def as_dict(self):
enc = MontyEncoder()
return {'mode': self.mode, 'comment': self.comment, 'num_kpts': self.num_kpts,
'kpts': enc.default(np.array(self.kpts)), 'kpt_shifts': self.kpt_shifts,
'kpts_weights': self.kpts_weights, 'use_symmetries': self.use_symmetries,
'use_time_reversal': self.use_time_reversal, 'chksymbreak': self.chksymbreak,
'@module': self.__class__.__module__, '@class': self.__class__.__name__}
def as_dict(self):
enc = MontyEncoder()
return {'mode': self.mode.name, 'comment': self.comment,
'num_kpts': self.num_kpts,
'kpts': enc.default(np.array(self.kpts)), 'kpt_shifts': self.kpt_shifts,
'kpts_weights': self.kpts_weights, 'use_symmetries': self.use_symmetries,
'use_time_reversal': self.use_time_reversal, 'chksymbreak': self.chksymbreak,
'@module': self.__class__.__module__, '@class': self.__class__.__name__}
def as_dict(self):
"""
Note: stores the real and imaginary part of the dielectric tensor
separately, due to issues with JSON serializing complex numbers.
Returns:
dict: Dictionary representation of the DielTensor instance.
"""
d = dict()
d["energies"] = MontyEncoder().default(self.energies)
d["real_diel"] = MontyEncoder().default(self.dielectric_tensor.real)
d["imag_diel"] = MontyEncoder().default(self.dielectric_tensor.imag)
return d
def test_entry(self):
enc = MontyEncoder()
dec = MontyDecoder()
entry = ComputedEntry("Fe2O3", 2.3)
jsonstr = enc.encode(entry)
d = dec.decode(jsonstr)
self.assertEqual(type(d), ComputedEntry)
#Check list of entries
entries = [entry, entry, entry]
jsonstr = enc.encode(entries)
d = dec.decode(jsonstr)
for i in d:
self.assertEqual(type(i), ComputedEntry)
self.assertEqual(len(d), 3)
def test_entry(self):
enc = MontyEncoder()
dec = MontyDecoder()
entry = ComputedEntry("Fe2O3", 2.3)
jsonstr = enc.encode(entry)
d = dec.decode(jsonstr)
self.assertEqual(type(d), ComputedEntry)
#Check list of entries
entries = [entry, entry, entry]
jsonstr = enc.encode(entries)
d = dec.decode(jsonstr)
for i in d:
self.assertEqual(type(i), ComputedEntry)
self.assertEqual(len(d), 3)
def as_dict(self):
d = super(Slab, self).as_dict()
d["@module"] = self.__class__.__module__
d["@class"] = self.__class__.__name__
d["oriented_unit_cell"] = self.oriented_unit_cell.as_dict()
d["miller_index"] = self.miller_index
d["shift"] = self.shift
d["scale_factor"] = MontyEncoder().default(self.scale_factor)
d["reconstruction"] = self.reconstruction
d["energy"] = self.energy
return d
def as_dict(self):
"""Serialize the reference as a dict."""
schema = self.output_schema
schema_dict = MontyEncoder().default(
schema) if schema is not None else None
data = {
"@module": self.__class__.__module__,
"@class": self.__class__.__name__,
"@version": None,
"uuid": self.uuid,
"attributes": self.attributes,
"output_schema": schema_dict,
}
return data
def test_serialization():
import json
from monty.json import MontyDecoder, MontyEncoder
from jobflow import Job
test_job = Job(function=add, function_args=(1, ), function_kwargs={"b": 2})
uuid = test_job.uuid
encoded_job = json.loads(MontyEncoder().encode(test_job))
decoded_job = MontyDecoder().process_decoded(encoded_job)
assert decoded_job.uuid == uuid
def test_serialization():
import json
from monty.json import MontyDecoder, MontyEncoder
from jobflow import Flow
flow = Flow([])
flow_host = Flow([flow])
host_uuid = flow_host.uuid
encoded_flow = json.loads(MontyEncoder().encode(flow_host))
decoded_flow = MontyDecoder().process_decoded(encoded_flow)
assert decoded_flow.jobs[0].host == host_uuid
def test_pandas(self):
cls = ClassContainingDataFrame(df=pd.DataFrame([{
"a": 1,
"b": 1
}, {
"a": 1,
"b": 2
}]))
d = json.loads(MontyEncoder().encode(cls))
self.assertEqual(d["df"]["@module"], "pandas")
self.assertEqual(d["df"]["@class"], "DataFrame")
obj = ClassContainingDataFrame.from_dict(d)
self.assertIsInstance(obj, ClassContainingDataFrame)
self.assertIsInstance(obj.df, pd.DataFrame)
self.assertEqual(list(obj.df.a), [1, 1])
def default(obj):
"""
For use with msgpack.packb(obj, default=default). Supports Monty's as_dict
protocol, numpy arrays and datetime.
"""
return MontyEncoder().default(obj)
def run_task(self, fw_spec):
# get the database connection
db_file = env_chk(self["db_file"], fw_spec)
mmdb = VaspCalcDb.from_db_file(db_file, admin=True)
wf_uuid = self["approx_neb_wf_uuid"]
# check if provided approx_neb_wf_uuid is unique
# e.g. not already used in approx_neb collection
approx_neb_db = mmdb.db["approx_neb"]
if approx_neb_db.count_documents({"wf_uuid": wf_uuid}) != 0:
raise ValueError(
"Provided approx_neb_wf_uuid is not unique. A unique workflow id is required for querying in the approx_neb workflow."
)
# update host task doc (from host_task_id) with unique wf_uuid
# (tracks approx_neb workflows generated from this host task doc)
t_id = self.get("host_task_id", fw_spec.get("host_task_id"))
host_tasks_doc = mmdb.collection.find_one_and_update(
{
"task_id": t_id,
"approx_neb.calc_type": "host"
},
{"$push": {
"approx_neb.wf_uuids": wf_uuid
}},
)
if host_tasks_doc == None:
raise ValueError(
"Error updating approx neb host with task_id: {}".format(t_id))
# Initialize and store select host task doc fields in approx_neb_doc
# (to be stored in the approx_neb collection)
approx_neb_doc = {
"wf_uuid": wf_uuid,
"host": {
"dir_name": host_tasks_doc["dir_name"],
"chemsys": host_tasks_doc["chemsys"],
"formula_pretty": host_tasks_doc["formula_pretty"],
"input_structure": host_tasks_doc["input"]["structure"],
"output": host_tasks_doc["output"],
"task_id": host_tasks_doc["task_id"],
},
"end_points": [],
}
# ensure tags and additional_fields are the same
# in both the approx_neb and tasks collections
additional_fields = self.get("additional_fields", {})
if isinstance(additional_fields, dict):
for key, value in additional_fields.items():
if key not in approx_neb_doc.keys():
approx_neb_doc[key] = value
tags = self.get("tags")
if tags:
approx_neb_doc["tags"] = tags
# insert approx_neb_doc in the approx_neb collection of provided database
# includes fix to ensure approx_neb_doc is a json serializable dict
approx_neb_doc = MontyEncoder().encode(approx_neb_doc)
approx_neb_doc = loads(approx_neb_doc)
approx_neb_doc["last_updated"] = datetime.utcnow()
mmdb.collection = mmdb.db["approx_neb"]
mmdb.collection.insert_one(approx_neb_doc)
# Update fw_spec with approx_neb_doc and store wf_uuid
# in launches collection for record keeping
return FWAction(stored_data={
"wf_uuid": wf_uuid,
"approx_neb_doc": approx_neb_doc
})