def strip_ldosblock(param):
param_copy = param.get_dict()
translated_para = FDFDict(param_copy)
translated_para.pop('%block localdensityofstates')
return Dict(dict=translated_para)
def test_fdfdict_no_argument():
"""
Simple test of a FDFDict class instance with no
argument passed, also tests `get_last_untranslated_key`
"""
f = FDFDict()
# insertion and saving
f["A-_.B"] = 1
assert ("a_b" in list(f._storage.keys()))
assert f["a_:.b"] == 1
assert f.get_last_untranslated_key("a::_b") == "A-_.B"
def initialize(self):
super().initialize()
self.ctx.collectwcinfo = []
# We are going to overwrite the initial structure if volume_per_atom is provided
if "volume_per_atom" in self.ctx.inputs:
self.ctx.inputs.structure = scale_to_vol(
self.ctx.inputs.structure, self.ctx.inputs.volume_per_atom)
test_input_params = FDFDict(self.ctx.inputs.parameters.get_dict())
for k, v in sorted(test_input_params.get_filtered_items()):
if k in ('mdvariablecell', 'mdrelaxcellonly'):
if v is True or v == "T" or v == "true" or v == ".true.":
self.report(
'WARNING: Relaxation with variable cell detected! '
'No action taken, but are you sure this is what you want?'
)
def test_fdfdict_wrong_argument():
"""
Simple test of a FDFDict class instance with an unaccepted
argument passed
"""
import pytest
with pytest.raises(RuntimeError):
f = FDFDict("w")
def test_fdfdict():
"""
Simple test of a FDFDict class instance with a dictionary passed
"""
inp_dict = {"w": 3, "e": 4, "w--": 5}
f = FDFDict(inp_dict)
assert f["w"] == 5
f["e-"] = 4
assert f.get_last_untranslated_key("e") == "e-"
assert "w" in f
for i in f:
assert i in ["w", "e"]
def test_fdfdict_methods():
"""
Simple test of a FDFDict class instance with a dictionary passed
"""
inp_dict = {"w": 3, "e": 4, "w--": 5}
f = FDFDict(inp_dict)
assert sorted(f.values()) == sorted([5, 4])
assert sorted(f.keys()) == sorted(["w", "e"])
assert sorted(f.untranslated_keys()) == sorted(["w--", "e"])
assert sorted(f.items()) == sorted([("w", 5), ("e", 4)])
assert sorted(f.untranslated_items()) == sorted([("w--", 5), ("e", 4)])
assert f.get_dict() == {"w": 5, "e": 4}
assert f.get_untranslated_dict() == {"w--": 5, "e": 4}
def bandskpoints_warnings(value):
"""
Called in validate_inputs. Only issue warnings.
"""
import warnings
if "bandskpoints" in value:
bandskpoints = value["bandskpoints"]
input_params = FDFDict(value["parameters"].get_dict())
# We send a warning if user set a cell in `bandskpoints` and this cell is not the input cell.
# This cell is anyway ignored since the inputs or output structure are used for the kpoints path.
kpcell = bandskpoints.get_attribute("cell", None)
if kpcell:
if kpcell != value["structure"].cell:
warnings.warn(
'The cell set in `bandskpoints` is ignored! Overridden by output or input structure.'
)
#second we rise a warning about consequences when the cell is relaxed
var_cell_keys = [
FDFDict.translate_key("md-variable-cell"),
FDFDict.translate_key("md-constant-volume")
]
var_cell_keys.append(FDFDict.translate_key("md-relax-cell-only"))
for key in input_params:
if key in var_cell_keys:
logline = (
"Requested calculation of bands after a relaxation with variable cell! "
+
"Are you sure you are happy about the selected kpoints for bands? Cell symmetry might change! "
+ "It is suggested to use the `BandGapWorkChain`.")
if isinstance(input_params[key], str):
if FDFDict.translate_key(input_params[key]) in [
"t", "true", "yes", ".true."
]:
warnings.warn(logline)
break
else:
if input_params[key] is True:
warnings.warn(logline)
break
def preprocess(self):
"""
In the preprocess, we make decisions on bandskpoints if they are not requested in input.
In case of single point calculation, bandskpoints are added using seekpath.
In case of relaxation, the relaxation is run, but an extra step at the end of
the calculation will calculate the bands.
"""
self.ctx.need_fin_step = False
var_geom = False
self.ctx.need_to_generate_bandskp = False
#The relaxation in Siesta is triggered by md-steps keyword.
#I verified taht md-type of run alone do not trigger veriable geometry.
fdf_par = FDFDict(self.inputs.parameters.get_dict())
for item in fdf_par:
if item in [
FDFDict.translate_key("mdsteps"),
FDFDict.translate_key("mdnumcgsteps")
]:
if fdf_par[item] != 0:
var_geom = True
break
if "bandskpoints" not in self.inputs:
if var_geom:
self.report(
"The kpoints path for the calculation of bands will be automatically generated "
"using seekpath. Because a relaxation was requested, the bands calculation will "
"be performed on a separate final step. The cell of the final step might change "
"due to seekpath. This cell is returned in `output_structure`."
)
self.ctx.need_fin_step = True
else:
self.report(
"The kpoints path for the calculation of bands will be automatically generated "
"using seekpath. Because of seekpath, the cell might change."
)
self.ctx.need_to_generate_bandskp = True
def validate_parameters(value, _):
"""
Validate parameters input port. Looks for blocked keywords (defined as attribute of SiestaCalculation)
and that pao infos are not here (they belong to the basis Dict).
"""
if value:
input_params = FDFDict(value.get_dict())
for key in input_params:
if "pao" in key:
return "you can't have PAO options in the parameters input port, they belong to the basis input port."
if key in SiestaCalculation._aiida_blocked_keywords:
message = (
f"you can't specify explicitly the '{input_params.get_last_untranslated_key(key)}' flag "
+ "in the input parameters.")
return message
def handle_error_split_norm(self, node):
"""
The split_norm parameter was too small. We need to change it and restart.
The minimum split_norm is stored in the logs of the old calculation.
This error happens only at the beginning of the run, therefore no real restart needed.
Just a new calculation with a new split_norm.
"""
self.report(
f'SiestaCalculation<{node.pk}> crashed with split_norm issue.')
#Retrive the minimum split norm from the logs of failed calc.
logs = orm.Log.objects.get_logs_for(node)
for log in logs:
if "Error in split_norm option" in log.message:
mylog = log.message.split()
new_split_norm = float(mylog[-1]) + 0.001
# We want to understand the presence of "pao-split-norm" in input and:
# 1) if present, we change its value to the minimum allowed
# 2) if not present, we activate pao-SplitTailNorm
# As we don't know in which sintax the user passed "pao-split-norm (remember that every fdf variant
# is allowed), we translate the original dict to a FDFDict that is aware of equivalent keyword.
transl_basis = FDFDict(self.ctx.inputs["basis"].get_dict())
glob_split_norm = False
for key in transl_basis:
if key == "paosplitnorm":
glob_split_norm = True
if glob_split_norm:
self.report('Resetting the pao-split-norm global value')
transl_basis["pao-split-norm"] = new_split_norm
else:
self.report(
'Adding pao-SplitTailNorm to solve the split_norm problem')
transl_basis["pao-SplitTailNorm"] = True
new_basis = orm.Dict(dict=transl_basis)
self.ctx.inputs["basis"] = new_basis
return ProcessHandlerReport(do_break=True)
def checks(self): # noqa: MC0001 - is mccabe too complex funct -
"""
Checks on inputs and definition of few variables useful in the next steps
"""
stm_code = self.inputs.stm_code
code = self.inputs.code
mode = self.inputs.stm_mode.value
spinstm = self.inputs.stm_spin.value
param_dict = self.inputs.parameters.get_dict()
translatedkey = FDFDict(param_dict)
if code.computer.pk != stm_code.computer.pk:
raise ValueError(
"The siesta code and the stm code must be on the same computer!"
)
allowedmodes = ["constant-height", "constant-current"]
if mode not in allowedmodes:
raise ValueError(
f"The allowed options for the port 'stm_mode' are {allowedmodes}"
)
#Some logic regarding the spin. Example: the user can request in input
#of the workchain a stm_spin "non-collinear", but the underline
#siesta calculation was performed with no spin, a warning must be issued
#and the stm analysis is carried in non-spin mode. The opposite however
#is allowed, stm_spin "none" can be generated also from the "non-collinear"
#calculations. Sintax of both Siesta4.0 and Siesta>4.1 is supported.
allowedspins = ["none", "collinear", "non-collinear"]
if spinstm not in allowedspins:
raise ValueError(
f"The allowed options for the port 'stm_spin' are {allowedspins}"
)
spinsiesta = "none"
newversintax = False
oldversintax = False
for k, v in sorted(translatedkey.get_filtered_items()):
if k == "spinpolarized":
oldversintax = True
if v is True or v == "T" or v == "true" or v == ".true.":
spinsiesta = "coll"
if oldversintax:
for k, v in sorted(translatedkey.get_filtered_items()):
if k == "noncollinearspin":
if v is True or v == "T" or v == "true" or v == ".true.":
spinsiesta = "noncoll"
for k, v in sorted(translatedkey.get_filtered_items()):
if k == "spin":
newversintax = True
translatevalue = FDFDict.translate_key(v)
if translatevalue in ('noncollinear', 'spinorbit'):
spinsiesta = "noncoll"
if translatevalue == "polarized":
spinsiesta = "coll"
if newversintax and oldversintax:
self.report(
'WARNING: in the siesta input parameters, both keywork "spin" and '
'"spinpolarized" have been detected. This might confuse the WorkChain and return '
'unexpected outputs')
if spinsiesta == "none" and spinstm != "none":
self.report(
'WARNING: Requested STM with spin option, '
'but the DFT run was performed with no spin. "stm_spin" is reset to "none"'
)
spinstm = "none"
if spinsiesta == "coll" and spinstm == "non-collinear":
self.report(
'WARNING: Requested STM with spin non-collinear option, '
'but the DFT run was performed with collinear spin. "stm_spin" is reset to "collinear"'
)
spinstm = "collinear"
self.ctx.spinstm = spinstm
#LDOS check, the inputs "emax" and "emin" define the energy range for the calculation of the
#ldos. If a block "localdensityofstates" is found in the parameters of the siesta calculation,
#a warining is issued and the block is stripped.
self.ctx.ldosdefinedinparam = False
for k, v in sorted(translatedkey.get_filtered_items()):
if k == "%block localdensityofstates":
self.ctx.ldosdefinedinparam = True
self.report(
'WARNING: A local-density-of-state block was defined in the input parameters, however '
'this will be ignored as the input ports "emin" and "emax" define the energy of LDOS'
)
def drop_md_keys(param):
for item in param.copy().keys():
trans_item = FDFDict.translate_key(item)
if trans_item.startswith("md"):
param.pop(item)
return param
def prepare_for_submission(
self, folder): # noqa: MC0001 - is mccabe too complex funct -
"""
Create the input files from the input nodes passed to this instance of the `CalcJob`.
:param folder: an `aiida.common.folders.Folder` to temporarily write files on disk
:return: `aiida.common.datastructures.CalcInfo` instance
"""
# ============================ Initializations =============================
# All input ports are validated, here asses their presence in case optional.
code = self.inputs.code
# self.initialize preprocess structure and basis. Decides whether use ions or pseudos
structure, basis_dict, floating_species_names, ion_or_pseudo_str = self.initialize(
)
ion_or_pseudo = self.inputs[ion_or_pseudo_str]
parameters = self.inputs.parameters
if 'kpoints' in self.inputs:
kpoints = self.inputs.kpoints
else:
kpoints = None
# As internal convention, the keys of the settings dict are uppercase
if 'settings' in self.inputs:
settings = self.inputs.settings.get_dict()
settings_dict = {str(k).upper(): v for (k, v) in settings.items()}
else:
settings_dict = {}
if 'bandskpoints' in self.inputs:
bandskpoints = self.inputs.bandskpoints
else:
bandskpoints = None
if 'parent_calc_folder' in self.inputs:
parent_calc_folder = self.inputs.parent_calc_folder
else:
parent_calc_folder = None
lua_inputs = self.inputs.lua
if 'script' in lua_inputs:
lua_script = lua_inputs.script
else:
lua_script = None
if 'parameters' in lua_inputs:
lua_parameters = lua_inputs.parameters
else:
lua_parameters = None
if 'input_files' in lua_inputs:
lua_input_files = lua_inputs.input_files
else:
lua_input_files = None
if 'retrieve_list' in lua_inputs:
lua_retrieve_list = lua_inputs.retrieve_list
else:
lua_retrieve_list = None
# List of files to copy in the folder where the calculation runs, e.g. pseudo files
local_copy_list = []
# List of files for restart
remote_copy_list = []
# ================ Preprocess of input parameters =================
input_params = FDFDict(parameters.get_dict())
input_params.update(
{'system-name': self.inputs.metadata.options.prefix})
input_params.update(
{'system-label': self.inputs.metadata.options.prefix})
input_params.update({'use-tree-timer': 'T'})
input_params.update({'xml-write': 'T'})
input_params.update({'number-of-species': len(structure.kinds)})
input_params.update({'number-of-atoms': len(structure.sites)})
input_params.update({'geometry-must-converge': 'T'})
input_params.update({'lattice-constant': '1.0 Ang'})
input_params.update({'atomic-coordinates-format': 'Ang'})
if lua_script is not None:
input_params.update({'md-type-of-run': 'Lua'})
input_params.update({'lua-script': lua_script.filename})
local_copy_list.append(
(lua_script.uuid, lua_script.filename, lua_script.filename))
if lua_input_files is not None:
# Copy the whole contents of the FolderData object
for file in lua_input_files.list_object_names():
local_copy_list.append((lua_input_files.uuid, file, file))
if ion_or_pseudo_str == "ions":
input_params.update({'user-basis': 'T'})
# NOTES:
# 1) The lattice-constant parameter must be 1.0 Ang to impose the units and consider
# that the dimenstions of the lattice vectors are already correct with no need of alat.
# This breaks the band-k-points "pi/a" option. The use of this option is banned.
# 2) The implicit coordinate convention of the StructureData class corresponds to the "Ang"
# convention in Siesta. That is why "atomic-coordinates-format" is blocked and reset.
# 3) The Siesta code doesn't raise any warining if the geometry is not converged, unless
# the keyword geometry-must-converge is set. That's why it is always added.
# ============================ Preparation of input data =================================
# -------------------------------- CELL_PARAMETERS ---------------------------------------
cell_parameters_card = "%block lattice-vectors\n"
for vector in structure.cell:
cell_parameters_card += ("{0:18.10f} {1:18.10f} {2:18.10f}"
"\n".format(*vector))
cell_parameters_card += "%endblock lattice-vectors\n"
# ----------------------------ATOMIC_SPECIES & PSEUDOS/IONS-------------------------------
atomic_species_card_list = []
# Dictionary to get the atomic number of a given element
datmn = {v['symbol']: k for k, v in elements.items()}
spind = {}
spcount = 0
for kind in structure.kinds:
spcount += 1 # species count
spind[kind.name] = spcount
atomic_number = datmn[kind.symbol]
# Siesta expects negative atomic numbers for floating species
if kind.name in floating_species_names:
atomic_number = -atomic_number
#Create the core of the chemicalspecieslabel block
atomic_species_card_list.append("{0:5} {1:5} {2:5}\n".format(
spind[kind.name], atomic_number, kind.name.rjust(6)))
psp_or_ion = ion_or_pseudo[kind.name]
# Add pseudo (ion) file to the list of files to copy (create), with the appropiate name.
# In the case of sub-species (different kind.name but same kind.symbol, e.g., 'C_surf',
# sharing the same pseudo with 'C'), we copy the file ('C.psf') twice, once as 'C.psf',
# and once as 'C_surf.psf'. This is required by Siesta.
# It is passed as list of tuples with format ('node_uuid', 'filename', 'relativedestpath').
# Since no subfolder is present in Siesta for pseudos, filename == relativedestpath.
if isinstance(psp_or_ion, IonData):
file_name = kind.name + ".ion"
with folder.open(file_name, 'w', encoding='utf8') as handle:
handle.write(psp_or_ion.get_content_ascii_format())
if isinstance(psp_or_ion, (PsfData, DeprecatedPsfData)):
local_copy_list.append(
(psp_or_ion.uuid, psp_or_ion.filename, kind.name + ".psf"))
if isinstance(psp_or_ion, (PsmlData, DeprecatedPsmlData)):
local_copy_list.append((psp_or_ion.uuid, psp_or_ion.filename,
kind.name + ".psml"))
atomic_species_card_list = (["%block chemicalspecieslabel\n"] +
list(atomic_species_card_list))
atomic_species_card = "".join(atomic_species_card_list)
atomic_species_card += "%endblock chemicalspecieslabel\n"
# Free memory
del atomic_species_card_list
# -------------------------------------- ATOMIC_POSITIONS -----------------------------------
atomic_positions_card_list = [
"%block atomiccoordinatesandatomicspecies\n"
]
countatm = 0
for site in structure.sites:
countatm += 1
atomic_positions_card_list.append(
"{0:18.10f} {1:18.10f} {2:18.10f} {3:4} {4:6} {5:6}\n".format(
site.position[0], site.position[1], site.position[2],
spind[site.kind_name], site.kind_name.rjust(6), countatm))
atomic_positions_card = "".join(atomic_positions_card_list)
del atomic_positions_card_list # Free memory
atomic_positions_card += "%endblock atomiccoordinatesandatomicspecies\n"
# --------------------------------------- K-POINTS ----------------------------------------
# It is optional, if not specified, gamma point only is performed (default of siesta)
if kpoints is not None:
mesh, offset = kpoints.get_kpoints_mesh()
kpoints_card_list = ["%block kgrid_monkhorst_pack\n"]
kpoints_card_list.append("{0:6} {1:6} {2:6} {3:18.10f}\n".format(
mesh[0], 0, 0, offset[0]))
kpoints_card_list.append("{0:6} {1:6} {2:6} {3:18.10f}\n".format(
0, mesh[1], 0, offset[1]))
kpoints_card_list.append("{0:6} {1:6} {2:6} {3:18.10f}\n".format(
0, 0, mesh[2], offset[2]))
kpoints_card = "".join(kpoints_card_list)
kpoints_card += "%endblock kgrid_monkhorst_pack\n"
del kpoints_card_list
# ------------------------------------ K-POINTS-FOR-BANDS ----------------------------------
# Two possibility are supported in Siesta: BandLines ad BandPoints.
# User can't choose directly one of the two options, BandsLine is set automatically
# if bandskpoints has labels, BandsPoints if bandskpoints has no labels.
# BandLinesScale=pi/a not supported because a=1 always. BandLinesScale ReciprocalLatticeVectors
# always set.
if bandskpoints is not None:
#the band line scale
bandskpoints_card_list = [
"BandLinesScale ReciprocalLatticeVectors\n"
]
#set the BandPoints
if bandskpoints.labels is None:
bandskpoints_card_list.append("%block BandPoints\n")
for kpo in bandskpoints.get_kpoints():
bandskpoints_card_list.append(
"{0:8.3f} {1:8.3f} {2:8.3f} \n".format(
kpo[0], kpo[1], kpo[2]))
fbkpoints_card = "".join(bandskpoints_card_list)
fbkpoints_card += "%endblock BandPoints\n"
#set the BandLines
else:
bandskpoints_card_list.append("%block BandLines\n")
savindx = []
listforbands = bandskpoints.get_kpoints()
for indx, label in bandskpoints.labels:
savindx.append(indx)
rawindex = 0
for indx, label in bandskpoints.labels:
rawindex = rawindex + 1
x, y, z = listforbands[indx]
if rawindex == 1:
bandskpoints_card_list.append(
"{0:3} {1:8.3f} {2:8.3f} {3:8.3f} {4:1} \n".format(
1, x, y, z, label))
else:
bandskpoints_card_list.append(
"{0:3} {1:8.3f} {2:8.3f} {3:8.3f} {4:1} \n".format(
indx - savindx[rawindex - 2], x, y, z, label))
fbkpoints_card = "".join(bandskpoints_card_list)
fbkpoints_card += "%endblock BandLines\n"
del bandskpoints_card_list
# ================================= Operations for restart =================================
# The presence of a 'parent_calc_folder' input node signals that we want to
# get something from there, as indicated in the self._restart_copy_from attribute.
# In Siesta's case, for now, just the density-matrix file is copied
# to the current calculation's working folder.
# ISSUE: Is this mechanism flexible enough? An alternative would be to
# pass the information about which file(s) to copy in the metadata.options dictionary
if parent_calc_folder is not None:
remote_copy_list.append(
(parent_calc_folder.computer.uuid,
os.path.join(parent_calc_folder.get_remote_path(),
self._restart_copy_from), self._restart_copy_to))
input_params.update({'dm-use-save-dm': "T"})
# ===================================== FDF file creation ====================================
# To have easy access to inputs metadata options
metadataoption = self.inputs.metadata.options
# input_filename = self.inputs.metadata.options.input_filename
input_filename = folder.get_abs_path(metadataoption.input_filename)
# Print to file
with open(input_filename, 'w') as infile:
# Parameters
for k, v in sorted(input_params.get_filtered_items()):
infile.write("%s %s\n" % (k, v))
# Basis set info is processed just like the general parameters section.
if basis_dict: #It migh also be empty dict. In such case we do not write.
infile.write("#\n# -- Basis Set Info follows\n#\n")
for k, v in basis_dict.items():
infile.write("%s %s\n" % (k, v))
# Write previously generated cards now
infile.write("#\n# -- Structural Info follows\n#\n")
infile.write(atomic_species_card)
infile.write(cell_parameters_card)
infile.write(atomic_positions_card)
if kpoints is not None:
infile.write("#\n# -- K-points Info follows\n#\n")
infile.write(kpoints_card)
if bandskpoints is not None:
infile.write("#\n# -- Bandlines/Bandpoints Info follows\n#\n")
infile.write(fbkpoints_card)
# Write max wall-clock time
# This should prevent SiestaCalculation from being terminated by scheduler, however the
# strategy is not 100% effective since SIESTA checks the simulation time versus max-walltime
# only at the end of each SCF and geometry step. The scheduler might kill the process in between.
infile.write("#\n# -- Max wall-clock time block\n#\n")
infile.write(
f"maxwalltime {metadataoption.max_wallclock_seconds}\n")
# ================================== Lua parameters file ===================================
if lua_parameters is not None:
lua_config_filename = folder.get_abs_path("config.lua")
# Generate a 'config.lua' file with Lua syntax
with open(lua_config_filename, 'w') as f_lua:
f_lua.write("--- Lua script parameters \n")
for k, v in lua_parameters.get_dict().items():
if isinstance(v, str):
f_lua.write('%s = "%s"\n' % (k, v))
else:
f_lua.write("%s = %s\n" % (k, v))
# ============================= Code and Calc info =========================================
# Code information object and Calc information object are now
# only used to set up the CMDLINE (the bash line that launches siesta)
# and to set up the list of files to retrieve.
cmdline_params = settings_dict.pop('CMDLINE', [])
codeinfo = CodeInfo()
codeinfo.cmdline_params = list(cmdline_params)
codeinfo.stdin_name = metadataoption.input_filename
codeinfo.stdout_name = metadataoption.output_filename
codeinfo.code_uuid = code.uuid
calcinfo = CalcInfo()
calcinfo.uuid = str(self.uuid)
calcinfo.local_copy_list = local_copy_list
calcinfo.remote_copy_list = remote_copy_list
calcinfo.codes_info = [codeinfo]
# Retrieve by default: the output file, the xml file, the messages file, and the json timing file.
# If bandskpoints, also the bands file is added to the retrieve list.
calcinfo.retrieve_list = []
xml_file = str(metadataoption.prefix) + ".xml"
bands_file = str(metadataoption.prefix) + ".bands"
calcinfo.retrieve_list.append(metadataoption.output_filename)
calcinfo.retrieve_list.append(xml_file)
calcinfo.retrieve_list.append(self._JSON_FILE)
calcinfo.retrieve_list.append(self._MESSAGES_FILE)
calcinfo.retrieve_list.append(self._BASIS_ENTHALPY_FILE)
calcinfo.retrieve_list.append("*.ion.xml")
if bandskpoints is not None:
calcinfo.retrieve_list.append(bands_file)
if lua_retrieve_list is not None:
calcinfo.retrieve_list += lua_retrieve_list.get_list()
# If we ever want to avoid having the config.lua file in the repository,
# since the information is already in the lua_parameters dictionary:
# if lua_parameters is not None:
# calcinfo.provenance_exclude_list = ['config.lua']
# Any other files specified in the settings dictionary
settings_retrieve_list = settings_dict.pop('ADDITIONAL_RETRIEVE_LIST',
[])
calcinfo.retrieve_list += settings_retrieve_list
return calcinfo
class SiestaCalculation(CalcJob):
"""
Siesta calculator class for AiiDA.
"""
# Class attributes: filepaths of certain outputs
_JSON_FILE = 'time.json'
_MESSAGES_FILE = 'MESSAGES'
_BASIS_ENTHALPY_FILE = 'BASIS_ENTHALPY'
# Class attributes: default of the input.spec...just default, but user could change the name
_DEFAULT_PREFIX = 'aiida'
_DEFAULT_INPUT_FILE = 'aiida.fdf'
_DEFAULT_OUTPUT_FILE = 'aiida.out'
# Class attribute: elements to copy from the parent in restarts (fow now, just the density matrix file)
_restart_copy_from = os.path.join('./', '*.DM')
# Class attribute: in restarts, it will copy the previous elements in the following folder
_restart_copy_to = './'
# Class attribute: blocked keywords
_readable_blocked = [
'system-name',
'system-label',
'number-of-species',
'number-of-atoms',
'lattice-constant',
'atomic-coordinates-format',
'use-tree-timer',
'xml-write',
'dm-use-save-dm',
'geometry-must-converge',
'user-basis',
'lua-script',
'max-walltime',
]
_aiida_blocked_keywords = [
FDFDict.translate_key(key) for key in _readable_blocked
]
@classmethod
def define(cls, spec):
super().define(spec)
# Input nodes
spec.input('code', valid_type=orm.Code, help='Input code')
spec.input('structure',
valid_type=orm.StructureData,
help='Input structure',
validator=validate_structure)
spec.input('kpoints',
valid_type=orm.KpointsData,
help='Input kpoints',
required=False,
validator=validate_kpoints)
spec.input('bandskpoints',
valid_type=orm.KpointsData,
help='Input kpoints for bands',
required=False,
validator=validate_bandskpoints)
spec.input('basis',
valid_type=orm.Dict,
help='Input basis',
required=False,
validator=validate_basis)
spec.input('settings',
valid_type=orm.Dict,
help='Input settings',
required=False)
spec.input('parameters',
valid_type=orm.Dict,
help='Input parameters',
validator=validate_parameters)
spec.input('parent_calc_folder',
valid_type=orm.RemoteData,
required=False,
help='Parent folder')
spec.input_namespace('pseudos',
valid_type=(PsfData, PsmlData, DeprecatedPsfData,
DeprecatedPsmlData),
help='Input pseudo potentials',
dynamic=True,
required=False,
validator=validate_pseudos)
spec.input_namespace('ions',
valid_type=IonData,
help='Input ion file',
dynamic=True,
required=False)
# Input namespace for Lua-related material.
# Parameters are in a separate dictionary to enable a reduced set of 'universal' scripts for particular uses.
# Input files (e.g., image files for NEB) should be packaged in a FolderData object.
# Files to be retrieved should be specified in a list o# path specifications.
spec.input_namespace('lua', help='Script and files for the Lua engine')
spec.input('lua.script', valid_type=orm.SinglefileData, required=False)
spec.input('lua.parameters', valid_type=orm.Dict, required=False)
spec.input('lua.input_files',
valid_type=orm.FolderData,
required=False)
spec.input('lua.retrieve_list', valid_type=orm.List, required=False)
# Metadada.options host the inputs that are not stored as a separate node, but attached to `CalcJobNode`
# as attributes. They are optional, since a default is specified, but they might be changed by the user.
# The first one is siesta specific. The others are defined in the CalcJob, here we change the default.
spec.input('metadata.options.prefix',
valid_type=str,
default=cls._DEFAULT_PREFIX)
spec.inputs['metadata']['options'][
'input_filename'].default = cls._DEFAULT_INPUT_FILE
spec.inputs['metadata']['options'][
'output_filename'].default = cls._DEFAULT_OUTPUT_FILE
spec.inputs['metadata']['options'][
'parser_name'].default = 'siesta.parser'
# Global validator for the inputs
spec.inputs.validator = validate_inputs
# Output nodes
spec.output('output_parameters',
valid_type=Dict,
required=True,
help='The calculation results')
spec.output('output_structure',
valid_type=StructureData,
required=False,
help='Optional relaxed structure')
spec.output('bands',
valid_type=BandsData,
required=False,
help='Optional band structure')
spec.output('forces_and_stress',
valid_type=ArrayData,
required=False,
help='Optional forces and stress')
spec.output_namespace('ion_files',
valid_type=IonData,
dynamic=True,
required=False)
# Option that allows access through node.res should be existing output node and a Dict
spec.default_output_node = 'output_parameters'
# Exit codes for specific errors. Useful for error handeling in workchains
spec.exit_code(453,
'BANDS_PARSE_FAIL',
message='Failure while parsing the bands file')
spec.exit_code(
452,
'BANDS_FILE_NOT_PRODUCED',
message='Bands analysis was requested, but file is not present')
spec.exit_code(450,
'SCF_NOT_CONV',
message='Calculation did not reach scf convergence!')
spec.exit_code(
451,
'GEOM_NOT_CONV',
message='Calculation did not reach geometry convergence!')
spec.exit_code(
350,
'UNEXPECTED_TERMINATION',
message='Statement "Job completed" not detected, unknown error')
spec.exit_code(449,
'SPLIT_NORM',
message='Split_norm parameter too small')
spec.exit_code(
448,
'BASIS_POLARIZ',
message='Problems in the polarization of a basis element')
def initialize(self):
"""
Some initialization (called at the beginning of `prepare_for_submission`:
1) Create an internal structure where possible `floating_sites` are added.
2) Create a list containing floating_species_names.
3) Remove the `floating_sites` to the basis dictionary.
4) Checks whether info on basis and pseudos are passed directly as ion files,
in that case, cancel any info passed in the basis input.
"""
value = self.inputs
structure = internal_structure(value["structure"])
floating_species_names = []
basis_dict = None
if 'basis' in value:
basis_dict = value["basis"].get_dict()
floating = basis_dict.pop('floating_sites', None)
if floating is not None:
for item in floating:
structure.append_atom(position=item["position"],
symbols=item["symbols"],
name=item["name"])
floating_species_names.append(item["name"])
if 'ions' in value:
basis_dict = None
ion_or_pseudo = 'ions'
else:
ion_or_pseudo = 'pseudos'
return structure, basis_dict, floating_species_names, ion_or_pseudo
def prepare_for_submission(
self, folder): # noqa: MC0001 - is mccabe too complex funct -
"""
Create the input files from the input nodes passed to this instance of the `CalcJob`.
:param folder: an `aiida.common.folders.Folder` to temporarily write files on disk
:return: `aiida.common.datastructures.CalcInfo` instance
"""
# ============================ Initializations =============================
# All input ports are validated, here asses their presence in case optional.
code = self.inputs.code
# self.initialize preprocess structure and basis. Decides whether use ions or pseudos
structure, basis_dict, floating_species_names, ion_or_pseudo_str = self.initialize(
)
ion_or_pseudo = self.inputs[ion_or_pseudo_str]
parameters = self.inputs.parameters
if 'kpoints' in self.inputs:
kpoints = self.inputs.kpoints
else:
kpoints = None
# As internal convention, the keys of the settings dict are uppercase
if 'settings' in self.inputs:
settings = self.inputs.settings.get_dict()
settings_dict = {str(k).upper(): v for (k, v) in settings.items()}
else:
settings_dict = {}
if 'bandskpoints' in self.inputs:
bandskpoints = self.inputs.bandskpoints
else:
bandskpoints = None
if 'parent_calc_folder' in self.inputs:
parent_calc_folder = self.inputs.parent_calc_folder
else:
parent_calc_folder = None
lua_inputs = self.inputs.lua
if 'script' in lua_inputs:
lua_script = lua_inputs.script
else:
lua_script = None
if 'parameters' in lua_inputs:
lua_parameters = lua_inputs.parameters
else:
lua_parameters = None
if 'input_files' in lua_inputs:
lua_input_files = lua_inputs.input_files
else:
lua_input_files = None
if 'retrieve_list' in lua_inputs:
lua_retrieve_list = lua_inputs.retrieve_list
else:
lua_retrieve_list = None
# List of files to copy in the folder where the calculation runs, e.g. pseudo files
local_copy_list = []
# List of files for restart
remote_copy_list = []
# ================ Preprocess of input parameters =================
input_params = FDFDict(parameters.get_dict())
input_params.update(
{'system-name': self.inputs.metadata.options.prefix})
input_params.update(
{'system-label': self.inputs.metadata.options.prefix})
input_params.update({'use-tree-timer': 'T'})
input_params.update({'xml-write': 'T'})
input_params.update({'number-of-species': len(structure.kinds)})
input_params.update({'number-of-atoms': len(structure.sites)})
input_params.update({'geometry-must-converge': 'T'})
input_params.update({'lattice-constant': '1.0 Ang'})
input_params.update({'atomic-coordinates-format': 'Ang'})
if lua_script is not None:
input_params.update({'md-type-of-run': 'Lua'})
input_params.update({'lua-script': lua_script.filename})
local_copy_list.append(
(lua_script.uuid, lua_script.filename, lua_script.filename))
if lua_input_files is not None:
# Copy the whole contents of the FolderData object
for file in lua_input_files.list_object_names():
local_copy_list.append((lua_input_files.uuid, file, file))
if ion_or_pseudo_str == "ions":
input_params.update({'user-basis': 'T'})
# NOTES:
# 1) The lattice-constant parameter must be 1.0 Ang to impose the units and consider
# that the dimenstions of the lattice vectors are already correct with no need of alat.
# This breaks the band-k-points "pi/a" option. The use of this option is banned.
# 2) The implicit coordinate convention of the StructureData class corresponds to the "Ang"
# convention in Siesta. That is why "atomic-coordinates-format" is blocked and reset.
# 3) The Siesta code doesn't raise any warining if the geometry is not converged, unless
# the keyword geometry-must-converge is set. That's why it is always added.
# ============================ Preparation of input data =================================
# -------------------------------- CELL_PARAMETERS ---------------------------------------
cell_parameters_card = "%block lattice-vectors\n"
for vector in structure.cell:
cell_parameters_card += ("{0:18.10f} {1:18.10f} {2:18.10f}"
"\n".format(*vector))
cell_parameters_card += "%endblock lattice-vectors\n"
# ----------------------------ATOMIC_SPECIES & PSEUDOS/IONS-------------------------------
atomic_species_card_list = []
# Dictionary to get the atomic number of a given element
datmn = {v['symbol']: k for k, v in elements.items()}
spind = {}
spcount = 0
for kind in structure.kinds:
spcount += 1 # species count
spind[kind.name] = spcount
atomic_number = datmn[kind.symbol]
# Siesta expects negative atomic numbers for floating species
if kind.name in floating_species_names:
atomic_number = -atomic_number
#Create the core of the chemicalspecieslabel block
atomic_species_card_list.append("{0:5} {1:5} {2:5}\n".format(
spind[kind.name], atomic_number, kind.name.rjust(6)))
psp_or_ion = ion_or_pseudo[kind.name]
# Add pseudo (ion) file to the list of files to copy (create), with the appropiate name.
# In the case of sub-species (different kind.name but same kind.symbol, e.g., 'C_surf',
# sharing the same pseudo with 'C'), we copy the file ('C.psf') twice, once as 'C.psf',
# and once as 'C_surf.psf'. This is required by Siesta.
# It is passed as list of tuples with format ('node_uuid', 'filename', 'relativedestpath').
# Since no subfolder is present in Siesta for pseudos, filename == relativedestpath.
if isinstance(psp_or_ion, IonData):
file_name = kind.name + ".ion"
with folder.open(file_name, 'w', encoding='utf8') as handle:
handle.write(psp_or_ion.get_content_ascii_format())
if isinstance(psp_or_ion, (PsfData, DeprecatedPsfData)):
local_copy_list.append(
(psp_or_ion.uuid, psp_or_ion.filename, kind.name + ".psf"))
if isinstance(psp_or_ion, (PsmlData, DeprecatedPsmlData)):
local_copy_list.append((psp_or_ion.uuid, psp_or_ion.filename,
kind.name + ".psml"))
atomic_species_card_list = (["%block chemicalspecieslabel\n"] +
list(atomic_species_card_list))
atomic_species_card = "".join(atomic_species_card_list)
atomic_species_card += "%endblock chemicalspecieslabel\n"
# Free memory
del atomic_species_card_list
# -------------------------------------- ATOMIC_POSITIONS -----------------------------------
atomic_positions_card_list = [
"%block atomiccoordinatesandatomicspecies\n"
]
countatm = 0
for site in structure.sites:
countatm += 1
atomic_positions_card_list.append(
"{0:18.10f} {1:18.10f} {2:18.10f} {3:4} {4:6} {5:6}\n".format(
site.position[0], site.position[1], site.position[2],
spind[site.kind_name], site.kind_name.rjust(6), countatm))
atomic_positions_card = "".join(atomic_positions_card_list)
del atomic_positions_card_list # Free memory
atomic_positions_card += "%endblock atomiccoordinatesandatomicspecies\n"
# --------------------------------------- K-POINTS ----------------------------------------
# It is optional, if not specified, gamma point only is performed (default of siesta)
if kpoints is not None:
mesh, offset = kpoints.get_kpoints_mesh()
kpoints_card_list = ["%block kgrid_monkhorst_pack\n"]
kpoints_card_list.append("{0:6} {1:6} {2:6} {3:18.10f}\n".format(
mesh[0], 0, 0, offset[0]))
kpoints_card_list.append("{0:6} {1:6} {2:6} {3:18.10f}\n".format(
0, mesh[1], 0, offset[1]))
kpoints_card_list.append("{0:6} {1:6} {2:6} {3:18.10f}\n".format(
0, 0, mesh[2], offset[2]))
kpoints_card = "".join(kpoints_card_list)
kpoints_card += "%endblock kgrid_monkhorst_pack\n"
del kpoints_card_list
# ------------------------------------ K-POINTS-FOR-BANDS ----------------------------------
# Two possibility are supported in Siesta: BandLines ad BandPoints.
# User can't choose directly one of the two options, BandsLine is set automatically
# if bandskpoints has labels, BandsPoints if bandskpoints has no labels.
# BandLinesScale=pi/a not supported because a=1 always. BandLinesScale ReciprocalLatticeVectors
# always set.
if bandskpoints is not None:
#the band line scale
bandskpoints_card_list = [
"BandLinesScale ReciprocalLatticeVectors\n"
]
#set the BandPoints
if bandskpoints.labels is None:
bandskpoints_card_list.append("%block BandPoints\n")
for kpo in bandskpoints.get_kpoints():
bandskpoints_card_list.append(
"{0:8.3f} {1:8.3f} {2:8.3f} \n".format(
kpo[0], kpo[1], kpo[2]))
fbkpoints_card = "".join(bandskpoints_card_list)
fbkpoints_card += "%endblock BandPoints\n"
#set the BandLines
else:
bandskpoints_card_list.append("%block BandLines\n")
savindx = []
listforbands = bandskpoints.get_kpoints()
for indx, label in bandskpoints.labels:
savindx.append(indx)
rawindex = 0
for indx, label in bandskpoints.labels:
rawindex = rawindex + 1
x, y, z = listforbands[indx]
if rawindex == 1:
bandskpoints_card_list.append(
"{0:3} {1:8.3f} {2:8.3f} {3:8.3f} {4:1} \n".format(
1, x, y, z, label))
else:
bandskpoints_card_list.append(
"{0:3} {1:8.3f} {2:8.3f} {3:8.3f} {4:1} \n".format(
indx - savindx[rawindex - 2], x, y, z, label))
fbkpoints_card = "".join(bandskpoints_card_list)
fbkpoints_card += "%endblock BandLines\n"
del bandskpoints_card_list
# ================================= Operations for restart =================================
# The presence of a 'parent_calc_folder' input node signals that we want to
# get something from there, as indicated in the self._restart_copy_from attribute.
# In Siesta's case, for now, just the density-matrix file is copied
# to the current calculation's working folder.
# ISSUE: Is this mechanism flexible enough? An alternative would be to
# pass the information about which file(s) to copy in the metadata.options dictionary
if parent_calc_folder is not None:
remote_copy_list.append(
(parent_calc_folder.computer.uuid,
os.path.join(parent_calc_folder.get_remote_path(),
self._restart_copy_from), self._restart_copy_to))
input_params.update({'dm-use-save-dm': "T"})
# ===================================== FDF file creation ====================================
# To have easy access to inputs metadata options
metadataoption = self.inputs.metadata.options
# input_filename = self.inputs.metadata.options.input_filename
input_filename = folder.get_abs_path(metadataoption.input_filename)
# Print to file
with open(input_filename, 'w') as infile:
# Parameters
for k, v in sorted(input_params.get_filtered_items()):
infile.write("%s %s\n" % (k, v))
# Basis set info is processed just like the general parameters section.
if basis_dict: #It migh also be empty dict. In such case we do not write.
infile.write("#\n# -- Basis Set Info follows\n#\n")
for k, v in basis_dict.items():
infile.write("%s %s\n" % (k, v))
# Write previously generated cards now
infile.write("#\n# -- Structural Info follows\n#\n")
infile.write(atomic_species_card)
infile.write(cell_parameters_card)
infile.write(atomic_positions_card)
if kpoints is not None:
infile.write("#\n# -- K-points Info follows\n#\n")
infile.write(kpoints_card)
if bandskpoints is not None:
infile.write("#\n# -- Bandlines/Bandpoints Info follows\n#\n")
infile.write(fbkpoints_card)
# Write max wall-clock time
# This should prevent SiestaCalculation from being terminated by scheduler, however the
# strategy is not 100% effective since SIESTA checks the simulation time versus max-walltime
# only at the end of each SCF and geometry step. The scheduler might kill the process in between.
infile.write("#\n# -- Max wall-clock time block\n#\n")
infile.write(
f"maxwalltime {metadataoption.max_wallclock_seconds}\n")
# ================================== Lua parameters file ===================================
if lua_parameters is not None:
lua_config_filename = folder.get_abs_path("config.lua")
# Generate a 'config.lua' file with Lua syntax
with open(lua_config_filename, 'w') as f_lua:
f_lua.write("--- Lua script parameters \n")
for k, v in lua_parameters.get_dict().items():
if isinstance(v, str):
f_lua.write('%s = "%s"\n' % (k, v))
else:
f_lua.write("%s = %s\n" % (k, v))
# ============================= Code and Calc info =========================================
# Code information object and Calc information object are now
# only used to set up the CMDLINE (the bash line that launches siesta)
# and to set up the list of files to retrieve.
cmdline_params = settings_dict.pop('CMDLINE', [])
codeinfo = CodeInfo()
codeinfo.cmdline_params = list(cmdline_params)
codeinfo.stdin_name = metadataoption.input_filename
codeinfo.stdout_name = metadataoption.output_filename
codeinfo.code_uuid = code.uuid
calcinfo = CalcInfo()
calcinfo.uuid = str(self.uuid)
calcinfo.local_copy_list = local_copy_list
calcinfo.remote_copy_list = remote_copy_list
calcinfo.codes_info = [codeinfo]
# Retrieve by default: the output file, the xml file, the messages file, and the json timing file.
# If bandskpoints, also the bands file is added to the retrieve list.
calcinfo.retrieve_list = []
xml_file = str(metadataoption.prefix) + ".xml"
bands_file = str(metadataoption.prefix) + ".bands"
calcinfo.retrieve_list.append(metadataoption.output_filename)
calcinfo.retrieve_list.append(xml_file)
calcinfo.retrieve_list.append(self._JSON_FILE)
calcinfo.retrieve_list.append(self._MESSAGES_FILE)
calcinfo.retrieve_list.append(self._BASIS_ENTHALPY_FILE)
calcinfo.retrieve_list.append("*.ion.xml")
if bandskpoints is not None:
calcinfo.retrieve_list.append(bands_file)
if lua_retrieve_list is not None:
calcinfo.retrieve_list += lua_retrieve_list.get_list()
# If we ever want to avoid having the config.lua file in the repository,
# since the information is already in the lua_parameters dictionary:
# if lua_parameters is not None:
# calcinfo.provenance_exclude_list = ['config.lua']
# Any other files specified in the settings dictionary
settings_retrieve_list = settings_dict.pop('ADDITIONAL_RETRIEVE_LIST',
[])
calcinfo.retrieve_list += settings_retrieve_list
return calcinfo
@classmethod
def inputs_generator(cls): # pylint: disable=no-self-argument,no-self-use
from aiida_siesta.utils.protocols_system.input_generators import SiestaCalculationInputGenerator
return SiestaCalculationInputGenerator(cls)