def add_structure(self, source, name=None, identifier=None, fmt=None):
"""add a structure to the mpfile"""
from pymatgen import Structure, MPRester
if isinstance(source, Structure):
structure = source
elif isinstance(source, dict):
structure = Structure.from_dict(source)
elif os.path.exists(source):
structure = Structure.from_file(source, sort=True)
elif isinstance(source, six.string_types):
if fmt is None:
raise ValueError('Need fmt to get structure from string!')
structure = Structure.from_str(source, fmt, sort=True)
else:
raise ValueError(source, 'not supported!')
if name is not None:
if not isinstance(name, six.string_types):
raise ValueError('structure name needs to be a string')
elif '.' in name:
raise ValueError('structure name cannot contain dots (.)')
mpr = MPRester()
if not mpr.api_key:
raise ValueError(
'API key not set. Run `pmg config --add PMG_MAPI_KEY <USER_API_KEY>`.'
)
matched_mpids = mpr.find_structure(structure)
formula = get_composition_from_string(structure.composition.formula)
if not matched_mpids:
if identifier is None:
identifier = formula
print(
'Structure not found in MP! Please submit via MPComplete to '
'obtain mp-id or manually choose an anchor mp-id! Continuing '
'with {} as identifier!'.format(identifier))
else:
print('Structure not found in MP! Forcing {} as identifier!'.
format(identifier))
elif identifier is None:
identifier = matched_mpids[0]
if len(matched_mpids) > 1:
print('Multiple matching structures found in MP. Using',
identifier)
elif identifier not in matched_mpids:
raise ValueError(
'Structure does not match {} but instead {}'.format(
identifier, matched_mpids))
idx = len(
self.document.get(identifier, {}).get(mp_level01_titles[3], {}))
sub_key = formula if name is None else name
if sub_key in self.document.get(identifier,
{}).get(mp_level01_titles[3], {}):
sub_key += '_{}'.format(idx)
self.document.rec_update(
nest_dict(structure.as_dict(),
[identifier, mp_level01_titles[3], sub_key]))
return identifier
def concat(self, mpfile):
"""concatenate single-section MPFile with this MPFile"""
try:
if len(mpfile.document) > 1:
raise ValueError(
'concatenation only possible with single section files')
except AttributeError:
raise ValueError('Provide a MPFile to concatenate')
mp_cat_id = mpfile.document.keys()[0]
general_title = mp_level01_titles[0]
if general_title in mpfile.document[mp_cat_id]:
general_data = mpfile.document[mp_cat_id].pop(general_title)
if general_title not in self.document:
self.document.rec_update(
nest_dict(general_data, [general_title]))
self.document.rec_update(
nest_dict(mpfile.document.pop(mp_cat_id),
[self.get_unique_mp_cat_id(mp_cat_id)]))
def add_data_table(self, identifier, dataframe, name):
"""add a datatable to the root-level section
Args:
identifier (str): MP category ID (`mp_cat_id`)
dataframe (pandas.DataFrame): tabular data as Pandas DataFrame
name (str): table name, optional if only one table in section
"""
# TODO: optional table name, required if multiple tables per root-level section
self.document.rec_update(
nest_dict(pandas_to_dict(dataframe), [identifier, name]))
def parse(self, file_string):
"""recursively parse sections according to number of separators"""
# split into section title line (even) and section body (odd entries)
sections = re.split(self.separator_regex(), file_string)
if len(sections) > 1:
# check for preceding bare section_body (without section title), and parse
if sections[0] != '': self.parse(sections[0])
# drop preceding bare section_body
sections = sections[1:] # https://docs.python.org/2/library/re.html#re.split
for section_index,section_body in enumerate(sections[1::2]):
clean_title = self.clean_title(sections[2*section_index])
# uniquify level-0 titles if necessary
if self.level == min_indent_level and clean_title in self.document:
clean_title += '--%d' % self.level0_counter
self.level0_counter += 1
self.increase_level(clean_title)
self.parse(section_body)
self.reduce_level()
else:
# separator level not found, convert section body to pandas object,
section_title = self.section_titles[-1]
is_data_section, pd_obj = self.read_csv(section_title, file_string)
logging.info(pd_obj)
# TODO: include validation
# use first csv table for default plot, first column as x-column
if is_data_section and mp_level01_titles[2] not in \
self.document[self.section_titles[0]]:
self.document.rec_update(nest_dict(
{'x': pd_obj.columns[0], 'table': section_title},
[self.section_titles[0], mp_level01_titles[2], 'default']
))
# add data section title to nest 'bare' data under data section
# => artificially increase and decrease level (see below)
is_bare_data = (is_data_section and self.is_bare_section(section_title))
if is_bare_data: self.increase_level(mp_level01_titles[1])
# update nested dict/document based on section level
self.document.rec_update(nest_dict(
self.to_dict(pd_obj), self.section_titles
))
if is_bare_data: self.reduce_level()
def insert_general_section(self, general_mpfile):
"""insert general section from `general_mpfile` into this MPFile"""
if general_mpfile is None: return
general_title = mp_level01_titles[0]
general_data = general_mpfile.document[general_title]
root_key = self.document.keys()[0]
first_subkey = self.document[root_key].keys()[0]
for key, value in general_data.items():
if key in self.document[root_key]:
self.document.rec_update(nest_dict(value, [root_key, key]))
else:
self.document[root_key].insert_before(first_subkey,
(key, value))
def concat(self, mpfile, uniquify=True):
"""concatenate single-section MPFile with this MPFile"""
try:
if len(mpfile.document) > 1:
raise ValueError(
'concatenation only possible with single section files')
except AttributeError:
raise ValueError('Provide a MPFile to concatenate')
mp_cat_id = mpfile.document.keys()[0]
general_title = mp_level01_titles[0]
if general_title in mpfile.document[mp_cat_id]:
general_data = mpfile.document[mp_cat_id].pop(general_title)
if general_title not in self.document:
self.document.rec_update(
nest_dict(general_data, [general_title]))
mp_cat_id_idx, mp_cat_id_uniq = 0, mp_cat_id
if uniquify:
while mp_cat_id_uniq in self.document.keys():
mp_cat_id_uniq = mp_cat_id + '--{}'.format(mp_cat_id_idx)
mp_cat_id_idx += 1
self.document.rec_update(
nest_dict(mpfile.document.pop(mp_cat_id), [mp_cat_id_uniq]))
def add_data_table(self, identifier, dataframe, name, plot_options=None):
"""add a datatable to the root-level section
Args:
identifier (str): MP category ID (`mp_cat_id`)
dataframe (pandas.DataFrame): tabular data as Pandas DataFrame
name (str): table name, optional if only one table in section
plot_options (dict): options for according plotly graph
"""
# TODO: optional table name, required if multiple tables per root-level section
table_start = mp_level01_titles[1] + '_'
if not name.startswith(table_start):
name = table_start + name
name = ''.join([replacements.get(c, c) for c in name])
self.document.rec_update(
nest_dict(Table(dataframe).to_dict(), [identifier, name]))
self.document[identifier].insert_default_plot_options(
dataframe, name, update_plot_options=plot_options)
def insert_general_section(self, general_mpfile):
"""insert general section from `general_mpfile` into this MPFile"""
if general_mpfile is None: return
general_title = mp_level01_titles[0]
general_data = general_mpfile.document[general_title]
root_key = self.document.keys()[0]
# need to reverse-loop to keep the order of the general_mpfile
for key, value in reversed(general_data.items()):
if key in self.document[root_key]:
self.document.rec_update(nest_dict(value, [root_key, key]))
else:
# this approach is due to the order sensitivity of key-value pairs
# before or after a `>>>..` row in the custom format (legacy)
# => ignoring it here would generate the wrong MPFile in get_string
for k, v in self.document[root_key].iteritems():
if isinstance(v, dict):
self.document[root_key].insert_before(k, (key, value))
break
def split(self):
general_mpfile = self.pop_first_section() \
if mp_level01_titles[0] in self.document.keys() else None
if not self.document:
raise ValueError('No contributions in MPFile! Either the file is'
' empty or only contains shared (meta-)data not'
' correlated to core identifier.')
while True:
try:
mpfile_single = self.pop_first_section()
mpid_orig = mpfile_single.ids[0]
mpid = mpid_orig.split('--')[0]
mpfile_single.document.rec_update(
nest_dict(mpfile_single.document.pop(mpid_orig), [mpid]))
if general_mpfile is not None:
mpfile_single.insert_general_section(general_mpfile)
yield mpfile_single
except KeyError:
break
def add_structure(self, source, name=None, identifier=None, fmt=None):
"""add a structure to the mpfile"""
if isinstance(source, Structure):
structure = source
elif isinstance(source, dict):
structure = Structure.from_dict(source)
elif os.path.exists(source):
structure = Structure.from_file(source)
elif isinstance(source, six.string_types):
if fmt is None:
raise ValueError('Need fmt to get structure from string!')
structure = Structure.from_str(source, fmt)
else:
raise ValueError(source, 'not supported!')
mpr = MPRester()
if not mpr.api_key:
raise ValueError(
'API key not set. Run `pmg config --add PMG_MAPI_KEY <USER_API_KEY>`.'
)
matched_mpids = mpr.find_structure(structure)
if not matched_mpids:
raise ValueError(
'Structure not found in MP. Please submit via MPComplete to obtain mp-id!'
)
elif identifier is None:
identifier = matched_mpids[0]
if len(matched_mpids) > 1:
print('Multiple matching structures found in MP. Using',
identifier)
elif identifier not in matched_mpids:
raise ValueError(
'Structure does not match {} but instead {}'.format(
identifier, matched_mpids))
idx = len(
self.document.get(identifier, {}).get(mp_level01_titles[3], {}))
sub_key = 's{}'.format(idx) if name is None else name
self.document.rec_update(
nest_dict(structure.as_dict(),
[identifier, mp_level01_titles[3], sub_key]))
def add_hierarchical_data(self, dct, identifier=mp_level01_titles[0]):
if len(self.ids) >= self.max_contribs:
raise StopIteration(
'Reached max. number of contributions in MPFile')
self.document.rec_update(nest_dict(RecursiveDict(dct), [identifier]))
def add_hierarchical_data(self, identifier, dct):
self.document.rec_update(nest_dict(dct, [identifier]))