODO: check what 'irep' entry does (it seems to be very specific to dmftproj)
pars['irep'] = 0
corr_shells.append(pars)
shion_to_corr_shell[ish].append(i)
# TODO: generalize this to the case of multiple shell groups
n_shells = n_corr_shells # No non-correlated shells at the moment
shells = corr_shells
# FIXME: atomic sorts in Wien2K are not the same as in VASP.
# A symmetry analysis from OUTCAR or symmetry file should be used
# to define equivalence classes of sites.
n_inequiv_shells, corr_to_inequiv, inequiv_to_corr = ConverterTools.det_shell_equivalence(self, corr_shells)
# NB!: these rotation matrices are specific to Wien2K! Set to identity in VASP
use_rotations = 1
rot_mat = [numpy.identity(corr_shells[icrsh]['dim'],numpy.complex_) for icrsh in range(n_corr_shells)]
rot_mat_time_inv = [0 for i in range(n_corr_shells)]
# TODO: implement transformation matrices
n_reps = [1 for i in range(n_inequiv_shells)]
dim_reps = [0 for i in range(n_inequiv_shells)]
T = []
for ish in range(n_inequiv_shells):
n_reps[ish] = 1 # Always 1 in VASP
ineq_first = inequiv_to_corr[ish]
dim_reps[ish] = [corr_shells[ineq_first]['dim']] # Just the dimension of the shell
# The transformation matrix:
# is of dimension 2l+1 without SO, and 2*(2l+1) with SO!
ll = 2 * corr_shells[inequiv_to_corr[ish]]['l']+1
lmax = ll * (corr_shells[inequiv_to_corr[ish]]['SO'] + 1)
# TODO: at the moment put T-matrices to identities
T.append(numpy.identity(lmax, numpy.complex_))
# if nc_flag:
## TODO: implement the noncollinear part
# raise NotImplementedError("Noncollinear calculations are not implemented")
# else:
hopping = numpy.zeros([n_k, n_spin_blocs, nb_max, nb_max], numpy.complex_)
band_window = [numpy.zeros((n_k, 2), dtype=int) for isp in xrange(n_spin_blocs)]
n_orbitals = numpy.zeros([n_k, n_spin_blocs], numpy.int)
for isp in xrange(n_spin_blocs):
for ik in xrange(n_k):
ib1, ib2 = int(rf.next()), int(rf.next())
band_window[isp][ik, :2] = ib1, ib2
nb = ib2 - ib1 + 1
n_orbitals[ik, isp] = nb
for ib in xrange(nb):
hopping[ik, isp, ib, ib] = rf.next()
# Projectors
print n_orbitals
print [crsh['dim'] for crsh in corr_shells]
proj_mat = numpy.zeros([n_k, n_spin_blocs, n_corr_shells, max([crsh['dim'] for crsh in corr_shells]), max(n_orbitals)], numpy.complex_)
# TODO: implement reading from more than one projector group
# In 'dmftproj' each ion represents a separate correlated shell.
# In my interface a 'projected shell' includes sets of ions.
# How to reconcile this? Two options:
#
# 1. Redefine 'projected shell' in my interface to make it correspond to one site only.
# In this case the list of ions must be defined at the level of the projector group.
#
# 2. Split my 'projected shell' to several 'correlated shells' here in the converter.
#
# At the moment I choose i.2 for its simplicity. But one should consider possible
# use cases and decide which solution is to be made permanent.
#
for ish, sh in enumerate(p_shells):
for isp in xrange(n_spin_blocs):
for ik in xrange(n_k):
for ion in xrange(len(sh['ion_list'])):
icsh = shion_to_corr_shell[ish][ion]
for ilm in xrange(sh['ndim']):
for ib in xrange(n_orbitals[ik, isp]):
# This is to avoid confusion with the order of arguments
pr = rf.next()
pi = rf.next()
proj_mat[ik, isp, icsh, ilm, ib] = complex(pr, pi)
things_to_set = ['n_shells','shells','n_corr_shells','corr_shells','n_spin_blocs','n_orbitals','n_k','SO','SP','energy_unit']
for it in things_to_set:
print "%s:"%(it), locals()[it]
setattr(self,it,locals()[it])
except StopIteration:
raise "VaspConverter: error reading %s"%self.gr_file
rf.close()
#
# try:
# energy_unit = R.next() # read the energy convertion factor
# n_k = int(R.next()) # read the number of k points
# k_dep_projection = 1
# SP = int(R.next()) # flag for spin-polarised calculation
# SO = int(R.next()) # flag for spin-orbit calculation
# charge_below = R.next() # total charge below energy window
# density_required = R.next() # total density required, for setting the chemical potential
# symm_op = 1 # Use symmetry groups for the k-sum
#
# # the information on the non-correlated shells is not important here, maybe skip:
# n_shells = int(R.next()) # number of shells (e.g. Fe d, As p, O p) in the unit cell,
# # corresponds to index R in formulas
# # now read the information about the shells (atom, sort, l, dim):
# shell_entries = ['atom', 'sort', 'l', 'dim']
# shells = [ {name: int(val) for name, val in zip(shell_entries, R)} for ish in range(n_shells) ]
#
# n_corr_shells = int(R.next()) # number of corr. shells (e.g. Fe d, Ce f) in the unit cell,
# # corresponds to index R in formulas
# # now read the information about the shells (atom, sort, l, dim, SO flag, irep):
# corr_shell_entries = ['atom', 'sort', 'l', 'dim', 'SO', 'irep']
# corr_shells = [ {name: int(val) for name, val in zip(corr_shell_entries, R)} for icrsh in range(n_corr_shells) ]
#
# # determine the number of inequivalent correlated shells and maps, needed for further reading
# n_inequiv_shells, corr_to_inequiv, inequiv_to_corr = ConverterTools.det_shell_equivalence(self,corr_shells)
#
# use_rotations = 1
# rot_mat = [numpy.identity(corr_shells[icrsh]['dim'],numpy.complex_) for icrsh in range(n_corr_shells)]
#
# # read the matrices
# rot_mat_time_inv = [0 for i in range(n_corr_shells)]
#
# for icrsh in range(n_corr_shells):
# for i in range(corr_shells[icrsh]['dim']): # read real part:
# for j in range(corr_shells[icrsh]['dim']):
# rot_mat[icrsh][i,j] = R.next()
# for i in range(corr_shells[icrsh]['dim']): # read imaginary part:
# for j in range(corr_shells[icrsh]['dim']):
# rot_mat[icrsh][i,j] += 1j * R.next()
#
# if (SP==1): # read time inversion flag:
# rot_mat_time_inv[icrsh] = int(R.next())
#
# # Read here the info for the transformation of the basis:
# n_reps = [1 for i in range(n_inequiv_shells)]
# dim_reps = [0 for i in range(n_inequiv_shells)]
# T = []
# for ish in range(n_inequiv_shells):
# n_reps[ish] = int(R.next()) # number of representatives ("subsets"), e.g. t2g and eg
# dim_reps[ish] = [int(R.next()) for i in range(n_reps[ish])] # dimensions of the subsets
#
# # The transformation matrix:
# # is of dimension 2l+1 without SO, and 2*(2l+1) with SO!
# ll = 2*corr_shells[inequiv_to_corr[ish]]['l']+1
# lmax = ll * (corr_shells[inequiv_to_corr[ish]]['SO'] + 1)
# T.append(numpy.zeros([lmax,lmax],numpy.complex_))
#
# # now read it from file:
# for i in range(lmax):
# for j in range(lmax):
# T[ish][i,j] = R.next()
# for i in range(lmax):
# for j in range(lmax):
# T[ish][i,j] += 1j * R.next()
#
# # Spin blocks to be read:
# n_spin_blocs = SP + 1 - SO
#
# # read the list of n_orbitals for all k points
# n_orbitals = numpy.zeros([n_k,n_spin_blocs],numpy.int)
# for isp in range(n_spin_blocs):
# for ik in range(n_k):
# n_orbitals[ik,isp] = int(R.next())
#
# # Initialise the projectors:
# proj_mat = numpy.zeros([n_k,n_spin_blocs,n_corr_shells,max([crsh['dim'] for crsh in corr_shells]),max(n_orbitals)],numpy.complex_)
#
# # Read the projectors from the file:
# for ik in range(n_k):
# for icrsh in range(n_corr_shells):
# n_orb = corr_shells[icrsh]['dim']
# # first Real part for BOTH spins, due to conventions in dmftproj:
# for isp in range(n_spin_blocs):
# for i in range(n_orb):
# for j in range(n_orbitals[ik][isp]):
# proj_mat[ik,isp,icrsh,i,j] = R.next()
# # now Imag part:
# for isp in range(n_spin_blocs):
# for i in range(n_orb):
# for j in range(n_orbitals[ik][isp]):
# proj_mat[ik,isp,icrsh,i,j] += 1j * R.next()
#
# # now define the arrays for weights and hopping ...
# bz_weights = numpy.ones([n_k],numpy.float_)/ float(n_k) # w(k_index), default normalisation
# hopping = numpy.zeros([n_k,n_spin_blocs,max(n_orbitals),max(n_orbitals)],numpy.complex_)
#
# # weights in the file
# for ik in range(n_k) : bz_weights[ik] = R.next()
#
# # if the sum over spins is in the weights, take it out again!!
# sm = sum(bz_weights)
# bz_weights[:] /= sm
#
# # Grab the H
# # we use now the convention of a DIAGONAL Hamiltonian -- convention for Wien2K.
# for isp in range(n_spin_blocs):
# for ik in range(n_k) :
# n_orb = n_orbitals[ik,isp]
# for i in range(n_orb):
# hopping[ik,isp,i,i] = R.next() * energy_unit
#
# # keep some things that we need for reading parproj:
# things_to_set = ['n_shells','shells','n_corr_shells','corr_shells','n_spin_blocs','n_orbitals','n_k','SO','SP','energy_unit']
# for it in things_to_set: setattr(self,it,locals()[it])
# except StopIteration : # a more explicit error if the file is corrupted.
# raise "Wien2k_converter : reading file %s failed!"%filename
#
# R.close()
# # Reading done!
# Save it to the HDF:
ar = HDFArchive(self.hdf_file,'a')
if not (self.dft_subgrp in ar): ar.create_group(self.dft_subgrp)
# The subgroup containing the data. If it does not exist, it is created. If it exists, the data is overwritten!
things_to_save = ['energy_unit','n_k','k_dep_projection','SP','SO','charge_below','density_required',
'symm_op','n_shells','shells','n_corr_shells','corr_shells','use_rotations','rot_mat',
'rot_mat_time_inv','n_reps','dim_reps','T','n_orbitals','proj_mat','bz_weights','hopping',
'n_inequiv_shells', 'corr_to_inequiv', 'inequiv_to_corr']
for it in things_to_save: ar[self.dft_subgrp][it] = locals()[it]
del ar
# Symmetries are used, so now convert symmetry information for *correlated* orbitals:
self.convert_symmetry_input(ctrl_head, orbits=self.corr_shells, symm_subgrp=self.symmcorr_subgrp)
# TODO: Implement misc_input
# self.convert_misc_input(bandwin_file=self.bandwin_file,struct_file=self.struct_file,outputs_file=self.outputs_file,
# misc_subgrp=self.misc_subgrp,SO=self.SO,SP=self.SP,n_k=self.n_k)
def convert_parproj_input(self):
"""
Reads the input for the partial charges projectors from case.parproj, and stores it in the symmpar_subgrp
group in the HDF5.
"""
if not (mpi.is_master_node()): return
mpi.report("Reading input from %s..."%self.parproj_file)
dens_mat_below = [ [numpy.zeros([self.shells[ish]['dim'],self.shells[ish]['dim']],numpy.complex_) for ish in range(self.n_shells)]
for isp in range(self.n_spin_blocs) ]
R = ConverterTools.read_fortran_file(self,self.parproj_file,self.fortran_to_replace)
n_parproj = [int(R.next()) for i in range(self.n_shells)]
n_parproj = numpy.array(n_parproj)
# Initialise P, here a double list of matrices:
proj_mat_all = numpy.zeros([self.n_k,self.n_spin_blocs,self.n_shells,max(n_parproj),max([sh['dim'] for sh in self.shells]),max(self.n_orbitals)],numpy.complex_)
rot_mat_all = [numpy.identity(self.shells[ish]['dim'],numpy.complex_) for ish in range(self.n_shells)]
rot_mat_all_time_inv = [0 for i in range(self.n_shells)]
for ish in range(self.n_shells):
# read first the projectors for this orbital:
for ik in range(self.n_k):
for ir in range(n_parproj[ish]):
for isp in range(self.n_spin_blocs):
for i in range(self.shells[ish]['dim']): # read real part:
for j in range(self.n_orbitals[ik][isp]):
proj_mat_all[ik,isp,ish,ir,i,j] = R.next()
for isp in range(self.n_spin_blocs):
for i in range(self.shells[ish]['dim']): # read imaginary part:
for j in range(self.n_orbitals[ik][isp]):
proj_mat_all[ik,isp,ish,ir,i,j] += 1j * R.next()
# now read the Density Matrix for this orbital below the energy window:
for isp in range(self.n_spin_blocs):
for i in range(self.shells[ish]['dim']): # read real part:
for j in range(self.shells[ish]['dim']):
dens_mat_below[isp][ish][i,j] = R.next()
for isp in range(self.n_spin_blocs):
for i in range(self.shells[ish]['dim']): # read imaginary part:
for j in range(self.shells[ish]['dim']):
dens_mat_below[isp][ish][i,j] += 1j * R.next()
if (self.SP==0): dens_mat_below[isp][ish] /= 2.0
# Global -> local rotation matrix for this shell:
for i in range(self.shells[ish]['dim']): # read real part:
for j in range(self.shells[ish]['dim']):
rot_mat_all[ish][i,j] = R.next()
for i in range(self.shells[ish]['dim']): # read imaginary part:
for j in range(self.shells[ish]['dim']):
rot_mat_all[ish][i,j] += 1j * R.next()
if (self.SP):
rot_mat_all_time_inv[ish] = int(R.next())
R.close()
# Reading done!
# Save it to the HDF:
ar = HDFArchive(self.hdf_file,'a')
if not (self.parproj_subgrp in ar): ar.create_group(self.parproj_subgrp)
# The subgroup containing the data. If it does not exist, it is created. If it exists, the data is overwritten!
things_to_save = ['dens_mat_below','n_parproj','proj_mat_all','rot_mat_all','rot_mat_all_time_inv']
for it in things_to_save: ar[self.parproj_subgrp][it] = locals()[it]
del ar
# Symmetries are used, so now convert symmetry information for *all* orbitals:
self.convert_symmetry_input(orbits=self.shells,symm_file=self.symmpar_file,symm_subgrp=self.symmpar_subgrp,SO=self.SO,SP=self.SP)
def convert_bands_input(self):
"""
Converts the input for momentum resolved spectral functions, and stores it in bands_subgrp in the
HDF5.
"""
if not (mpi.is_master_node()): return
mpi.report("Reading bands input from %s..."%self.band_file)
R = ConverterTools.read_fortran_file(self,self.band_file,self.fortran_to_replace)
try:
n_k = int(R.next())
# read the list of n_orbitals for all k points
n_orbitals = numpy.zeros([n_k,self.n_spin_blocs],numpy.int)
for isp in range(self.n_spin_blocs):
for ik in range(n_k):
n_orbitals[ik,isp] = int(R.next())
# Initialise the projectors:
proj_mat = numpy.zeros([n_k,self.n_spin_blocs,self.n_corr_shells,max([crsh['dim'] for crsh in self.corr_shells]),max(n_orbitals)],numpy.complex_)
# Read the projectors from the file:
for ik in range(n_k):
for icrsh in range(self.n_corr_shells):
n_orb = self.corr_shells[icrsh]['dim']
# first Real part for BOTH spins, due to conventions in dmftproj:
for isp in range(self.n_spin_blocs):
for i in range(n_orb):
for j in range(n_orbitals[ik,isp]):
proj_mat[ik,isp,icrsh,i,j] = R.next()
# now Imag part:
for isp in range(self.n_spin_blocs):
for i in range(n_orb):
for j in range(n_orbitals[ik,isp]):
proj_mat[ik,isp,icrsh,i,j] += 1j * R.next()
hopping = numpy.zeros([n_k,self.n_spin_blocs,max(n_orbitals),max(n_orbitals)],numpy.complex_)
# Grab the H
# we use now the convention of a DIAGONAL Hamiltonian!!!!
for isp in range(self.n_spin_blocs):
for ik in range(n_k) :
n_orb = n_orbitals[ik,isp]
for i in range(n_orb):
hopping[ik,isp,i,i] = R.next() * self.energy_unit
# now read the partial projectors:
n_parproj = [int(R.next()) for i in range(self.n_shells)]
n_parproj = numpy.array(n_parproj)
# Initialise P, here a double list of matrices:
proj_mat_all = numpy.zeros([n_k,self.n_spin_blocs,self.n_shells,max(n_parproj),max([sh['dim'] for sh in self.shells]),max(n_orbitals)],numpy.complex_)
for ish in range(self.n_shells):
for ik in range(n_k):
for ir in range(n_parproj[ish]):
for isp in range(self.n_spin_blocs):
for i in range(self.shells[ish]['dim']): # read real part:
for j in range(n_orbitals[ik,isp]):
proj_mat_all[ik,isp,ish,ir,i,j] = R.next()
for i in range(self.shells[ish]['dim']): # read imaginary part:
for j in range(n_orbitals[ik,isp]):
proj_mat_all[ik,isp,ish,ir,i,j] += 1j * R.next()
except StopIteration : # a more explicit error if the file is corrupted.
raise "Wien2k_converter : reading file band_file failed!"
R.close()
# Reading done!
# Save it to the HDF:
ar = HDFArchive(self.hdf_file,'a')
if not (self.bands_subgrp in ar): ar.create_group(self.bands_subgrp)
# The subgroup containing the data. If it does not exist, it is created. If it exists, the data is overwritten!
things_to_save = ['n_k','n_orbitals','proj_mat','hopping','n_parproj','proj_mat_all']
for it in things_to_save: ar[self.bands_subgrp][it] = locals()[it]
del ar
def convert_misc_input(self, bandwin_file, struct_file, outputs_file, misc_subgrp, SO, SP, n_k):
"""
Reads input for the band window from bandwin_file, which is case.oubwin,
structure from struct_file, which is case.struct,
symmetries from outputs_file, which is case.outputs.
"""
if not (mpi.is_master_node()): return
things_to_save = []
# Read relevant data from .oubwin/up/dn files
#############################################
# band_window: Contains the index of the lowest and highest band within the
# projected subspace (used by dmftproj) for each k-point.
if (SP == 0 or SO == 1):
files = [self.bandwin_file]
elif SP == 1:
files = [self.bandwin_file+'up', self.bandwin_file+'dn']
else: # SO and SP can't both be 1
assert 0, "convert_transport_input: Reding oubwin error! Check SP and SO!"
band_window = [numpy.zeros((n_k, 2), dtype=int) for isp in range(SP + 1 - SO)]
for isp, f in enumerate(files):
if os.path.exists(f):
mpi.report("Reading input from %s..."%f)
R = ConverterTools.read_fortran_file(self, f, self.fortran_to_replace)
assert int(R.next()) == n_k, "convert_misc_input: Number of k-points is inconsistent in oubwin file!"
assert int(R.next()) == SO, "convert_misc_input: SO is inconsistent in oubwin file!"
for ik in xrange(n_k):
R.next()
band_window[isp][ik,0] = R.next() # lowest band
band_window[isp][ik,1] = R.next() # highest band
R.next()
things_to_save.append('band_window')
R.close() # Reading done!
# Read relevant data from .struct file
######################################
# lattice_type: bravais lattice type as defined by Wien2k
# lattice_constants: unit cell parameters in a. u.
# lattice_angles: unit cell angles in rad
if (os.path.exists(self.struct_file)):
mpi.report("Reading input from %s..."%self.struct_file)
with open(self.struct_file) as R:
try:
R.readline()
lattice_type = R.readline().split()[0]
R.readline()
temp = R.readline()
print temp
lattice_constants = numpy.array([float(temp[0+10*i:10+10*i].strip()) for i in range(3)])
lattice_angles = numpy.array([float(temp[30+10*i:40+10*i].strip()) for i in range(3)]) * numpy.pi / 180.0
things_to_save.extend(['lattice_type', 'lattice_constants', 'lattice_angles'])
except IOError:
raise "convert_misc_input: reading file %s failed" %self.struct_file
# Read relevant data from .outputs file
#######################################
# rot_symmetries: matrix representation of all (space group) symmetry operations
if (os.path.exists(self.outputs_file)):
mpi.report("Reading input from %s..."%self.outputs_file)
rot_symmetries = []
with open(self.outputs_file) as R:
try:
while 1:
temp = R.readline().strip(' ').split()
if (temp[0] =='PGBSYM:'):
n_symmetries = int(temp[-1])
break
for i in range(n_symmetries):
while 1:
if (R.readline().strip().split()[0] == 'Symmetry'): break
sym_i = numpy.zeros((3, 3), dtype = float)
for ir in range(3):
temp = R.readline().strip().split()
for ic in range(3):
sym_i[ir, ic] = float(temp[ic])
R.readline()
rot_symmetries.append(sym_i)
things_to_save.extend(['n_symmetries', 'rot_symmetries'])
things_to_save.append('rot_symmetries')
except IOError:
raise "convert_misc_input: reading file %s failed" %self.outputs_file
# Save it to the HDF:
ar=HDFArchive(self.hdf_file,'a')
if not (misc_subgrp in ar): ar.create_group(misc_subgrp)
for it in things_to_save: ar[misc_subgrp][it] = locals()[it]
del ar
def convert_transport_input(self):
"""
Reads the input files necessary for transport calculations
and stores the data in the HDFfile
"""
if not (mpi.is_master_node()): return
# Check if SP, SO and n_k are already in h5
ar = HDFArchive(self.hdf_file, 'a')
if not (self.dft_subgrp in ar): raise IOError, "convert_transport_input: No %s subgroup in hdf file found! Call convert_dmft_input first." %self.dft_subgrp
SP = ar[self.dft_subgrp]['SP']
SO = ar[self.dft_subgrp]['SO']
n_k = ar[self.dft_subgrp]['n_k']
del ar
# Read relevant data from .pmat/up/dn files
###########################################
# band_window_optics: Contains the index of the lowest and highest band within the
# band window (used by optics) for each k-point.
# velocities_k: velocity (momentum) matrix elements between all bands in band_window_optics
# and each k-point.
if (SP == 0 or SO == 1):
files = [self.pmat_file]
elif SP == 1:
files = [self.pmat_file+'up', self.pmat_file+'dn']
else: # SO and SP can't both be 1
assert 0, "convert_transport_input: Reading velocity file error! Check SP and SO!"
velocities_k = [[] for f in files]
band_window_optics = []
for isp, f in enumerate(files):
if not os.path.exists(f) : raise IOError, "convert_transport_input: File %s does not exist" %f
mpi.report("Reading input from %s..."%f)
R = ConverterTools.read_fortran_file(self, f, {'D':'E','(':'',')':'',',':' '})
band_window_optics_isp = []
for ik in xrange(n_k):
R.next()
nu1 = int(R.next())
nu2 = int(R.next())
band_window_optics_isp.append((nu1, nu2))
n_bands = nu2 - nu1 + 1
for _ in range(4): R.next()
if n_bands <= 0:
velocity_xyz = numpy.zeros((1, 1, 3), dtype = complex)
else:
velocity_xyz = numpy.zeros((n_bands, n_bands, 3), dtype = complex)
for nu_i in range(n_bands):
for nu_j in range(nu_i, n_bands):
for i in range(3):
velocity_xyz[nu_i][nu_j][i] = R.next() + R.next()*1j
if (nu_i != nu_j): velocity_xyz[nu_j][nu_i][i] = velocity_xyz[nu_i][nu_j][i].conjugate()
velocities_k[isp].append(velocity_xyz)
band_window_optics.append(numpy.array(band_window_optics_isp))
R.close() # Reading done!
# Put data to HDF5 file
ar = HDFArchive(self.hdf_file, 'a')
if not (self.transp_subgrp in ar): ar.create_group(self.transp_subgrp)
# The subgroup containing the data. If it does not exist, it is created. If it exists, the data is overwritten!!!
things_to_save = ['band_window_optics', 'velocities_k']
for it in things_to_save: ar[self.transp_subgrp][it] = locals()[it]
del ar
def convert_symmetry_input(self, ctrl_head, orbits, symm_subgrp):
"""
Reads input for the symmetrisations from symm_file, which is case.sympar or case.symqmc.
"""
# In VASP interface the symmetries are read directly from *.ctrl file
# For the moment the symmetry parameters are just stubs
n_symm = 0
n_atoms = 1
perm = [0]
n_orbits = len(orbits)
SP = ctrl_head['ns']
SO = ctrl_head['nc_flag']
time_inv = [0]
mat = [numpy.identity(1)]
mat_tinv = [numpy.identity(1)]
# if not (mpi.is_master_node()): return
# mpi.report("Reading input from %s..."%symm_file)
#
# n_orbits = len(orbits)
#
# R = ConverterTools.read_fortran_file(self,symm_file,self.fortran_to_replace)
#
# try:
# n_symm = int(R.next()) # Number of symmetry operations
# n_atoms = int(R.next()) # number of atoms involved
# perm = [ [int(R.next()) for i in range(n_atoms)] for j in range(n_symm) ] # list of permutations of the atoms
# if SP:
# time_inv = [ int(R.next()) for j in range(n_symm) ] # time inversion for SO coupling
# else:
# time_inv = [ 0 for j in range(n_symm) ]
#
# # Now read matrices:
# mat = []
# for i_symm in range(n_symm):
#
# mat.append( [ numpy.zeros([orbits[orb]['dim'], orbits[orb]['dim']],numpy.complex_) for orb in range(n_orbits) ] )
# for orb in range(n_orbits):
# for i in range(orbits[orb]['dim']):
# for j in range(orbits[orb]['dim']):
# mat[i_symm][orb][i,j] = R.next() # real part
# for i in range(orbits[orb]['dim']):
# for j in range(orbits[orb]['dim']):
# mat[i_symm][orb][i,j] += 1j * R.next() # imaginary part
#
# mat_tinv = [numpy.identity(orbits[orb]['dim'],numpy.complex_)
# for orb in range(n_orbits)]
#
# if ((SO==0) and (SP==0)):
# # here we need an additional time inversion operation, so read it:
# for orb in range(n_orbits):
# for i in range(orbits[orb]['dim']):
# for j in range(orbits[orb]['dim']):
# mat_tinv[orb][i,j] = R.next() # real part
# for i in range(orbits[orb]['dim']):
# for j in range(orbits[orb]['dim']):
# mat_tinv[orb][i,j] += 1j * R.next() # imaginary part
# Save it to the HDF:
ar=HDFArchive(self.hdf_file,'a')
if not (symm_subgrp in ar): ar.create_group(symm_subgrp)
things_to_save = ['n_symm','n_atoms','perm','orbits','SO','SP','time_inv','mat','mat_tinv']
for it in things_to_save:
print "%s:"%(it), locals()[it]
ar[symm_subgrp][it] = locals()[it]
del ar