def test_conversion_multiindex(self):
d = {'comp_str': ["Fe2", "MnO2"]}
df_1lvl = DataFrame(data=d)
df_1lvl = StrToComposition().featurize_dataframe(df_1lvl,
'comp_str',
multiindex=True)
self.assertEqual(
df_1lvl[("StrToComposition", "composition")].tolist(),
[Composition("Fe2"), Composition("MnO2")])
df_2lvl = DataFrame(data=d)
df_2lvl.columns = MultiIndex.from_product(
(["custom"], df_2lvl.columns.values))
df_2lvl = StrToComposition().featurize_dataframe(
df_2lvl, ("custom", "comp_str"), multiindex=True)
self.assertEqual(
df_2lvl[("StrToComposition", "composition")].tolist(),
[Composition("Fe2"), Composition("MnO2")])
df_2lvl = DataFrame(data=d)
df_2lvl.columns = MultiIndex.from_product(
(["custom"], df_2lvl.columns.values))
sto = StrToComposition(target_col_id='test')
df_2lvl = sto.featurize_dataframe(df_2lvl, ("custom", "comp_str"),
multiindex=True)
self.assertEqual(
df_2lvl[("StrToComposition", "test")].tolist(),
[Composition("Fe2"), Composition("MnO2")])
# if two level multiindex provided as target, it should be written there
# here we test converting multiindex in place
df_2lvl = DataFrame(data=d)
df_2lvl.columns = MultiIndex.from_product(
(["custom"], df_2lvl.columns.values))
sto = StrToComposition(target_col_id=None, overwrite_data=True)
df_2lvl = sto.featurize_dataframe(df_2lvl, ("custom", "comp_str"),
multiindex=True)
self.assertEqual(
df_2lvl[("custom",
"comp_str")][df_2lvl[("custom",
"comp_str")].columns[3]].tolist(),
[Composition("Fe2"), Composition("MnO2")])
# Try inplace multiindex conversion with return errors
df_2lvl = DataFrame(data=d)
df_2lvl.columns = MultiIndex.from_product(
(["custom"], df_2lvl.columns.values))
sto = StrToComposition(target_col_id=None, overwrite_data=True)
df_2lvl = sto.featurize_dataframe(df_2lvl, ("custom", "comp_str"),
multiindex=True,
return_errors=True,
ignore_errors=True)
self.assertTrue(
all(df_2lvl[("custom", "StrToComposition Exceptions")].isnull()))
def has_polymorphs(self):
"""Determine if a task's raw data contains polymorphs.
Returns:
(bool) If true, contains polymorphs.
"""
checker_key = "pmg_composition"
self._check_is_loaded()
if self.metadata.input_type == "composition":
stc = StrToComposition(target_col_id=checker_key, reduce=True)
comps = stc.featurize_dataframe(self.df,
"composition")[checker_key].values
elif self.metadata.input_type == "structure":
stc = StructureToComposition(target_col_id=checker_key,
reduce=True)
comps = stc.featurize_dataframe(self.df,
"structure")[checker_key].values
else:
raise ValueError(
"Cannot check for polymorphs without input type in "
"(structure, composition)!")
unique_comps = set(comps)
if len(unique_comps) != len(comps):
return True
else:
return False
def test_conversion_multiindex(self):
d = {'comp_str': ["Fe2", "MnO2"]}
df_1lvl = DataFrame(data=d)
df_1lvl = StrToComposition().featurize_dataframe(
df_1lvl, 'comp_str', multiindex=True)
self.assertEqual(df_1lvl[("StrToComposition", "composition")].tolist(),
[Composition("Fe2"), Composition("MnO2")])
df_2lvl = DataFrame(data=d)
df_2lvl.columns = MultiIndex.from_product((["custom"],
df_2lvl.columns.values))
df_2lvl = StrToComposition().featurize_dataframe(
df_2lvl, ("custom", "comp_str"), multiindex=True)
self.assertEqual(df_2lvl[("StrToComposition", "composition")].tolist(),
[Composition("Fe2"), Composition("MnO2")])
df_2lvl = DataFrame(data=d)
df_2lvl.columns = MultiIndex.from_product((["custom"],
df_2lvl.columns.values))
sto = StrToComposition(target_col_id='test')
df_2lvl = sto.featurize_dataframe(
df_2lvl, ("custom", "comp_str"), multiindex=True)
self.assertEqual(df_2lvl[("StrToComposition", "test")].tolist(),
[Composition("Fe2"), Composition("MnO2")])
# if two level multiindex provided as target, it should be written there
# here we test converting multiindex in place
df_2lvl = DataFrame(data=d)
df_2lvl.columns = MultiIndex.from_product((["custom"],
df_2lvl.columns.values))
sto = StrToComposition(target_col_id=None, overwrite_data=True)
df_2lvl = sto.featurize_dataframe(
df_2lvl, ("custom", "comp_str"), multiindex=True, inplace=False)
self.assertEqual(df_2lvl[("custom", "comp_str")].tolist(),
[Composition("Fe2"), Composition("MnO2")])
# Try inplace multiindex conversion with return errors
df_2lvl = DataFrame(data=d)
df_2lvl.columns = MultiIndex.from_product((["custom"],
df_2lvl.columns.values))
sto = StrToComposition(target_col_id=None, overwrite_data=True)
df_2lvl = sto.featurize_dataframe(
df_2lvl, ("custom", "comp_str"), multiindex=True,
return_errors=True, ignore_errors=True)
self.assertTrue(
all(df_2lvl[("custom", "StrToComposition Exceptions")].isnull()))
class FeatureGenerator:
"""
A wraper class to generate multiple type of elemental features
"""
def __init__(self):
self.feature_calculators = MultipleFeaturizer([
cf.ElementProperty.from_preset(preset_name="magpie"),
cf.Stoichiometry(),
cf.ValenceOrbital(props=['frac']),
cf.IonProperty(fast=True),
cf.BandCenter(),
cf.ElementFraction(),
])
self.str2composition = StrToComposition()
def generate(self, df: pd.DataFrame, ignore_errors: bool = False):
"""
generate feature from a dataframe with a "formula" column that contains
chemical formulas of the compositions.
"""
df = self.str2composition.featurize_dataframe(
df, "formula", ignore_errors=ignore_errors)
df = df.dropna()
df = self.feature_calculators.featurize_dataframe(
df, col_id='composition', ignore_errors=ignore_errors)
df["NComp"] = df["composition"].apply(len)
return df
def test_conversion_overwrite(self):
# Test with overwrite
d = {'comp_str': ["Fe2", "MnO2"]}
df = DataFrame(data=d)
stc = StrToComposition(target_col_id='comp_str', overwrite_data=False)
with self.assertRaises(ValueError):
df = stc.featurize_dataframe(df, 'comp_str', inplace=True)
with self.assertRaises(ValueError):
df = stc.featurize_dataframe(df, 'comp_str', inplace=False)
stc = StrToComposition(target_col_id='comp_str', overwrite_data=True)
dfres_ipt = df.copy()
stc.featurize_dataframe(dfres_ipt, 'comp_str', inplace=True)
self.assertListEqual(dfres_ipt.columns.tolist(), ["comp_str"])
dfres_ipf = stc.featurize_dataframe(df, 'comp_str', inplace=False)
self.assertListEqual(dfres_ipf.columns.tolist(), ["comp_str"])
def test_str_to_composition(self):
d = {'comp_str': ["Fe2", "MnO2"]}
df = DataFrame(data=d)
df = StrToComposition().featurize_dataframe(df, 'comp_str')
self.assertEqual(df["composition"].tolist(),
[Composition("Fe2"), Composition("MnO2")])
stc = StrToComposition(reduce=True, target_col_id='composition_red')
df = stc.featurize_dataframe(df, 'comp_str')
self.assertEqual(df["composition_red"].tolist(),
[Composition("Fe"), Composition("MnO2")])
class FeatureGenerator:
"""
A wraper class to generate multiple type of elemental features
"""
def __init__(self):
self.feature_calculators = MultipleFeaturizer([
cf.ElementProperty.from_preset(preset_name="magpie"),
cf.Stoichiometry(),
cf.ValenceOrbital(props=['frac']),
cf.IonProperty(fast=True),
cf.BandCenter(),
cf.ElementFraction(),
])
self.str2composition = StrToComposition()
def generate(self,
df: pd.DataFrame,
ignore_errors: bool = False,
drop_mode=True):
"""
generate feature from a dataframe with a "formula" column that contains
chemical formulas of the compositions.
df : a dataframe with a column name formula
ignore_errors : ignore errors when generating features
drop_mode : drop property that generated from mode aggregation function
"""
df = self.str2composition.featurize_dataframe(
df, "formula", ignore_errors=ignore_errors)
df = df.dropna()
df = self.feature_calculators.featurize_dataframe(
df, col_id='composition', ignore_errors=ignore_errors)
df["NComp"] = df["composition"].apply(len)
if drop_mode:
df = df.drop(columns=[
c for c in df.columns if "mode" in c and c.startswith("Magpie")
])
return df
def _tidy_column(self, df, featurizer_type):
"""
Various conversions to homogenize columns for featurization input.
For example, take a column of compositions and ensure they are decorated
with oxidation states, are not strings, etc.
Args:
df (pandas.DataFrame)
featurizer_type: The key defining the featurizer input. For example,
composition featurizers should have featurizer_type of
"composition".
Returns:
df (pandas.DataFrame): DataFrame with featurizer_type column
ready for featurization.
"""
# todo: Make the following conversions more robust (no [0] type checking)
type_tester = df[featurizer_type].iloc[0]
if featurizer_type == self.composition_col:
# Convert formulas to composition objects
if isinstance(type_tester, str):
self.logger.info(
self._log_prefix +
"Compositions detected as strings. Attempting "
"conversion to Composition objects...")
stc = StrToComposition(overwrite_data=True,
target_col_id=featurizer_type)
df = stc.featurize_dataframe(df,
featurizer_type,
multiindex=self.multiindex,
ignore_errors=True,
inplace=False)
elif isinstance(type_tester, dict):
self.logger.info(self._log_prefix +
"Compositions detected as dicts. Attempting "
"conversion to Composition objects...")
df[featurizer_type] = [
Composition.from_dict(d) for d in df[featurizer_type]
]
# Convert non-oxidstate containing comps to oxidstate comps
if self.guess_oxistates:
self.logger.info(
self._log_prefix +
"Guessing oxidation states of compositions, as"
" they were not present in input.")
cto = CompositionToOxidComposition(
target_col_id=featurizer_type,
overwrite_data=True,
return_original_on_error=True,
max_sites=-50)
try:
df = cto.featurize_dataframe(df,
featurizer_type,
multiindex=self.multiindex,
inplace=False)
except Exception as e:
self.logger.info(self._log_prefix +
"Could not decorate oxidation states due "
"to {}. Excluding featurizers based on "
"composition oxistates".format(e))
classes_require_oxi = [
c.__class__.__name__
for c in CompositionFeaturizers().need_oxi
]
self.exclude.extend(classes_require_oxi)
else:
# Convert structure/bs/dos dicts to objects (robust already)
if isinstance(type_tester, (dict, str)):
self.logger.info(self._log_prefix.capitalize() +
"{} detected as string or dict. Attempting "
"conversion to {} objects..."
"".format(featurizer_type, featurizer_type))
if isinstance(type_tester, str):
raise ValueError("{} column is type {}. Cannot convert."
"".format(featurizer_type,
type(type_tester)))
dto = DictToObject(overwrite_data=True,
target_col_id=featurizer_type)
df = dto.featurize_dataframe(df,
featurizer_type,
inplace=False)
# Decorate with oxidstates
if featurizer_type == self.structure_col and \
self.guess_oxistates:
self.logger.info(
self._log_prefix +
"Guessing oxidation states of structures if they were "
"not present in input.")
sto = StructureToOxidStructure(
target_col_id=featurizer_type,
overwrite_data=True,
return_original_on_error=True,
max_sites=-50)
try:
df = sto.featurize_dataframe(
df,
featurizer_type,
multiindex=self.multiindex,
inplace=False)
except Exception as e:
self.logger.info(
self._log_prefix +
"Could not decorate oxidation states on structures "
"due to {}.".format(e))
return df
"""
['_id', 'material_id', 'formula', 'nsites', 'space_group', 'volume',
'structure', 'elastic_anisotropy', 'G_Reuss', 'G_VRH', 'G_Voigt',
'K_Reuss', 'K_VRH', 'K_Voigt', 'poisson_ratio', 'compliance_tensor',
'elastic_tensor', 'elastic_tensor_original', 'cif', 'kpoint_density',
'poscar']
"""
unwanted_columns = ['_id', 'material_id', 'nsites', 'volume',
'cif', 'kpoint_density', 'poscar']
df = df.drop(unwanted_columns, axis=1)
from matminer.featurizers.conversions import StrToComposition
sc_feat = StrToComposition()
df = sc_feat.featurize_dataframe(df, col_id='formula')
from matminer.featurizers.composition import ElementProperty
ep_feat = ElementProperty.from_preset(preset_name='magpie')
df = ep_feat.featurize_dataframe(df, col_id='composition')
from matminer.featurizers.conversions import CompositionToOxidComposition
co_feat = CompositionToOxidComposition()
df = co_feat.featurize_dataframe(df, col_id='composition')
from matminer.featurizers.composition import OxidationStates
os_feat = OxidationStates()
df = os_feat.featurize_dataframe(df, col_id='composition_oxid')
