def _prepare_group_for_upload(cls, group_name, group_description=None, dry_run=False): """Prepare a (possibly new) group to upload a POTCAR family to.""" if not dry_run: group, group_created = Group.get_or_create( name=group_name, type_string=cls.potcar_family_type_string) else: group = cls.get_potcar_group(group_name) group_created = bool(not group) if not group: group = Group(name=group_name) if group.user.pk != get_current_user().pk: raise UniquenessError( 'There is already a POTCAR family group with name {}, but it belongs to user {}, therefore you cannot modify it' .format(group_name, group.user.email)) if group_description: group.description = group_description elif group_created: raise ValueError( 'A new POTCAR family {} should be created but no description was given!' .format(group_name)) return group
def upload_basisset_family(cls, folder, group_name, group_description, stop_if_existing=True, extension=".basis", dry_run=False): """ Upload a set of Basis Set files in a given group. :param folder: a path containing all Basis Set files to be added. Only files ending in the set extension (case-insensitive) are considered. :param group_name: the name of the group to create. If it exists and is non-empty, a UniquenessError is raised. :param group_description: a string to be set as the group description. Overwrites previous descriptions, if the group was existing. :param stop_if_existing: if True, check for the md5 of the files and, if the file already exists in the DB, raises a MultipleObjectsError. If False, simply adds the existing BasisSetData node to the group. :param extension: the filename extension to look for :param dry_run: If True, do not change the database. """ from aiida.common import aiidalogger from aiida.orm import Group from aiida.common.exceptions import UniquenessError, NotExistent from aiida_crystal17.aiida_compatability import get_automatic_user automatic_user = get_automatic_user() if not os.path.isdir(folder): raise ValueError("folder must be a directory") # only files, and only those ending with specified exension; # go to the real file if it is a symlink files = [ os.path.realpath(os.path.join(folder, i)) for i in os.listdir(folder) if os.path.isfile(os.path.join(folder, i)) and i.lower().endswith(extension) ] nfiles = len(files) try: group = Group.get(name=group_name, type_string=BASISGROUP_TYPE) group_created = False except NotExistent: group = Group(name=group_name, type_string=BASISGROUP_TYPE, user=automatic_user) group_created = True if group.user.email != automatic_user.email: raise UniquenessError( "There is already a BasisFamily group with name {}" ", but it belongs to user {}, therefore you " "cannot modify it".format(group_name, group.user.email)) # Always update description, even if the group already existed group.description = group_description # NOTE: GROUP SAVED ONLY AFTER CHECKS OF UNICITY basis_and_created = _retrieve_basis_sets(files, stop_if_existing) # check whether basisset are unique per element elements = [(i[0].element, i[0].md5sum) for i in basis_and_created] # If group already exists, check also that I am not inserting more than # once the same element if not group_created: for aiida_n in group.nodes: # Skip non-basis sets if not isinstance(aiida_n, BasisSetData): continue elements.append((aiida_n.element, aiida_n.md5sum)) elements = set( elements) # Discard elements with the same MD5, that would # not be stored twice elements_names = [e[0] for e in elements] if not len(elements_names) == len(set(elements_names)): duplicates = set( [x for x in elements_names if elements_names.count(x) > 1]) duplicates_string = ", ".join(i for i in duplicates) raise UniquenessError( "More than one Basis found for the elements: " + duplicates_string + ".") # At this point, save the group, if still unstored if group_created and not dry_run: group.store() # save the basis set in the database, and add them to group for basisset, created in basis_and_created: if created: if not dry_run: basisset.store() aiidalogger.debug("New node {0} created for file {1}".format( # pylint: disable=logging-format-interpolation basisset.uuid, basisset.filename)) else: aiidalogger.debug("Reusing node {0} for file {1}".format( # pylint: disable=logging-format-interpolation basisset.uuid, basisset.filename)) # Add elements to the group all together if not dry_run: group.add_nodes(basis for basis, created in basis_and_created) nuploaded = len([_ for _, created in basis_and_created if created]) return nfiles, nuploaded
def upload_psf_family(folder, group_name, group_description, stop_if_existing=True): """ Upload a set of PSF files in a given group. :param folder: a path containing all PSF files to be added. Only files ending in .PSF (case-insensitive) are considered. :param group_name: the name of the group to create. If it exists and is non-empty, a UniquenessError is raised. :param group_description: a string to be set as the group description. Overwrites previous descriptions, if the group was existing. :param stop_if_existing: if True, check for the md5 of the files and, if the file already exists in the DB, raises a MultipleObjectsError. If False, simply adds the existing PsfData node to the group. """ import os import aiida.common from aiida.common import aiidalogger from aiida.orm import Group from aiida.common.exceptions import UniquenessError, NotExistent from aiida.backends.utils import get_automatic_user from aiida.orm.querybuilder import QueryBuilder if not os.path.isdir(folder): raise ValueError("folder must be a directory") # only files, and only those ending with .psf or .PSF; # go to the real file if it is a symlink files = [ os.path.realpath(os.path.join(folder, i)) for i in os.listdir(folder) if os.path.isfile(os.path.join(folder, i)) and i.lower().endswith('.psf') ] nfiles = len(files) try: group = Group.get(name=group_name, type_string=PSFGROUP_TYPE) group_created = False except NotExistent: group = Group(name=group_name, type_string=PSFGROUP_TYPE, user=get_automatic_user()) group_created = True if group.user != get_automatic_user(): raise UniquenessError("There is already a PsfFamily group with name {}" ", but it belongs to user {}, therefore you " "cannot modify it".format( group_name, group.user.email)) # Always update description, even if the group already existed group.description = group_description # NOTE: GROUP SAVED ONLY AFTER CHECKS OF UNICITY pseudo_and_created = [] for f in files: md5sum = aiida.common.utils.md5_file(f) qb = QueryBuilder() qb.append(PsfData, filters={'attributes.md5': {'==': md5sum}}) existing_psf = qb.first() #existing_psf = PsfData.query(dbattributes__key="md5", # dbattributes__tval = md5sum) if existing_psf is None: # return the psfdata instances, not stored pseudo, created = PsfData.get_or_create(f, use_first=True, store_psf=False) # to check whether only one psf per element exists # NOTE: actually, created has the meaning of "to_be_created" pseudo_and_created.append((pseudo, created)) else: if stop_if_existing: raise ValueError("A PSF with identical MD5 to " " {} cannot be added with stop_if_existing" "".format(f)) existing_psf = existing_psf[0] pseudo_and_created.append((existing_psf, False)) # check whether pseudo are unique per element elements = [(i[0].element, i[0].md5sum) for i in pseudo_and_created] # If group already exists, check also that I am not inserting more than # once the same element if not group_created: for aiida_n in group.nodes: # Skip non-pseudos if not isinstance(aiida_n, PsfData): continue elements.append((aiida_n.element, aiida_n.md5sum)) elements = set(elements) # Discard elements with the same MD5, that would # not be stored twice elements_names = [e[0] for e in elements] if not len(elements_names) == len(set(elements_names)): duplicates = set( [x for x in elements_names if elements_names.count(x) > 1]) duplicates_string = ", ".join(i for i in duplicates) raise UniquenessError("More than one PSF found for the elements: " + duplicates_string + ".") # At this point, save the group, if still unstored if group_created: group.store() # save the psf in the database, and add them to group for pseudo, created in pseudo_and_created: if created: pseudo.store() aiidalogger.debug("New node {} created for file {}".format( pseudo.uuid, pseudo.filename)) else: aiidalogger.debug("Reusing node {} for file {}".format( pseudo.uuid, pseudo.filename)) # Add elements to the group all togetehr group.add_nodes(pseudo for pseudo, created in pseudo_and_created) nuploaded = len([_ for _, created in pseudo_and_created if created]) return nfiles, nuploaded