def im_sol_sub_def_energy_parse():
m_description = 'Command to parse solute substitution defect ' \
'energies for intermetallics from the VASP DFT ' \
'calculations.'
parser = ArgumentParser(description=m_description)
parser.add_argument("--mpid",
type=str.lower,
help="Materials Project id of the intermetallic structure.\n" \
"For more info on Materials Project, please refer to " \
"www.materialsproject.org")
parser.add_argument("--solute", help="Solute Element")
parser.add_argument("--mapi_key",
default = None,
help="Your Materials Project REST API key.\n" \
"For more info, please refer to " \
"www.materialsproject.org/opne")
args = parser.parse_args()
energy_dict = solute_def_parse_energy(args.mpid, args.solute,
args.mapi_key)
if energy_dict:
fl_nm = args.mpid+'_solute-'+args.solute+'_raw_defect_energy.json'
dumpfn(energy_dict, fl_nm, indent=2, cls=MontyEncoder)
def im_vac_antisite_def_energy_parse():
m_description = 'Command to parse vacancy and antisite defect ' \
'energies for intermetallics from the VASP DFT ' \
'calculations.'
parser = ArgumentParser(description=m_description)
parser.add_argument("--mpid",
type=str.lower,
help="Materials Project id of the intermetallic structure.\n" \
"For more info on Materials Project, please refer to " \
"www.materialsproject.org")
parser.add_argument("--mapi_key",
default = None,
help="Your Materials Project REST API key.\n" \
"For more info, please refer to " \
"www.materialsproject.org/opne")
args = parser.parse_args()
print args
energy_dict = vac_antisite_def_parse_energy(args.mpid, args.mapi_key)
print type(energy_dict)
for key,value in energy_dict.items():
print key
print type(key), type(value)
for key2, val2 in value.items():
print type(key2), type(val2)
if energy_dict:
fl_nm = args.mpid+'_raw_defect_energy.json'
dumpfn(energy_dict, fl_nm, cls=MontyEncoder, indent=2)
def run_task(self, fw_spec):
transformations = []
transformation_params = self.get("transformation_params",
[{} for i in range(len(self["transformations"]))])
for t in self["transformations"]:
found = False
for m in ["advanced_transformations", "defect_transformations",
"site_transformations", "standard_transformations"]:
mod = import_module("pymatgen.transformations.{}".format(m))
try:
t_cls = getattr(mod, t)
except AttributeError:
continue
t_obj = t_cls(**transformation_params.pop(0))
transformations.append(t_obj)
found = True
if not found:
raise ValueError("Could not find transformation: {}".format(t))
# TODO: @matk86 - should prev_calc_dir use CONTCAR instead of POSCAR? Note that if
# current dir, maybe it is POSCAR indeed best ... -computron
structure = self['structure'] if not self.get('prev_calc_dir', None) else \
Poscar.from_file(os.path.join(self['prev_calc_dir'], 'POSCAR')).structure
ts = TransformedStructure(structure)
transmuter = StandardTransmuter([ts], transformations)
final_structure = transmuter.transformed_structures[-1].final_structure.copy()
vis_orig = self["vasp_input_set"]
vis_dict = vis_orig.as_dict()
vis_dict["structure"] = final_structure.as_dict()
vis_dict.update(self.get("override_default_vasp_params", {}) or {})
vis = vis_orig.__class__.from_dict(vis_dict)
vis.write_input(".")
dumpfn(transmuter.transformed_structures[-1], "transformations.json")
def run(self, job_cmd=None):
"""
run the vasp jobs through custodian
if the job list is empty,
run a single job with the initial input set
"""
for j in self.jobs:
if job_cmd is not None:
j.job_cmd = job_cmd
else:
j.job_cmd = self.job_cmd
c_params = {'jobs': [j.as_dict() for j in self.jobs],
'handlers': [h.as_dict() for h in self.handlers],
'max_errors': 5}
c = Custodian(self.handlers, self.jobs, max_errors=5)
c.run()
for j in self.jobs:
self.cal_log.append({"job": j.as_dict(),
'job_id': j.job_id,
"corrections": [],
'final_energy': None})
self.job_ids.append(j.job_id)
if self.checkpoint_file:
dumpfn(self.cal_log, self.checkpoint_file,
cls=MontyEncoder, indent=4)
else:
dumpfn(self.cal_log, Calibrate.LOG_FILE, cls=MontyEncoder,
indent=4)
def setup(self):
"""
Performs initial setup for VaspJob, including overriding any settings
and backing up.
"""
decompress_dir('.')
if self.backup:
for f in VASP_INPUT_FILES:
shutil.copy(f, "{}.orig".format(f))
if self.auto_npar:
try:
incar = Incar.from_file("INCAR")
# Only optimized NPAR for non-HF and non-RPA calculations.
if not (incar.get("LHFCALC") or incar.get("LRPA") or
incar.get("LEPSILON")):
if incar.get("IBRION") in [5, 6, 7, 8]:
# NPAR should not be set for Hessian matrix
# calculations, whether in DFPT or otherwise.
del incar["NPAR"]
else:
import multiprocessing
# try sge environment variable first
# (since multiprocessing counts cores on the current
# machine only)
ncores = os.environ.get('NSLOTS') or \
multiprocessing.cpu_count()
ncores = int(ncores)
for npar in range(int(math.sqrt(ncores)),
ncores):
if ncores % npar == 0:
incar["NPAR"] = npar
break
incar.write_file("INCAR")
except:
pass
if self.auto_continue:
if os.path.exists("continue.json"):
actions = loadfn("continue.json").get("actions")
logger.info("Continuing previous VaspJob. Actions: {}".format(actions))
backup(VASP_BACKUP_FILES, prefix="prev_run")
VaspModder().apply_actions(actions)
else:
# Default functionality is to copy CONTCAR to POSCAR and set
# ISTART to 1 in the INCAR, but other actions can be specified
if self.auto_continue is True:
actions = [{"file": "CONTCAR",
"action": {"_file_copy": {"dest": "POSCAR"}}},
{"dict": "INCAR",
"action": {"_set": {"ISTART": 1}}}]
else:
actions = self.auto_continue
dumpfn({"actions": actions}, "continue.json")
if self.settings_override is not None:
VaspModder().apply_actions(self.settings_override)
def generate_single_job_dict():
"""
Used to generate test dictionary for single jobs.
"""
single_job_dict = {}
for file in single_job_out_names:
single_job_dict[file] = QCOutput(os.path.join(test_dir, file)).data
dumpfn(single_job_dict, "single_job.json")
def run(self):
"""
Override of Custodian.run() to include instructions to copy the
temp_dir to the scratch partition on slave compute nodes if requested.
"""
cwd = os.getcwd()
with ScratchDir(self.scratch_dir, create_symbolic_link=True,
copy_to_current_on_exit=True,
copy_from_current_on_enter=True) as temp_dir:
self._manage_node_scratch(temp_dir_path=temp_dir,
job_start=True)
self.total_errors = 0
start = datetime.datetime.now()
logger.info("Run started at {} in {}.".format(
start, temp_dir))
v = sys.version.replace("\n", " ")
logger.info("Custodian running on Python version {}".format(v))
try:
# skip jobs until the restart
for job_n, job in islice(enumerate(self.jobs, 1),
self.restart, None):
self._run_job(job_n, job, temp_dir)
# Checkpoint after each job so that we can recover from
# last point and remove old checkpoints
if self.checkpoint:
super(SSHCustodian, self)._save_checkpoint(cwd, job_n)
except CustodianError as ex:
logger.error(ex.message)
if ex.raises:
raise RuntimeError("{} errors reached: {}. Exited..."
.format(self.total_errors, ex))
finally:
# Log the corrections to a json file.
logger.info("Logging to {}...".format(super(SSHCustodian,
self).LOG_FILE))
dumpfn(self.run_log, super(SSHCustodian, self).LOG_FILE,
cls=MontyEncoder, indent=4)
end = datetime.datetime.now()
logger.info("Run ended at {}.".format(end))
run_time = end - start
logger.info("Run completed. Total time taken = {}."
.format(run_time))
# Remove duplicate copy of log file, provided it ends with
# ".log"
for x in ([x for x in os.listdir(temp_dir)
if re.match(r'\w*\.log', x)]):
os.remove(os.path.join(temp_dir, x))
self._manage_node_scratch(temp_dir_path=temp_dir,
job_start=False)
if self.gzipped_output:
gzip_dir(".")
# Cleanup checkpoint files (if any) if run is successful.
super(SSHCustodian, self)._delete_checkpoints(cwd)
return self.run_log
def run(self):
"""
Runs all the jobs jobs.
Returns:
All errors encountered as a list of list.
[[error_dicts for job 1], [error_dicts for job 2], ....]
"""
cwd = os.getcwd()
with ScratchDir(self.scratch_dir, create_symbolic_link=True,
copy_to_current_on_exit=True,
copy_from_current_on_enter=True) as temp_dir:
self.total_errors = 0
start = datetime.datetime.now()
logger.info("Run started at {} in {}.".format(
start, temp_dir))
v = sys.version.replace("\n", " ")
logger.info("Custodian running on Python version {}".format(v))
logger.info("Hostname: {}, Cluster: {}".format(
*get_execution_host_info()))
try:
# skip jobs until the restart
for job_n, job in islice(enumerate(self.jobs, 1),
self.restart, None):
self._run_job(job_n, job)
# Checkpoint after each job so that we can recover from last
# point and remove old checkpoints
if self.checkpoint:
self.restart = job_n
Custodian._save_checkpoint(cwd, job_n)
except CustodianError as ex:
logger.error(ex.message)
if ex.raises:
raise RuntimeError("{} errors reached: {}. Exited..."
.format(self.total_errors, ex))
finally:
# Log the corrections to a json file.
logger.info("Logging to {}...".format(Custodian.LOG_FILE))
dumpfn(self.run_log, Custodian.LOG_FILE, cls=MontyEncoder,
indent=4)
end = datetime.datetime.now()
logger.info("Run ended at {}.".format(end))
run_time = end - start
logger.info("Run completed. Total time taken = {}."
.format(run_time))
if self.gzipped_output:
gzip_dir(".")
# Cleanup checkpoint files (if any) if run is successful.
Custodian._delete_checkpoints(cwd)
return self.run_log
def generate_multi_job_dict():
"""
Used to generate test dictionary for multiple jobs
"""
multi_job_dict = {}
for file in multi_job_out_names:
outputs = QCOutput.multiple_outputs_from_file(QCOutput, os.path.join(test_dir, file), keep_sub_files=False)
data = []
for sub_output in outputs:
data.append(sub_output.data)
multi_job_dict[file] = data
dumpfn(multi_job_dict, "multi_job.json")
def update_checkpoint(launchpad, launch_id, checkpoint):
"""
Helper function to update checkpoint
Args:
launchpad (LaunchPad): LaunchPad to ping with checkpoint data
launch_id (int): launch id to update
checkpoint (dict): checkpoint data
"""
if launchpad:
launchpad.ping_launch(launch_id, checkpoint=checkpoint)
else:
offline_info = loadfn("FW_offline.json")
offline_info.update({"checkpoint": checkpoint})
dumpfn(offline_info, "FW_offline.json")
def add_config_var(args):
d = {}
if os.path.exists(SETTINGS_FILE):
shutil.copy(SETTINGS_FILE, SETTINGS_FILE + ".bak")
print("Existing %s backed up to %s"
% (SETTINGS_FILE, SETTINGS_FILE + ".bak"))
d = loadfn(SETTINGS_FILE)
toks = args.var_spec
if len(toks) % 2 != 0:
print("Bad variable specification!")
sys.exit(-1)
for i in range(int(len(toks) / 2)):
d[toks[2 * i]] = toks[2 * i + 1]
dumpfn(d, SETTINGS_FILE, default_flow_style=False)
print("New %s written!" % (SETTINGS_FILE))
def _do_check(self, handlers, terminate_func=None):
"""
checks the specified handlers. Returns True iff errors caught
"""
corrections = []
for h in handlers:
try:
if h.check():
if h.max_num_corrections is not None \
and h.n_applied_corrections >= h.max_num_corrections:
msg = "Maximum number of corrections {} reached " \
"for handler {}".format(h.max_num_corrections, h)
if h.raise_on_max:
self.run_log[-1]["handler"] = h
self.run_log[-1]["max_errors_per_handler"] = True
raise MaxCorrectionsPerHandlerError(msg, True, h.max_num_corrections, h)
else:
logger.warning(msg+" Correction not applied.")
continue
if terminate_func is not None and h.is_terminating:
logger.info("Terminating job")
terminate_func()
# make sure we don't terminate twice
terminate_func = None
d = h.correct()
d["handler"] = h
logger.error("\n" + pformat(d, indent=2, width=-1))
corrections.append(d)
h.n_applied_corrections += 1
except Exception:
if not self.skip_over_errors:
raise
else:
import traceback
logger.error("Bad handler %s " % h)
logger.error(traceback.format_exc())
corrections.append(
{"errors": ["Bad handler %s " % h],
"actions": []})
self.total_errors += len(corrections)
self.errors_current_job += len(corrections)
self.run_log[-1]["corrections"].extend(corrections)
# We do a dump of the run log after each check.
dumpfn(self.run_log, Custodian.LOG_FILE, cls=MontyEncoder,
indent=4)
return len(corrections) > 0
def test_DiagnosticProperties_class(self):
with ScratchDir("."):
os.environ["BEEP_PROCESSING_DIR"] = TEST_FILE_DIR
pcycler_run_loc = os.path.join(
TEST_FILE_DIR,
"PreDiag_000240_000227_truncated_structure.json")
pcycler_run = auto_load_processed(pcycler_run_loc)
featurizer = DiagnosticProperties.from_run(pcycler_run_loc,
os.getcwd(),
pcycler_run)
path, local_filename = os.path.split(featurizer.name)
folder = os.path.split(path)[-1]
dumpfn(featurizer, featurizer.name)
self.assertEqual(folder, "DiagnosticProperties")
self.assertEqual(featurizer.X.shape, (30, 9))
print(list(featurizer.X.iloc[2, :]))
self.assertListEqual(list(featurizer.X.iloc[2, :]), [
141, 0.9859837086597274, 7.885284043, 4.323121513988055,
21.12108276469096, 30, 100, 'reset', 'discharge_energy'
])
def test_mpk(self):
d = {"hello": "world"}
# Test automatic format detection
dumpfn(d, "monte_test.mpk")
d2 = loadfn("monte_test.mpk")
self.assertEqual(
d, {k.decode('utf-8'): v.decode('utf-8')
for k, v in d2.items()})
os.remove("monte_test.mpk")
# Test to ensure basename is respected, and not directory
with ScratchDir('.'):
os.mkdir("mpk_test")
os.chdir("mpk_test")
fname = os.path.abspath("test_file.json")
dumpfn({"test": 1}, fname)
with open("test_file.json", "r") as f:
reloaded = json.loads(f.read())
self.assertEqual(reloaded['test'], 1)
def save_model(self, filename: str) -> None:
"""
Save the model to a keras model hdf5 and a json config for additional
converters
Args:
filename: (str) output file name
Returns:
None
"""
self.model.save(filename)
dumpfn(
{
"graph_converter": self.graph_converter,
"target_scaler": self.target_scaler,
"metadata": self.metadata
},
filename + ".json",
)
def save_model(self, filename):
"""
Save the model to a keras model hdf5 and a json config for additional
converters
Args:
filename: (str) output file name
Returns:
None
"""
self.model.save(filename)
dumpfn(
{
'graph_converter': self.graph_converter,
'target_scaler': self.target_scaler,
'metadata': self.metadata
},
filename + '.json'
)
def add_config_var(args):
"""
Add configuration args.
:param args:
"""
d = {}
if os.path.exists(SETTINGS_FILE):
shutil.copy(SETTINGS_FILE, SETTINGS_FILE + ".bak")
print("Existing {} backed up to {}".format(SETTINGS_FILE,
SETTINGS_FILE + ".bak"))
d = loadfn(SETTINGS_FILE)
toks = args.var_spec
if len(toks) % 2 != 0:
print("Bad variable specification!")
sys.exit(-1)
for i in range(int(len(toks) / 2)):
d[toks[2 * i]] = toks[2 * i + 1]
dumpfn(d, SETTINGS_FILE)
print("New %s written!" % (SETTINGS_FILE))
def test_formula_query():
op = FormulaQuery()
assert op.query("Si2O4") == {
"criteria": {
"composition_reduced.O": 2.0,
"composition_reduced.Si": 1.0,
"nelements": 2,
}
}
with ScratchDir("."):
dumpfn(op, "temp.json")
new_op = loadfn("temp.json")
assert new_op.query("Si2O4") == {
"criteria": {
"composition_reduced.O": 2.0,
"composition_reduced.Si": 1.0,
"nelements": 2,
}
}
def test_multi_material_id_query():
op = MultiMaterialIDQuery()
assert op.query(material_ids="mp-149, mp-13") == {
"criteria": {
"material_id": {
"$in": ["mp-149", "mp-13"]
}
}
}
with ScratchDir("."):
dumpfn(op, "temp.json")
new_op = loadfn("temp.json")
assert new_op.query(material_ids="mp-149, mp-13") == {
"criteria": {
"material_id": {
"$in": ["mp-149", "mp-13"]
}
}
}
def create_aflow_test_docs():
auids = [
'aflow:0132ab6b9cddd429', # Has an elastic tensor file
'aflow:0136cbe39e59c471', # An average joe material
'aflow:d0c93a9396dc599e'
]
query = AflowAPIQuery.from_pymongo({'auid': {
'$in': auids
}},
AflowIngester._available_kws,
50,
property_reduction=True)
if query.N != len(auids):
auids_retrieved = [
material['auid'] for page in query.responses.values()
for material in page.values()
]
auids_not_retrieved = set(auids) - set(auids_retrieved)
raise ValueError(
"Not all materials retrieved. Perhaps they have been deprecated? "
"Unavailabie auids:\n{}".format(auids_not_retrieved))
data = []
for item in query:
raw_data = item.raw
try:
contcar_data = item.files['CONTCAR.relax.vasp']()
except Exception:
contcar_data = None
try:
elastic_tensor_data = item.files['AEL_elastic_tensor.json']()
elastic_tensor_data = json.loads(elastic_tensor_data)
except Exception:
elastic_tensor_data = None
raw_data['CONTCAR_relax_vasp'] = contcar_data
raw_data['AEL_elastic_tensor_json'] = elastic_tensor_data
data.append(raw_data)
dumpfn(data, os.path.join(TEST_DATA_DIR, 'aflow_store.json'))
def references_to_bib(refs):
"""
Takes a list of reference strings and converts them to bibtex
entries
Args:
refs ([str]): list of string references, which can be
bibtex entries, digital object identifiers ("doi:DOI_GOES_HERE")
or urls ("url:URL_GOES_HERE")
Returns:
(list): list of bibtex formatted strings
"""
parsed_refs = []
for ref in refs:
if ref in _REFERENCE_CACHE:
parsed_ref = _REFERENCE_CACHE[ref]
elif ref.startswith('@'):
parsed_ref = ref
elif ref.startswith('url:'):
# uses arbitrary key
url = ref.split('url:')[1]
parsed_ref = """@misc{{url:{0},
url = {{{1}}}
}}""".format(str(abs(url.__hash__()))[0:6], url)
elif ref.startswith('doi:'):
doi = ref.split('doi:')[1]
parsed_ref = content_negotiation(doi, format='bibentry')
else:
raise ValueError(
'Unknown reference style for '
'reference: {} (please either '
'supply a BibTeX string, or a string '
'starting with url: followed by a URL or '
'starting with doi: followed by a DOI)'.format(ref))
if ref not in _REFERENCE_CACHE:
_REFERENCE_CACHE[ref] = parsed_ref
dumpfn(_REFERENCE_CACHE, _REFERENCE_CACHE_PATH)
parsed_refs.append(parsed_ref)
return parsed_refs
def test_grain_boundary_structure_query():
op = GBStructureQuery()
assert op.query(
sigma=5,
type=GBTypeEnum.twist,
chemsys="Si-Fe",
pretty_formula="Fe2Si4",
gb_plane="1,1,1",
rotation_axis="1,0,1",
) == {
"criteria": {
"sigma": 5,
"type": "twist",
"chemsys": "Fe-Si",
"pretty_formula": "FeSi2",
"gb_plane": [1, 1, 1],
"rotation_axis": [1, 0, 1],
}
}
with ScratchDir("."):
dumpfn(op, "temp.json")
new_op = loadfn("temp.json")
assert new_op.query(
sigma=5,
type=GBTypeEnum.twist,
chemsys="Si-Fe",
pretty_formula="Fe2Si4",
gb_plane="1,1,1",
rotation_axis="1,0,1",
) == {
"criteria": {
"sigma": 5,
"type": "twist",
"chemsys": "Fe-Si",
"pretty_formula": "FeSi2",
"gb_plane": [1, 1, 1],
"rotation_axis": [1, 0, 1],
}
}
def post_equi(confs, inter_param):
# find all POSCARs and their name like mp-xxx
# ...
conf_dirs = []
for conf in confs:
conf_dirs.extend(glob.glob(conf))
conf_dirs.sort()
task_dirs = []
for ii in conf_dirs:
task_dirs.append(os.path.abspath(os.path.join(ii, 'relaxation', 'relax_task')))
task_dirs.sort()
# generate a list of task names like mp-xxx/relaxation
# ...
# dump the relaxation result.
for ii in task_dirs:
poscar = os.path.join(ii, 'POSCAR')
inter = make_calculator(inter_param, poscar)
res = inter.compute(ii)
dumpfn(res, os.path.join(ii, 'result.json'), indent=4)
def test_molecule_elements_query():
op = MoleculeElementsQuery()
assert op.query(elements="Si, O, P") == {
"criteria": {
"elements": {
"$all": ["Si", "O", "P"]
}
}
}
with ScratchDir("."):
dumpfn(op, "temp.json")
new_op = loadfn("temp.json")
assert new_op.query(elements="Si, O, P") == {
"criteria": {
"elements": {
"$all": ["Si", "O", "P"]
}
}
}
def generate_characterize_path_files(rn, old_solved_PRs, dist_and_path):
pickle_in = open(
os.path.join(
test_dir,
"unittest_RN_before_characterize_path.pkl",
),
"wb",
)
pickle.dump(rn, pickle_in)
pickle_in = open(
os.path.join(test_dir, "unittest_characterize_path_PRs_IN.pkl"),
"wb",
)
pickle.dump(old_solved_PRs, pickle_in)
dumpfn(
dist_and_path,
os.path.join(
test_dir,
"unittest_characterize_path_path_IN.json",
),
)
def test_from_csv(self):
csv_file = os.path.join(TEST_FILE_DIR, "parameter_test.csv")
# Test basic functionality
with ScratchDir('.') as scratch_dir:
makedirs_p(os.path.join(scratch_dir, "procedures"))
makedirs_p(os.path.join(scratch_dir, "names"))
generate_protocol_files_from_csv(csv_file, scratch_dir)
self.assertEqual(
len(os.listdir(os.path.join(scratch_dir, "procedures"))), 3)
# Test avoid overwriting file functionality
with ScratchDir('.') as scratch_dir:
makedirs_p(os.path.join(scratch_dir, "procedures"))
makedirs_p(os.path.join(scratch_dir, "names"))
dumpfn({"hello": "world"}, "procedures/name_000007.000")
generate_protocol_files_from_csv(csv_file, scratch_dir)
post_file = loadfn("procedures/name_000007.000")
self.assertEqual(post_file, {"hello": "world"})
self.assertEqual(
len(os.listdir(os.path.join(scratch_dir, "procedures"))), 3)
def test_has_props_query():
op = HasPropsQuery()
assert op.query(has_props="electronic_structure, thermo") == {
"criteria": {
"has_props": {
"$all": ["electronic_structure", "thermo"]
}
}
}
with ScratchDir("."):
dumpfn(op, "temp.json")
new_op = loadfn("temp.json")
assert new_op.query(has_props="electronic_structure, thermo") == {
"criteria": {
"has_props": {
"$all": ["electronic_structure", "thermo"]
}
}
}
def test_shear_modulus_query():
op = ShearModulusQuery()
q = op.query(
g_voigt_min=0,
g_voigt_max=5,
g_reuss_min=0,
g_reuss_max=5,
g_vrh_min=0,
g_vrh_max=5,
)
fields = ["elasticity.g_voigt", "elasticity.g_reuss", "elasticity.g_vrh"]
assert q == {
"criteria": {field: {
"$gte": 0,
"$lte": 5
}
for field in fields}
}
with ScratchDir("."):
dumpfn(op, "temp.json")
new_op = loadfn("temp.json")
q = new_op.query(
g_voigt_min=0,
g_voigt_max=5,
g_reuss_min=0,
g_reuss_max=5,
g_vrh_min=0,
g_vrh_max=5,
)
assert q == {
"criteria": {field: {
"$gte": 0,
"$lte": 5
}
for field in fields}
}
def test_insertion_voltage_step_query():
op = InsertionVoltageStepQuery()
q = op.query(
stability_charge_min=0,
stability_charge_max=5,
stability_discharge_min=0,
stability_discharge_max=5,
)
fields = [
"stability_charge",
"stability_discharge",
]
assert q == {
"criteria": {field: {
"$gte": 0,
"$lte": 5
}
for field in fields}
}
with ScratchDir("."):
dumpfn(op, "temp.json")
new_op = loadfn("temp.json")
q = new_op.query(
stability_charge_min=0,
stability_charge_max=5,
stability_discharge_min=0,
stability_discharge_max=5,
)
assert q == {
"criteria": {field: {
"$gte": 0,
"$lte": 5
}
for field in fields}
}
def test_serialization(self):
with ScratchDir("."):
os.environ["BEEP_PROCESSING_DIR"] = os.getcwd()
dataset = BeepDataset.from_features('test_dataset', ['PreDiag'], FEATURIZER_CLASSES,
feature_dir=os.path.join(TEST_FILE_DIR, 'data-share/features'))
dumpfn(dataset, 'temp_dataset.json')
dataset = loadfn('temp_dataset.json')
self.assertEqual(dataset.name, 'test_dataset')
self.assertEqual(dataset.data.shape, (2, 56))
# from pdb import set_trace; set_trace()
self.assertListEqual(list(dataset.data.seq_num), [196, 197])
self.assertIsNone(dataset.X_test)
self.assertSetEqual(set(dataset.feature_sets.keys()), {'RPTdQdVFeatures', 'DiagnosticSummaryStats'})
self.assertEqual(dataset.missing.feature_class.iloc[0], 'HPPCResistanceVoltageFeatures')
self.assertIsInstance(dataset.filenames, list)
os.environ["BEEP_PROCESSING_DIR"] = os.getcwd()
dataset2 = BeepDataset.from_features('test_dataset', ['PreDiag'], [RPTdQdVFeatures],
feature_dir=os.path.join(TEST_FILE_DIR, 'data-share/features'))
dumpfn(dataset2, "temp_dataset_2.json")
dataset2 = loadfn('temp_dataset_2.json')
self.assertEqual(dataset2.missing.columns.to_list(), ["filename", "feature_class"])
def run_task(self, fw_spec):
db_file = env_chk(self["db_file"], fw_spec)
wf_uuid = self["wf_uuid"]
mc_settings = self.get("mc_settings", {})
# Get Heisenberg models from db
mmdb = VaspCalcDb.from_db_file(db_file, admin=True)
mmdb.collection = mmdb.db["exchange"]
# Get documents
docs = list(
mmdb.collection.find({"wf_meta.wf_uuid": wf_uuid},
["heisenberg_model", "nn_cutoff"]))
hmodels = [
HeisenbergModel.from_dict(d["heisenberg_model"]) for d in docs
]
cutoffs = [hmodel.cutoff for hmodel in hmodels]
ordered_hmodels = [
h for _, h in sorted(zip(cutoffs, hmodels), reverse=False)
]
# Take the model with smallest NN cutoff
hmodel = ordered_hmodels[0]
# Get a converged Heisenberg model if one was found
# if fw_spec["converged_heisenberg_model"]:
# hmodel = HeisenbergModel.from_dict(fw_spec["converged_heisenberg_model"])
vc = VampireCaller(hm=hmodel, **mc_settings)
vampire_output = vc.output
# Update FW spec
update_spec = {"vampire_output": vampire_output}
# Write to file
dumpfn(vampire_output.as_dict(), "vampire_output.json")
return FWAction(update_spec=update_spec)
def test_RPTdQdVFeatures_class(self):
with ScratchDir("."):
os.environ["BEEP_PROCESSING_DIR"] = TEST_FILE_DIR
pcycler_run_loc = os.path.join(
TEST_FILE_DIR,
"PreDiag_000240_000227_truncated_structure.json")
pcycler_run = loadfn(pcycler_run_loc)
params_dict = {
"diag_ref": 0,
"diag_nr": 2,
"charge_y_n": 1,
"rpt_type": "rpt_2C",
"plotting_y_n": 0,
}
featurizer = RPTdQdVFeatures.from_run(pcycler_run_loc, os.getcwd(),
pcycler_run, params_dict)
path, local_filename = os.path.split(featurizer.name)
folder = os.path.split(path)[-1]
dumpfn(featurizer, featurizer.name)
self.assertEqual(folder, "RPTdQdVFeatures")
self.assertEqual(featurizer.X.shape[1], 11)
self.assertEqual(featurizer.metadata["parameters"], params_dict)
def do_query(args):
m = MPRester()
try:
criteria = json.loads(args.criteria)
except json.decoder.JSONDecodeError:
criteria = args.criteria
if args.structure:
count = 0
for d in m.query(criteria, properties=["structure", "task_id"]):
s = d["structure"]
formula = re.sub(r"\s+", "", s.formula)
if args.structure == "poscar":
fname = "POSCAR.%s_%s" % (d["task_id"], formula)
else:
fname = "%s-%s.%s" % (d["task_id"], formula, args.structure)
s.to(filename=fname)
count += 1
print("%d structures written!" % count)
elif args.entries:
entries = m.get_entries(criteria)
dumpfn(entries, args.entries)
print("%d entries written to %s!" % (len(entries), args.entries))
else:
props = ["e_above_hull", "spacegroup"]
props += args.data
entries = m.get_entries(criteria, property_data=props)
t = []
headers = ["mp-id", "Formula", "Spacegroup", "E/atom (eV)",
"E above hull (eV)"] + args.data
for e in entries:
row = [e.entry_id, e.composition.reduced_formula,
e.data["spacegroup"]["symbol"],
e.energy_per_atom, e.data["e_above_hull"]]
row += [e.data[s] for s in args.data]
t.append(row)
t = sorted(t, key=lambda x: x[headers.index("E above hull (eV)")])
print(tabulate(t, headers=headers, tablefmt="pipe", floatfmt=".3f"))
def test_substrate_structure_operator():
op = SubstrateStructureQuery()
assert op.query(film_orientation="0,1, 1",
substrate_orientation="1, 0,1") == {
"criteria": {
"film_orient": "0 1 1",
"orient": "1 0 1"
}
}
with ScratchDir("."):
dumpfn(op, "temp.json")
new_op = loadfn("temp.json")
assert new_op.query(film_orientation="0,1, 1",
substrate_orientation="1, 0,1") == {
"criteria": {
"film_orient": "0 1 1",
"orient": "1 0 1"
}
}
def test_to_from_dict(self):
d = self.PxIon.as_dict()
ion_entry = self.PxIon.from_dict(d)
self.assertEqual(ion_entry.entry.name, "MnO4[-1]", "Wrong Entry!")
d = self.PxSol.as_dict()
sol_entry = self.PxSol.from_dict(d)
self.assertEqual(sol_entry.name, "Mn2O3(s)", "Wrong Entry!")
self.assertEqual(
sol_entry.energy,
self.PxSol.energy,
"as_dict and from_dict energies unequal",
)
# Ensure computed entry data persists
entry = ComputedEntry("TiO2", energy=-20, data={"test": "test"})
pbx_entry = PourbaixEntry(entry=entry)
with ScratchDir("."):
dumpfn(pbx_entry, "pbx_entry.json")
reloaded = loadfn("pbx_entry.json")
self.assertIsInstance(reloaded.entry, ComputedEntry)
self.assertIsNotNone(reloaded.entry.data)
def pmg_dump(obj, filename, **kwargs):
"""
Dump an object to a json file using MontyEncoder. Note that these
objects can be lists, dicts or otherwise nested pymatgen objects that
support the as_dict() and from_dict MSONable protocol.
Args:
obj (object): Object to dump.
filename (str): Filename of file to open. Can be gzipped or bzipped.
\*\*kwargs: Any of the keyword arguments supported by the json.dump
method.
"""
return dumpfn(obj, filename, **kwargs)
def run_task(self, fw_spec):
transformations = []
transformation_params = self.get(
"transformation_params",
[{} for i in range(len(self["transformations"]))])
for t in self["transformations"]:
found = False
for m in [
"advanced_transformations", "defect_transformations",
"site_transformations", "standard_transformations"
]:
mod = import_module("pymatgen.transformations.{}".format(m))
try:
t_cls = getattr(mod, t)
except AttributeError:
continue
t_obj = t_cls(**transformation_params.pop(0))
transformations.append(t_obj)
found = True
if not found:
raise ValueError("Could not find transformation: {}".format(t))
# TODO: @matk86 - should prev_calc_dir use CONTCAR instead of POSCAR? Note that if
# current dir, maybe it is POSCAR indeed best ... -computron
structure = self['structure'] if not self.get('prev_calc_dir', None) else \
Poscar.from_file(os.path.join(self['prev_calc_dir'], 'POSCAR')).structure
ts = TransformedStructure(structure)
transmuter = StandardTransmuter([ts], transformations)
final_structure = transmuter.transformed_structures[
-1].final_structure.copy()
vis_orig = self["vasp_input_set"]
vis_dict = vis_orig.as_dict()
vis_dict["structure"] = final_structure.as_dict()
vis_dict.update(self.get("override_default_vasp_params", {}) or {})
vis = vis_orig.__class__.from_dict(vis_dict)
vis.write_input(".")
dumpfn(transmuter.transformed_structures[-1], "transformations.json")
def test_xas_task_id_operator():
op = XASTaskIDQuery()
assert op.query(task_ids="mp-149, mp-13") == {
"criteria": {
"task_id": {
"$in": ["mp-149", "mp-13"]
}
}
}
with ScratchDir("."):
dumpfn(op, "temp.json")
new_op = loadfn("temp.json")
assert new_op.query(task_ids="mp-149, mp-13") == {
"criteria": {
"task_id": {
"$in": ["mp-149", "mp-13"]
}
}
}
def test_path_finding(self):
molecule_entries = loadfn(
os.path.join(test_dir, "ronalds_MoleculeEntry.json"))
li_plus_mol_entry = find_mol_entry_from_xyz_and_charge(
molecule_entries, (os.path.join(test_dir, "Li.xyz")), 1)
ec_mol_entry = find_mol_entry_from_xyz_and_charge(
molecule_entries, (os.path.join(test_dir, "EC.xyz")), 0)
ledc_mol_entry = find_mol_entry_from_xyz_and_charge(
molecule_entries, (os.path.join(test_dir, "LEDC.xyz")), 0)
result = path_finding_wrapper(molecule_entries,
[li_plus_mol_entry, ec_mol_entry],
ledc_mol_entry)
dumpfn(result, "/tmp/lol")
result_canonicalized = loadfn("/tmp/lol")
expected = loadfn(os.path.join(test_dir, "ronalds_PRs.json"))
assert result_canonicalized == expected
def _get_mpid_cache(self):
path = os.path.join(os.path.dirname(module_path), "mpid_cache.json")
if os.path.isfile(path):
mpid_cache = loadfn(path)
else:
with MPRester() as mpr:
# restrict random mpids to those likely experimentally known
# and not too large
entries = mpr.query(
{"nsites": {"$lte": 16}},
["task_id", "icsd_ids"],
chunk_size=0,
mp_decode=False,
)
mpid_cache = [
entry["task_id"] for entry in entries if len(entry["icsd_ids"]) > 2
]
dumpfn(mpid_cache, path)
self.mpid_cache = mpid_cache
def pmg_dump(obj, filename, **kwargs):
"""
Dump an object to a json file using MontyEncoder. Note that these
objects can be lists, dicts or otherwise nested pymatgen objects that
support the as_dict() and from_dict MSONable protocol.
Args:
obj (object): Object to dump.
filename (str): Filename of file to open. Can be gzipped or bzipped.
\*\*kwargs: Any of the keyword arguments supported by the json.dump
method.
"""
return dumpfn(obj, filename, **kwargs)
def test_run_builder(mongostore):
memorystore = MemoryStore("temp")
builder = CopyBuilder(mongostore, memorystore)
mongostore.update([{
mongostore.key: i,
mongostore.last_updated_field: datetime.utcnow()
} for i in range(10)])
runner = CliRunner()
with runner.isolated_filesystem():
dumpfn(builder, "test_builder.json")
result = runner.invoke(run, ["-v", "test_builder.json"])
assert result.exit_code == 0
assert "CopyBuilder" in result.output
assert "SerialProcessor" in result.output
result = runner.invoke(run, ["-v", "-n", "2", "test_builder.json"])
assert result.exit_code == 0
assert "CopyBuilder" in result.output
assert "MultiProcessor" in result.output
def test_possible_oxi_state_query():
op = PossibleOxiStateQuery()
assert op.query(possible_species="Cr2+, O2-") == {
"criteria": {
"possible_species": {
"$all": ["Cr2+", "O2-"]
}
}
}
with ScratchDir("."):
dumpfn(op, "temp.json")
new_op = loadfn("temp.json")
assert op.query(possible_species="Cr2+, O2-") == {
"criteria": {
"possible_species": {
"$all": ["Cr2+", "O2-"]
}
}
}
def do_query(args):
m = MPRester()
try:
criteria = json.loads(args.criteria)
except json.decoder.JSONDecodeError:
criteria = args.criteria
if args.structure:
count = 0
for d in m.query(criteria, properties=["structure", "task_id"]):
s = d["structure"]
formula = re.sub(r"\s+", "", s.formula)
if args.structure == "poscar":
fname = "POSCAR.%s_%s" % (d["task_id"], formula)
else:
fname = "%s-%s.%s" % (d["task_id"], formula, args.structure)
s.to(filename=fname)
count += 1
print("%d structures written!" % count)
elif args.entries:
entries = m.get_entries(criteria)
dumpfn(entries, args.entries)
print("%d entries written to %s!" % (len(entries), args.entries))
else:
props = ["e_above_hull", "spacegroup"]
props += args.data
entries = m.get_entries(criteria, property_data=props)
t = []
headers = ["mp-id", "Formula", "Spacegroup", "E/atom (eV)",
"E above hull (eV)"] + args.data
for e in entries:
row = [e.entry_id, e.composition.reduced_formula,
e.data["spacegroup"]["symbol"],
e.energy_per_atom, e.data["e_above_hull"]]
row += [e.data[s] for s in args.data]
t.append(row)
t = sorted(t, key=lambda x: x[headers.index("E above hull (eV)")])
print(tabulate(t, headers=headers, tablefmt="pipe", floatfmt=".3f"))
def test_dumpfn_loadfn(self):
d = {"hello": "world"}
dumpfn(d, "monte_test.json", indent=4)
d2 = loadfn("monte_test.json")
self.assertEqual(d, d2)
os.remove("monte_test.json")
dumpfn(d, "monte_test.yaml", default_flow_style=False)
d2 = loadfn("monte_test.yaml")
self.assertEqual(d, d2)
dumpfn(d, "monte_test.yaml", Dumper=Dumper)
d2 = loadfn("monte_test.yaml")
os.remove("monte_test.yaml")
dumpfn(d, "monte_test.mpk")
d2 = loadfn("monte_test.mpk")
self.assertEqual(d, {k.decode('utf-8'): v.decode('utf-8') for k, v in d2.items()})
os.remove("monte_test.mpk")
def test_dumpf_loadf(self):
d = {"hello": "world"}
dumpfn(d, "monte_test.json", indent=4)
d2 = loadfn("monte_test.json")
self.assertEqual(d, d2)
os.remove("monte_test.json")
dumpfn(d, "monte_test.yaml", default_flow_style=False)
d2 = loadfn("monte_test.yaml")
self.assertEqual(d, d2)
dumpfn(d, "monte_test.yaml", Dumper=Dumper)
d2 = loadfn("monte_test.yaml")
os.remove("monte_test.yaml")
def write_config(path=None):
path = os.path.join(os.path.expanduser('~'), ".fireworks", 'FW_config.yaml') if path is None else path
dumpfn(config_to_dict(), path)
def solute_def_parse_energy(args):
mpid = args.mpid
solute = args.solute
mapi_key = args.mapi_key
if not mpid:
print ("============\nERROR: Provide an mpid\n============")
return
if not solute:
print ("============\nERROR: Provide solute element\n============")
return
if not mapi_key:
with MPRester() as mp:
structure = mp.get_structure_by_material_id(mpid)
else:
with MPRester(mapi_key) as mp:
structure = mp.get_structure_by_material_id(mpid)
energy_dict = {}
solutes = []
def_folders = glob.glob(os.path.join(
mpid,"solute*subspecie-{}".format(solute)))
def_folders += glob.glob(os.path.join(mpid,"bulk"))
for defdir in def_folders:
fldr_name = os.path.split(defdir)[1]
vr_file = os.path.join(defdir,'vasprun.xml')
if not os.path.exists(vr_file):
print (fldr_name, ": vasprun.xml doesn't exist in the folder. " \
"Abandoning parsing of energies for {}".format(mpid))
break # Further processing for the mpid is not useful
try:
vr = Vasprun(vr_file)
except:
print (fldr_name, ":Failure, couldn't parse vaprun.xml file. "
"Abandoning parsing of energies for {}".format(mpid))
break
if not vr.converged:
print (fldr_name, ": Vasp calculation not converged. "
"Abandoning parsing of energies for {}".format(mpid))
break # Further processing for the mpid is not useful
fldr_fields = fldr_name.split("_")
if 'bulk' in fldr_fields:
bulk_energy = vr.final_energy
bulk_sites = vr.structures[-1].num_sites
elif 'solute' in fldr_fields:
site_index = int(fldr_fields[1])
site_multiplicity = int(fldr_fields[2].split("-")[1])
site_specie = fldr_fields[3].split("-")[1]
substitution_specie = fldr_fields[4].split("-")[1]
energy = vr.final_energy
solutes.append({'site_index':site_index,
'site_specie':site_specie,'energy':energy,
'substitution_specie':substitution_specie,
'site_multiplicity':site_multiplicity
})
else:
if not solutes:
print("Solute folders do not exist")
return {}
print("Solute {} calculations successful for {}".format(solute,mpid))
for solute in solutes:
solute_flip_energy = solute['energy']-bulk_energy
solute['energy'] = solute_flip_energy
solutes.sort(key=lambda entry: entry['site_index'])
energy_dict[mpid] = {'solutes':solutes}
fl_nm = mpid+'_solute-'+args.solute+'_raw_defect_energy.json'
dumpfn(energy_dict, fl_nm, indent=2, cls=MontyEncoder)
def vac_antisite_def_parse_energy(args):
mpid = args.mpid
mapi_key = args.mapi_key
if not mpid:
print("============\nERROR: Provide an mpid\n============")
return
if not mapi_key:
with MPRester() as mp:
structure = mp.get_structure_by_material_id(mpid)
else:
with MPRester(mapi_key) as mp:
structure = mp.get_structure_by_material_id(mpid)
energy_dict = {}
antisites = []
vacancies = []
def_folders = glob.glob(os.path.join(mpid,"vacancy*"))
def_folders += glob.glob(os.path.join(mpid,"antisite*"))
def_folders += glob.glob(os.path.join(mpid,"bulk"))
for defdir in def_folders:
fldr_name = os.path.split(defdir)[1]
vr_file = os.path.join(defdir,'vasprun.xml')
if not os.path.exists(vr_file):
print (fldr_name, ": vasprun.xml doesn't exist in the folder. " \
"Abandoning parsing of energies for {}".format(mpid))
break # Further processing for the mpid is not useful
try:
vr = Vasprun(vr_file)
except:
print (fldr_name, ":Failure, couldn't parse vaprun.xml file. "
"Abandoning parsing of energies for {}".format(mpid))
break
if not vr.converged:
print (fldr_name, ": Vasp calculation not converged. "
"Abandoning parsing of energies for {}".format(mpid))
break # Further processing for the mpid is not useful
fldr_fields = fldr_name.split("_")
if 'bulk' in fldr_fields:
bulk_energy = vr.final_energy
bulk_sites = vr.structures[-1].num_sites
elif 'vacancy' in fldr_fields:
site_index = int(fldr_fields[1])
site_multiplicity = int(fldr_fields[2].split("-")[1])
site_specie = fldr_fields[3].split("-")[1]
energy = vr.final_energy
vacancies.append({'site_index':site_index,
'site_specie':site_specie,'energy':energy,
'site_multiplicity':site_multiplicity
})
elif 'antisite' in fldr_fields:
site_index = int(fldr_fields[1])
site_multiplicity = int(fldr_fields[2].split("-")[1])
site_specie = fldr_fields[3].split("-")[1]
substitution_specie = fldr_fields[4].split("-")[1]
energy = vr.final_energy
antisites.append({'site_index':site_index,
'site_specie':site_specie,'energy':energy,
'substitution_specie':substitution_specie,
'site_multiplicity':site_multiplicity
})
else:
print("All calculations successful for ", mpid)
e0 = bulk_energy/bulk_sites*structure.num_sites
for vac in vacancies:
vac_flip_energy = vac['energy']-bulk_energy
vac['energy'] = vac_flip_energy
vacancies.sort(key=lambda entry: entry['site_index'])
for antisite in antisites:
as_flip_energy = antisite['energy']-bulk_energy
antisite['energy'] = as_flip_energy
antisites.sort(key=lambda entry: entry['site_index'])
energy_dict[str(mpid)] = {u"structure":structure,
'e0':e0,'vacancies':vacancies,'antisites':antisites}
fl_nm = args.mpid+'_raw_defect_energy.json'
dumpfn(energy_dict, fl_nm, cls=MontyEncoder, indent=2)
def run_interrupted(self):
"""
Runs custodian in a interuppted mode, which sets up and
validates jobs but doesn't run the executable
Returns:
number of remaining jobs
Raises:
CustodianError on unrecoverable errors, and jobs that fail
validation
"""
try:
cwd = os.getcwd()
start = datetime.datetime.now()
v = sys.version.replace("\n", " ")
logger.info("Custodian started in singleshot mode at {} in {}."
.format(start, cwd))
logger.info("Custodian running on Python version {}".format(v))
# load run log
if os.path.exists(Custodian.LOG_FILE):
self.run_log = loadfn(Custodian.LOG_FILE, cls=MontyDecoder)
if len(self.run_log) == 0:
# starting up an initial job - setup input and quit
job_n = 0
job = self.jobs[job_n]
logger.info("Setting up job no. 1 ({}) ".format(job.name))
job.setup()
self.run_log.append({"job": job.as_dict(), "corrections": [], 'job_n': job_n})
return len(self.jobs)
else:
# Continuing after running calculation
job_n = self.run_log[-1]['job_n']
job = self.jobs[job_n]
# If we had to fix errors from a previous run, insert clean log
# dict
if len(self.run_log[-1]['corrections']) > 0:
logger.info("Reran {}.run due to fixable errors".format(job.name))
# check error handlers
logger.info("Checking error handlers for {}.run".format(job.name))
if self._do_check(self.handlers):
logger.info("Failed validation based on error handlers")
# raise an error for an unrecoverable error
for x in self.run_log[-1]["corrections"]:
if not x["actions"] and x["handler"].raises_runtime_error:
s = "Unrecoverable error for handler: {}. " \
"Raising RuntimeError".format(x["handler"])
raise CustodianError(s, True, x["handler"])
logger.info("Corrected input based on error handlers")
# Return with more jobs to run if recoverable error caught
# and corrected for
return len(self.jobs) - job_n
# check validators
logger.info("Checking validator for {}.run".format(job.name))
for v in self.validators:
if v.check():
logger.info("Failed validation based on validator")
s = "Validation failed: {}".format(v)
raise CustodianError(s, True, v)
logger.info("Postprocessing for {}.run".format(job.name))
job.postprocess()
# IF DONE WITH ALL JOBS - DELETE ALL CHECKPOINTS AND RETURN
# VALIDATED
if len(self.jobs) == (job_n + 1):
self.finished = True
return 0
# Setup next job_n
job_n += 1
job = self.jobs[job_n]
self.run_log.append({"job": job.as_dict(), "corrections": [],
'job_n': job_n})
job.setup()
return len(self.jobs) - job_n
except CustodianError as ex:
logger.error(ex.message)
if ex.raises:
raise RuntimeError("{} errors reached: {}. Exited..."
.format(self.total_errors, ex))
finally:
#Log the corrections to a json file.
logger.info("Logging to {}...".format(Custodian.LOG_FILE))
dumpfn(self.run_log, Custodian.LOG_FILE, cls=MontyEncoder,
indent=4)
end = datetime.datetime.now()
logger.info("Run ended at {}.".format(end))
run_time = end - start
logger.info("Run completed. Total time taken = {}."
.format(run_time))
if self.finished and self.gzipped_output:
gzip_dir(".")
def test_as_dict(self):
dumpfn(self.entry_set, "temp_entry_set.json")
entry_set = loadfn("temp_entry_set.json")
self.assertEqual(len(entry_set), len(self.entry_set))
os.remove("temp_entry_set.json")
def save(self, fname="Transport_Properties.json"):
dumpfn(self.props_dict, fname)
def update_checkpoint(job_ids=None, jfile=None, **kwargs):
"""
rerun the jobs with job ids in the job_ids list. The jobs are
read from the json checkpoint file, jfile.
If no job_ids are given then the checkpoint file will
be updated with corresponding final energy
Args:
job_ids: list of job ids to update or q resolve
jfile: check point file
"""
cal_log = loadfn(jfile, cls=MontyDecoder)
cal_log_new = []
all_jobs = []
run_jobs = []
handlers = []
final_energy = None
incar = None
kpoints = None
qadapter = None
#if updating the specs of the job
for k, v in kwargs.items():
if k == 'incar':
incar = v
if k == 'kpoints':
kpoints = v
if k == 'que':
qadapter = v
for j in cal_log:
job = j["job"]
job.job_id = j['job_id']
all_jobs.append(job)
if job_ids and (j['job_id'] in job_ids or job.job_dir in job_ids):
logger.info('setting job {0} in {1} to rerun'.format(j['job_id'], job.job_dir))
contcar_file = job.job_dir+os.sep+'CONTCAR'
poscar_file = job.job_dir+os.sep+'POSCAR'
if os.path.isfile(contcar_file) and len(open(contcar_file).readlines()) != 0 :
logger.info('setting poscar file from {}'
.format(contcar_file))
job.vis.poscar = Poscar.from_file(contcar_file)
else:
logger.info('setting poscar file from {}'
.format(poscar_file))
job.vis.poscar = Poscar.from_file(poscar_file)
if incar:
logger.info('incar overridden')
job.vis.incar = incar
if kpoints:
logger.info('kpoints overridden')
job.vis.kpoints = kpoints
if qadapter:
logger.info('qadapter overridden')
job.vis.qadapter = qadapter
run_jobs.append(job)
if run_jobs:
c = Custodian(handlers, run_jobs, max_errors=5)
c.run()
for j in all_jobs:
final_energy = j.get_final_energy()
cal_log_new.append({"job": j.as_dict(),
'job_id': j.job_id,
"corrections": [],
'final_energy': final_energy})
dumpfn(cal_log_new, jfile, cls=MontyEncoder,
indent=4)