# -*- coding: utf-8 -*-
###############################################################################
# Copyright (c), Forschungszentrum Jülich GmbH, IAS-1/PGI-1, Germany. #
# All rights reserved. #
# This file is part of the AiiDA-FLEUR package. #
# #
# The code is hosted on GitHub at https://github.com/JuDFTteam/aiida-fleur #
# For further information on the license, see the LICENSE.txt file #
# For further information please visit http://www.flapw.de or #
# http://aiida-fleur.readthedocs.io/en/develop/ #
###############################################################################
"""
In this module is the :class:`~aiida_fleur.data.fleurinp.FleurinpData` class, and methods for FLEUR
input manipulation plus methods for extration of AiiDA data structures.
"""
# TODO: maybe add a modify method which returns a fleurinpmodifier class
# TODO: inpxml to dict: maybe kpts should not be written to the dict? same with symmetry
# TODO: test for large input files, I believe the recursion is still quite slow..
# TODO: 2D cell get kpoints and get structure also be carefull with tria = T!!!
# TODO : maybe save when get_structure or get_kpoints was executed on fleurinp,
# because otherwise return this node instead of creating a new one!
# TODO: get rid of duplicate code for parsing the inp.xml to an etree
from __future__ import absolute_import
from __future__ import print_function
import os
import io
import re
import six
from lxml import etree
import warnings
from aiida.orm import Data, Node, load_node, CalcJobNode
from aiida.common.exceptions import InputValidationError, ValidationError
from aiida.engine.processes.functions import calcfunction as cf
[docs]def get_fleurinp_from_folder_data(folder_node, store=False, additional_files=None):
"""
Create FleurinpData object from the given RemoteData object
:param remote_node: RemoteData to use for the generation of the FleurinpData
:param store: bool, if True the FleurinpData object will be stored after generation
:returns: FleurinpData object with the input xml files from the retrieved folder
of the calculation associated RemoteData
"""
if additional_files is None:
additional_files = []
input_xml_files = [file for file in folder_node.list_object_names() if file.endswith('.xml') and 'out' not in file]
fleurinp = FleurinpData(files=input_xml_files + additional_files, node=folder_node)
if store:
fleurinp.store()
return fleurinp
[docs]def get_fleurinp_from_remote_data(remote_node, store=False, additional_files=None):
"""
Create FleurinpData object from the given RemoteData object
:param remote_node: RemoteData to use for the generation of the FleurinpData
:param store: bool, if True the FleurinpData object will be stored after generation
:returns: FleurinpData object with the input xml files from the retrieved folder
of the calculation associated RemoteData
"""
for link in remote_node.get_incoming().all():
if isinstance(link.node, CalcJobNode):
parent_calc_node = link.node
retrieved = parent_calc_node.get_outgoing().get_node_by_label('retrieved')
return get_fleurinp_from_folder_data(retrieved, store=store, additional_files=additional_files)
[docs]class FleurinpData(Data):
"""
AiiDA data object representing everything a FLEUR calculation needs.
It is initialized with an absolute path to an ``inp.xml`` file or a
FolderData node containing ``inp.xml``.
Other files can also be added that will be copied to the remote machine, where the
calculation takes place.
It stores the files in the repository and stores the input parameters of the
``inp.xml`` file of FLEUR in the database as a python dictionary (as internal attributes).
When an ``inp.xml`` (name important!) file is added to files, FleurinpData searches
for a corresponding xml schema file in the PYTHONPATH environment variable.
Therefore, it is recommend to have the plug-in source code directory in the python environment.
If no corresponding schema file is found an error is raised.
FleurinpData also provides the user with
methods to extract AiiDA StructureData and
KpointsData nodes.
Remember that most attributes of AiiDA nodes can not be changed after they
have been stored in the database! Therefore, you have to use the FleurinpModifier class and its
methods if you want to change somthing in the ``inp.xml`` file. You will retrieve a new
FleurinpData that way and start a new calculation from it.
"""
# search in current folder and search in aiida source code
# we want to search in the Aiida source directory, get it from python path,
# maybe better from somewhere else.
# TODO: don not walk the whole python path, test if dir below is aiida?
# needs to be improved, schema file is often after new installation not found...
# installation with pip should always lead to a schema file in the python path, or even specific place
__version__ = '0.5.0'
# ignore machine dependent attributes in hash
_hash_ignored_attributes = [] #'_schema_file_path', '_search_paths']
[docs] def __init__(self, **kwargs):
"""
Initialize a FleurinpData object set the files given
"""
files = kwargs.pop('files', None)
for filename in files:
if 'inp.xml' in filename:
files.pop(files.index(filename))
files.append(filename)
break
node = kwargs.pop('node', None)
super().__init__(**kwargs)
if files:
if node:
self.set_files(files, node=node)
else:
self.set_files(files)
@property
def parser_info(self):
"""
Dict property, with the info and warnings from the inpxml_parser
"""
return self.get_extra('_parser_info', {})
@parser_info.setter
def parser_info(self, info_dict):
"""
Setter for has_schema
"""
self.set_extra('_parser_info', info_dict)
# files
@property
def files(self):
"""
Returns the list of the names of the files stored
"""
return self.get_attribute('files', [])
@files.setter
def files(self, filelist, node=None):
"""
Add a list of files to FleurinpData.
Alternative use setter method.
:param files: list of filepaths or filenames of node is specified
:param node: a Folder node containing files from the filelist
"""
for file1 in filelist:
self.set_file(file1, node=node)
[docs] def set_files(self, files, node=None):
"""
Add the list of files to the :class:`~aiida_fleur.data.fleurinp.FleurinpData` instance.
Can by used as an alternative to the setter.
:param files: list of abolute filepaths or filenames of node is specified
:param node: a :class:`~aiida.orm.FolderData` node containing files from the filelist
"""
for file1 in files:
self.set_file(file1, node=node)
[docs] def set_file(self, filename, dst_filename=None, node=None):
"""
Add a file to the :class:`~aiida_fleur.data.fleurinp.FleurinpData` instance.
:param filename: absolute path to the file or a filename of node is specified
:param node: a :class:`~aiida.orm.FolderData` node containing the file
"""
self._add_path(filename, dst_filename=dst_filename, node=node)
[docs] def open(self, path='inp.xml', mode='r', key=None): #pylint: disable=arguments-differ
"""
Returns an open file handle to the content of this data node.
:param key: name of the file to be opened
:param mode: the mode with which to open the file handle
:returns: A file handle in read mode
"""
if key is not None:
path = key
return super().open(path, mode=mode)
[docs] def get_content(self, filename='inp.xml'):
"""
Returns the content of the single file stored for this data node.
:returns: A string of the file content
"""
with self.open(path=filename, mode='r') as handle:
return handle.read()
[docs] def del_file(self, filename):
"""
Remove a file from FleurinpData instancefind
:param filename: name of the file to be removed from FleurinpData instance
"""
# remove from files attr list
if filename in self.get_attribute('files'):
try:
self.get_attribute('files').remove(filename)
# self._del_attribute(‘filename')
except AttributeError:
# There was no file set
pass
# remove from sandbox folder
if filename in self.list_object_names(): # get_folder_list():
self.delete_object(filename)
def _add_path(self, file1, dst_filename=None, node=None):
"""
Add a single file to the FleurinpData folder.
The destination name can be different. ``inp.xml`` is a special case.
file names are stored in the db, the whole file in the reporsitory.
:param file1: the file to be added, either string, absolute path, or filelike object whose contents to copy
Hint: Pass io.BytesIO(b"my string") to construct the file directly from a string.
:param dst_filename: string, new filename for given file in repo and db
:param node: aiida.orm.Node, usually FolderData if node is given get the 'file1' from the node
:raise: ValueError, InputValidationError
:return: None
"""
# TODO? Maybe only certain files should be allowed to be added
# contra: has to be maintained, also these files can be inputed from byte strings...
#_list_of_allowed_files = ['inp.xml', 'enpara', 'cdn1', 'sym.out', 'kpts']
if node:
if not isinstance(node, Node):
node = load_node(node) # if this fails it will raise
if file1 not in node.list_object_names():
# throw error, you try to add something that is not there
raise ValueError('file1 has to be in the specified node')
is_filelike = True
if dst_filename is None:
final_filename = file1
else:
final_filename = dst_filename
# Override file1 with bytestring of file
# since we have to use 'with', and node has no method to copy files
# we read the whole file and write it again later
# this is not so nice, but we assume that input files are rather small...
with node.open(file1, mode='rb') as file2:
file1 = io.BytesIO(file2.read())
elif isinstance(file1, six.string_types):
is_filelike = False
if not os.path.isabs(file1):
file1 = os.path.abspath(file1)
#raise ValueError("Pass an absolute path for file1: {}".format(file1))
if not os.path.isfile(file1):
raise ValueError('file1 must exist and must be a single file: {}'.format(file1))
if dst_filename is None:
final_filename = os.path.split(file1)[1]
else:
final_filename = dst_filename
else:
is_filelike = True
if dst_filename is None:
try:
final_filename = os.path.basename(file1.name) # Not sure if this still works for aiida>2.0
except AttributeError:
final_filename = 'inp.xml' # fall back to default
else:
final_filename = dst_filename
key = final_filename
old_file_list = self.list_object_names()
old_files_list = self.get_attribute('files', [])
# remove file from folder first if it exists
if final_filename not in old_file_list:
old_files_list.append(final_filename)
else:
try:
old_file_list.remove(final_filename)
except ValueError:
pass
if is_filelike:
try:
self.put_object_from_filelike(file1, key, mode='wb')
except TypeError:
self.put_object_from_filelike(file1, key)
else:
self.put_object_from_file(file1, key)
self.set_attribute('files', old_files_list) # We want to keep the other files
### Special case: 'inp.xml' ###
# here this is hardcoded, might want to change? get filename from elsewhere
if final_filename == 'inp.xml':
# get input file version number
inp_version_number = None
lines = self.get_content(filename=final_filename).split('\n')
for line in lines:
if re.search('fleurInputVersion', str(line)):
inp_version_number = re.findall(r'\d+.\d+', str(line))[0]
break
if inp_version_number is None:
raise InputValidationError('No fleurInputVersion number found '
'in given input file: {}. {}'
'Please check if this is a valid fleur input file. '
'It can not be validated and I can not use it. '
''.format(file1, lines))
self.set_attribute('inp_version', inp_version_number)
# finally set inp dict of Fleurinpdata
self._set_inp_dict()
def _set_inp_dict(self):
"""
Sets the inputxml_dict from the ``inp.xml`` file attached to FleurinpData
1. load ``inp.xml`` file
2. insert all files to include into the etree
3. call masci-tools input file parser (Validation happens inside here)
4. set inputxml_dict
"""
from masci_tools.io.parsers.fleur import inpxml_parser
#The schema_dict is not needed outside the inpxml_parser so we ignore it with the underscore
xmltree, _ = self.load_inpxml()
parser_info = {}
try:
inpxml_dict = inpxml_parser(xmltree, parser_info_out=parser_info)
except (ValueError, FileNotFoundError) as exc:
raise InputValidationError(f'inp.xml parser failed: {exc}') from exc
finally:
#Always try to provide the error/warning information
self.parser_info = parser_info
# set inpxml_dict attribute
self.set_attribute('inp_dict', inpxml_dict)
[docs] def load_inpxml(self, validate_xml_schema=True, return_included_tags=False, **kwargs):
"""
Returns the lxml etree and the schema dictionary corresponding to the version. If validate_xml_schema=True
the file will also be validated against the schema
Keyword arguments are passed on to the parser
"""
from masci_tools.io.io_fleurxml import load_inpxml
from masci_tools.util.xml.common_functions import validate_xml
self._validate()
with self.open(path='inp.xml', mode='rb') as inpxmlfile:
try:
xmltree, schema_dict = load_inpxml(inpxmlfile, **kwargs)
except ValueError as exc:
# prob inp.xml file broken
err_msg = ('The inp.xml file is probably broken, could not parse it to an xml etree.')
raise InputValidationError(err_msg) from exc
except FileNotFoundError as exc:
# prob inp.xml file broken
err_msg = ('The inp.xml file is probably broken, could not find corresponding input schema.')
raise InputValidationError(err_msg) from exc
xmltree, included_tags = self._include_files(xmltree)
develop_version = self.inp_version != schema_dict['inp_version']
if validate_xml_schema and not develop_version:
try:
validate_xml(xmltree,
schema_dict.xmlschema,
error_header='Input file is not validated against the schema')
except etree.DocumentInvalid as err:
raise InputValidationError(err) from err
elif develop_version:
self.logger.warning(f'You are using a Fleur input file with file version {self.inp_version}.\n'
'This version has no corresponding XML Schema stored in masci-tools.\n'
'Unexpected Errors can occur. If that is the case you can try to add the '
'XML Schema for this file version to masci-tools')
if return_included_tags:
return xmltree, schema_dict, included_tags
else:
return xmltree, schema_dict
def _include_files(self, xmltree):
"""
Tries to insert all .xml, which are not inp.xml file into the etree since they are
not naturally available for the parser (open vs self.open)
Creates a NamedTemporaryFile for each one and replaces the name in the etree_string
Then it is reparsed into a ElementTree and teh xi:include tags are executed
"""
from masci_tools.util.xml.common_functions import clear_xml
import tempfile
xmltree_string = etree.tostring(xmltree)
temp_files = []
for file in self.files:
if file.endswith('.xml') and file != 'inp.xml':
#Get file content from node
include_content = ''
with self.open(path=file, mode='r') as include_file:
include_content = include_file.read()
#Write content into temporary file
with tempfile.NamedTemporaryFile(mode='w', delete=False) as fo:
fo.write(include_content)
temp_files.append(fo.name)
#If the include tag for the given file is not present nothing is replaced
xmltree_string = xmltree_string.replace(bytes(file, 'utf-8'), bytes(fo.name, 'utf-8'))
#Regenerate the tree with tempfile names
xmltree_with_includes = etree.fromstring(xmltree_string).getroottree()
#Performs the inclusions and remove comments
cleared_tree, included_tags = clear_xml(xmltree_with_includes)
#Remove temporary files
for file in temp_files:
os.remove(file)
return cleared_tree, included_tags
# dict with inp paramters parsed from inp.xml
@property
def inp_dict(self):
"""
Returns the inp_dict (the representation of the ``inp.xml`` file) as it will
or is stored in the database.
"""
return self.get_attribute('inp_dict', {})
# version of the inp.xml file
@property
def inp_version(self):
"""
Returns the version string corresponding to the inp.xml file
"""
return self.get_attribute('inp_version', None)
def _validate(self):
"""
A validation method. Checks if an ``inp.xml`` file is in the FleurinpData.
"""
#from aiida.common.exceptions import ValidationError
# check if schema file path exists.
super()._validate()
if 'inp.xml' in self.files:
# has_inpxml = True # does nothing so far
pass
else:
raise ValidationError('inp.xml file not in attribute "files". '
'FleurinpData needs to have and inp.xml file!')
[docs] def get_fleur_modes(self):
"""
Analyses ``inp.xml`` file to set up a calculation mode. 'Modes' are paths a FLEUR
calculation can take, resulting in different output.
This files can be automatically addded to the retrieve_list of the calculation.
Common modes are: scf, jspin2, dos, band, pot8, lda+U, eels, ...
:return: a dictionary containing all possible modes.
"""
from masci_tools.util.xml.xml_getters import get_fleur_modes
xmltree, schema_dict = self.load_inpxml()
return get_fleur_modes(xmltree, schema_dict)
[docs] def get_nkpts(self):
"""
This routine returns the number of kpoints used in the fleur calculation
defined in this input
:returns: int with the number of kPoints
"""
from masci_tools.util.xml.xml_getters import get_nkpts
xmltree, schema_dict = self.load_inpxml()
return get_nkpts(xmltree, schema_dict)
[docs] def get_structuredata_ncf(self):
"""
This routine returns an AiiDA Structure Data type produced from the ``inp.xml``
file. not a calcfunction
:param self: a FleurinpData instance to be parsed into a StructureData
:returns: StructureData node, or None
"""
from aiida.orm import StructureData
from masci_tools.util.xml.xml_getters import get_structure_data
xmltree, schema_dict = self.load_inpxml()
atoms, cell, pbc = get_structure_data(xmltree, schema_dict)
struc = StructureData(cell=cell, pbc=pbc)
for pos, symbol in atoms:
struc.append_atom(position=pos, symbols=symbol)
# TODO DATA-DATA links are not wanted, you might want to use a cf instead
#struc.add_link_from(self, label='self.structure', link_type=LinkType.CREATE)
# label='self.structure'
# return {label : struc}
return struc
[docs] @cf
def get_structuredata(self):
"""
This routine return an AiiDA Structure Data type produced from the ``inp.xml``
file. If this was done before, it returns the existing structure data node.
This is a calcfunction and therefore keeps the provenance.
:param fleurinp: a FleurinpData instance to be parsed into a StructureData
:returns: StructureData node
"""
return self.get_structuredata_ncf()
[docs] def get_kpointsdata_ncf(self, name=None, index=None):
"""
This routine returns an AiiDA :class:`~aiida.orm.KpointsData` type produced from the
``inp.xml`` file. This only works if the kpoints are listed in the in inpxml.
This is NOT a calcfunction and does not keep the provenance!
:param name: str, optional, if given only the kpoint set with the given name
is returned
:param index: int, optional, if given only the kpoint set with the given index
is returned
:returns: :class:`~aiida.orm.KpointsData` node
"""
from aiida.orm import KpointsData
from masci_tools.util.xml.xml_getters import get_kpoints_data
# HINT, TODO:? in this routine, the 'cell' you might get in an other way
# exp: StructureData.cell, but for this you have to make a structureData Node,
# which might take more time for structures with lots of atoms.
# then just parsing the cell from the inp.xml
# as in the routine get_structureData
xmltree, schema_dict = self.load_inpxml()
if name is None and index is None:
kpoints, weights, cell, pbc = get_kpoints_data(xmltree, schema_dict)
else:
kpoints, weights, cell, pbc = get_kpoints_data(xmltree, schema_dict, name=name, index=index)
if isinstance(kpoints, dict):
kpoints_data = {}
for (label, kpoints_set), weights_set in zip(kpoints.items(), weights.values()):
kps = KpointsData()
kps.set_cell(cell)
kps.pbc = pbc
kps.set_kpoints(kpoints_set, cartesian=False, weights=weights_set)
#kpoints_data.add_link_from(self, label='fleurinp.kpts', link_type=LinkType.CREATE)
kps.label = 'fleurinp.kpts'
kpoints_data[label] = kps
else:
kpoints_data = KpointsData()
kpoints_data.set_cell(cell)
kpoints_data.pbc = pbc
kpoints_data.set_kpoints(kpoints, cartesian=False, weights=weights)
#kpoints_data.add_link_from(self, label='fleurinp.kpts', link_type=LinkType.CREATE)
kpoints_data.label = 'fleurinp.kpts'
return kpoints_data
[docs] @cf
def get_kpointsdata(self, name=None, index=None):
"""
This routine returns an AiiDA :class:`~aiida.orm.KpointsData` type produced from the
``inp.xml`` file. This only works if the kpoints are listed in the in inpxml.
This is a calcfunction and keeps the provenance!
:returns: :class:`~aiida.orm.KpointsData` node
"""
return self.get_kpointsdata_ncf(name=name, index=index)
[docs] def get_parameterdata_ncf(self, inpgen_ready=True, write_ids=True):
"""
This routine returns an AiiDA :class:`~aiida.orm.Dict` type produced from the ``inp.xml``
file. This node can be used for inpgen as `calc_parameters`.
This is NOT a calcfunction and does NOT keep the provenance!
:returns: :class:`~aiida.orm.Dict` node
"""
from aiida.orm import Dict
from masci_tools.util.xml.xml_getters import get_parameter_data
xmltree, schema_dict = self.load_inpxml()
parameter_data = get_parameter_data(xmltree, schema_dict, inpgen_ready=inpgen_ready, write_ids=write_ids)
return Dict(dict=parameter_data)
[docs] @cf
def get_parameterdata(self):
"""
This routine returns an AiiDA :class:`~aiida.orm.Dict` type produced from the ``inp.xml``
file. The returned node can be used for inpgen as `calc_parameters`.
This is a calcfunction and keeps the provenance!
:returns: :class:`~aiida.orm.Dict` node
"""
return self.get_parameterdata_ncf()
[docs] def get_tag(self, xpath):
"""
Tries to evaluate an xpath expression for ``inp.xml`` file. If it fails it logs it.
:param xpath: an xpath expression
:returns: A node list retrived using given xpath
"""
from masci_tools.util.xml.common_functions import eval_xpath
warnings.warn(
'The get_tag method is deprecated. Instead you can use the load_inpxml method to access '
'the xmltree and schema of the stored inp.xml. Then the required information can be accessed '
'via the XML functions in masci-tools or directly', DeprecationWarning)
xmltree, _ = self.load_inpxml()
root = xmltree.getroot()
return eval_xpath(root, xpath)