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:
msg = "Structure does not match {} but instead {}!".format(
identifier, matched_mpids)
raise ValueError(msg)
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 test(self, structure):
failures = []
if self.is_valid:
if not structure.is_valid():
failures.append("IS_VALID=False")
if self.potcar_exists:
elements = structure.composition.elements
if set(elements).intersection(set(self.NO_POTCARS)):
failures.append("POTCAR_EXISTS=False")
if self.max_natoms:
if structure.num_sites > self.max_natoms:
failures.append("MAX_NATOMS=Exceeded")
if self.is_ordered:
if not structure.is_ordered:
failures.append("IS_ORDERED=False")
if self.not_in_MP:
mpr = MPRester(self.MAPI_KEY)
mpids = mpr.find_structure(structure)
if mpids:
if self.require_bandstructure:
for mpid in mpids:
try:
bs = mpr.get_bandstructure_by_material_id(mpid)
if bs:
failures.append("NOT_IN_MP=False ({})".format(mpid))
except:
pass
else:
failures.append("NOT_IN_MP=False ({})".format(mpids[0]))
return True if not failures else False
def test(self, structure):
failures = []
if self.is_valid:
if not structure.is_valid():
failures.append("IS_VALID=False")
if self.potcar_exists:
elements = structure.composition.elements
if set(elements).intersection(set(self.NO_POTCARS)):
failures.append("POTCAR_EXISTS=False")
if self.max_natoms:
if structure.num_sites > self.max_natoms:
failures.append("MAX_NATOMS=Exceeded")
if self.is_ordered:
if not structure.is_ordered:
failures.append("IS_ORDERED=False")
if self.not_in_MP:
mpr = MPRester(self.MAPI_KEY)
mpids = mpr.find_structure(structure)
if mpids:
if self.require_bandstructure:
for mpid in mpids:
try:
bs = mpr.get_bandstructure_by_material_id(mpid)
if bs:
failures.append("NOT_IN_MP=False ({})".format(mpid))
except:
pass
else:
failures.append("NOT_IN_MP=False ({})".format(mpids[0]))
return True if not failures else False
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]))
class MaterialsEhullBuilder(AbstractBuilder):
def __init__(self, materials_write, mapi_key=None, update_all=False):
"""
Starting with an existing materials collection, adds stability information and
The Materials Project ID.
Args:
materials_write: mongodb collection for materials (write access needed)
mapi_key: (str) Materials API key (if MAPI_KEY env. var. not set)
update_all: (bool) - if true, updates all docs. If false, only updates
docs w/o a stability key
"""
self._materials = materials_write
self.mpr = MPRester(api_key=mapi_key)
self.update_all = update_all
def run(self):
logger.info("MaterialsEhullBuilder starting...")
self._build_indexes()
q = {"thermo.energy": {"$exists": True}}
if not self.update_all:
q["stability"] = {"$exists": False}
mats = [
m for m in self._materials.find(
q, {
"calc_settings": 1,
"structure": 1,
"thermo.energy": 1,
"material_id": 1
})
]
pbar = tqdm(mats)
for m in pbar:
pbar.set_description("Processing materials_id: {}".format(
m['material_id']))
try:
params = {}
for x in ["is_hubbard", "hubbards", "potcar_spec"]:
params[x] = m["calc_settings"][x]
structure = Structure.from_dict(m["structure"])
energy = m["thermo"]["energy"]
my_entry = ComputedEntry(structure.composition,
energy,
parameters=params)
# TODO: @computron This only calculates Ehull with respect to Materials Project.
# It should also account for the current database's results. -computron
self._materials.update_one({"material_id": m["material_id"]}, {
"$set": {
"stability": self.mpr.get_stability([my_entry])[0]
}
})
# TODO: @computron: also add additional properties like inverse hull energy?
# TODO: @computron it's better to use PD tool or reaction energy calculator
# Otherwise the compatibility schemes might have issues...one strategy might be
# use MP only to retrieve entries but compute the PD locally -computron
for el, elx in my_entry.composition.items():
entries = self.mpr.get_entries(el.symbol,
compatible_only=True)
min_e = min(
entries,
key=lambda x: x.energy_per_atom).energy_per_atom
energy -= elx * min_e
self._materials.update_one({"material_id": m["material_id"]}, {
"$set": {
"thermo.formation_energy_per_atom":
energy / structure.num_sites
}
})
mpids = self.mpr.find_structure(structure)
self._materials.update_one({"material_id": m["material_id"]},
{"$set": {
"mpids": mpids
}})
except:
import traceback
logger.exception("<---")
logger.exception(
"There was an error processing material_id: {}".format(m))
logger.exception(traceback.format_exc())
logger.exception("--->")
logger.info("MaterialsEhullBuilder finished processing.")
def reset(self):
logger.info("Resetting MaterialsEhullBuilder")
self._materials.update_many({}, {"$unset": {"stability": 1}})
self._build_indexes()
logger.info("Finished resetting MaterialsEhullBuilder")
def _build_indexes(self):
self._materials.create_index("stability.e_above_hull")
@classmethod
def from_file(cls, db_file, m="materials", **kwargs):
"""
Get a MaterialsEhullBuilder using only a db file
Args:
db_file: (str) path to db file
m: (str) name of "materials" collection
**kwargs: other parameters to feed into the builder, e.g. mapi_key
"""
db_write = get_database(db_file, admin=True)
return cls(db_write[m], **kwargs)
class MaterialsEhullBuilder(AbstractBuilder):
def __init__(self, materials_write, mapi_key=None, update_all=False):
"""
Starting with an existing materials collection, adds stability information and
The Materials Project ID.
Args:
materials_write: mongodb collection for materials (write access needed)
mapi_key: (str) Materials API key (if MAPI_KEY env. var. not set)
update_all: (bool) - if true, updates all docs. If false, only updates
docs w/o a stability key
"""
self._materials = materials_write
self.mpr = MPRester(api_key=mapi_key)
self.update_all = update_all
def run(self):
print("MaterialsEhullBuilder starting...")
self._build_indexes()
q = {"thermo.energy": {"$exists": True}}
if not self.update_all:
q["stability"] = {"$exists": False}
mats = [
m for m in self._materials.find(
q, {
"calc_settings": 1,
"structure": 1,
"thermo.energy": 1,
"material_id": 1
})
]
pbar = tqdm(mats)
for m in pbar:
pbar.set_description("Processing materials_id: {}".format(
m['material_id']))
try:
params = {}
for x in ["is_hubbard", "hubbards", "potcar_spec"]:
params[x] = m["calc_settings"][x]
structure = Structure.from_dict(m["structure"])
energy = m["thermo"]["energy"]
my_entry = ComputedEntry(structure.composition,
energy,
parameters=params)
self._materials.update_one({"material_id": m["material_id"]}, {
"$set": {
"stability": self.mpr.get_stability([my_entry])[0]
}
})
mpids = self.mpr.find_structure(structure)
self._materials.update_one({"material_id": m["material_id"]},
{"$set": {
"mpids": mpids
}})
except:
import traceback
print("<---")
print(
"There was an error processing material_id: {}".format(m))
traceback.print_exc()
print("--->")
print("MaterialsEhullBuilder finished processing.")
def reset(self):
self._materials.update_many({}, {"$unset": {"stability": 1}})
self._build_indexes()
def _build_indexes(self):
self._materials.create_index("stability.e_above_hull")
@classmethod
def from_file(cls, db_file, m="materials", **kwargs):
"""
Get a MaterialsEhullBuilder using only a db file
Args:
db_file: (str) path to db file
m: (str) name of "materials" collection
**kwargs: other parameters to feed into the builder, e.g. mapi_key
"""
db_write = get_database(db_file, admin=True)
return cls(db_write[m], **kwargs)
class MaterialsEhullBuilder:
def __init__(self, materials_write, mapi_key=None, update_all=False):
"""
Starting with an existing materials collection, adds stability information and
The Materials Project ID.
Args:
materials_write: mongodb collection for materials (write access needed)
mapi_key: (str) Materials API key (if MAPI_KEY env. var. not set)
update_all: (bool) - if true, updates all docs. If false, only updates
docs w/o a stability key
"""
self._materials = materials_write
self.mpr = MPRester(api_key=mapi_key)
self.update_all = update_all
def run(self):
print("MaterialsEhullBuilder starting...")
self._build_indexes()
q = {"thermo.energy": {"$exists": True}}
if not self.update_all:
q["stability"] = {"$exists": False}
mats = [m for m in self._materials.find(q, {"calc_settings": 1, "structure": 1,
"thermo.energy": 1, "material_id": 1})]
pbar = tqdm(mats)
for m in pbar:
pbar.set_description("Processing materials_id: {}".format(m['material_id']))
try:
params = {}
for x in ["is_hubbard", "hubbards", "potcar_spec"]:
params[x] = m["calc_settings"][x]
structure = Structure.from_dict(m["structure"])
energy = m["thermo"]["energy"]
my_entry = ComputedEntry(structure.composition, energy, parameters=params)
self._materials.update_one({"material_id": m["material_id"]},
{"$set": {"stability": self.mpr.get_stability([my_entry])[0]}})
mpids = self.mpr.find_structure(structure)
self._materials.update_one({"material_id": m["material_id"]}, {"$set": {"mpids": mpids}})
except:
import traceback
print("<---")
print("There was an error processing material_id: {}".format(m))
traceback.print_exc()
print("--->")
print("MaterialsEhullBuilder finished processing.")
def reset(self):
self._materials.update_many({}, {"$unset": {"stability": 1}})
self._build_indexes()
def _build_indexes(self):
self._materials.create_index("stability.e_above_hull")
@staticmethod
def from_db_file(db_file, m="materials", **kwargs):
"""
Get a MaterialsEhullBuilder using only a db file
Args:
db_file: (str) path to db file
m: (str) name of "materials" collection
**kwargs: other parameters to feed into the builder, e.g. mapi_key
"""
db_write = get_database(db_file, admin=True)
return MaterialsEhullBuilder(db_write[m], **kwargs)
project=project,
_skip=skip,
_limit=250,
_fields=["identifier"]).result()
existing += [c["identifier"] for c in contribs["data"]]
has_more = contribs["has_more"]
uploaded = len(existing)
print(uploaded, "already uploaded.")
if idx < uploaded:
continue
# structure = Structure.from_dict(input_prim)
if config["use_identifier"]:
structure = input_prim
matches = mpr.find_structure(structure)
if not matches:
print("no match for idx", idx)
matches = [str(idx)]
identifier = matches[0]
if identifier in existing:
continue
if identifier in contributions:
print(idx, identifier, "already parsed")
continue
contrib = {
"project": project,
"identifier": identifier,
"data": {
target: targ_prop
class MaterialsEhullBuilder(AbstractBuilder):
def __init__(self, materials_write, mapi_key=None, update_all=False):
"""
Starting with an existing materials collection, adds stability information and
The Materials Project ID.
Args:
materials_write: mongodb collection for materials (write access needed)
mapi_key: (str) Materials API key (if MAPI_KEY env. var. not set)
update_all: (bool) - if true, updates all docs. If false, only updates
docs w/o a stability key
"""
self._materials = materials_write
self.mpr = MPRester(api_key=mapi_key)
self.update_all = update_all
def run(self):
logger.info("MaterialsEhullBuilder starting...")
self._build_indexes()
q = {"thermo.energy": {"$exists": True}}
if not self.update_all:
q["stability"] = {"$exists": False}
mats = [m for m in self._materials.find(q, {"calc_settings": 1, "structure": 1,
"thermo.energy": 1, "material_id": 1})]
pbar = tqdm(mats)
for m in pbar:
pbar.set_description("Processing materials_id: {}".format(m['material_id']))
try:
params = {}
for x in ["is_hubbard", "hubbards", "potcar_spec"]:
params[x] = m["calc_settings"][x]
structure = Structure.from_dict(m["structure"])
energy = m["thermo"]["energy"]
my_entry = ComputedEntry(structure.composition, energy, parameters=params)
# TODO: @computron This only calculates Ehull with respect to Materials Project.
# It should also account for the current database's results. -computron
self._materials.update_one({"material_id": m["material_id"]},
{"$set": {"stability": self.mpr.get_stability([my_entry])[0]}})
# TODO: @computron: also add additional properties like inverse hull energy?
# TODO: @computron it's better to use PD tool or reaction energy calculator
# Otherwise the compatibility schemes might have issues...one strategy might be
# use MP only to retrieve entries but compute the PD locally -computron
for el, elx in my_entry.composition.items():
entries = self.mpr.get_entries(el.symbol, compatible_only=True)
min_e = min(entries, key=lambda x: x.energy_per_atom).energy_per_atom
energy -= elx * min_e
self._materials.update_one({"material_id": m["material_id"]},
{"$set": {"thermo.formation_energy_per_atom": energy / structure.num_sites}})
mpids = self.mpr.find_structure(structure)
self._materials.update_one({"material_id": m["material_id"]}, {"$set": {"mpids": mpids}})
except:
import traceback
logger.exception("<---")
logger.exception("There was an error processing material_id: {}".format(m))
logger.exception(traceback.format_exc())
logger.exception("--->")
logger.info("MaterialsEhullBuilder finished processing.")
def reset(self):
logger.info("Resetting MaterialsEhullBuilder")
self._materials.update_many({}, {"$unset": {"stability": 1}})
self._build_indexes()
logger.info("Finished resetting MaterialsEhullBuilder")
def _build_indexes(self):
self._materials.create_index("stability.e_above_hull")
@classmethod
def from_file(cls, db_file, m="materials", **kwargs):
"""
Get a MaterialsEhullBuilder using only a db file
Args:
db_file: (str) path to db file
m: (str) name of "materials" collection
**kwargs: other parameters to feed into the builder, e.g. mapi_key
"""
db_write = get_database(db_file, admin=True)
return cls(db_write[m], **kwargs)