class AbstractVaspDirCollParser(six.with_metaclass(abc.ABCMeta, object)):
"""Abstract base class for parsers of a collection of VASP directories
To implement a new parser, inherit from this class and
define the :meth:`compile` method.
"""
def __init__(self, rootdir):
"""read vasp output via drone and extract all data
:param rootdir: root directory containing collection of VASP dirs
:type rootdir: str
"""
from pymatgen.apps.borg.hive import SimpleVaspToComputedEntryDrone
from pymatgen.apps.borg.queen import BorgQueen
self.rootdir = rootdir
self.drone = SimpleVaspToComputedEntryDrone(inc_structure=True)
self.queen = BorgQueen(self.drone, rootdir, 1) # TODO: make sure uw2_si2 also works in parallel
self.data = self.queen.get_data()
def find_entry_for_directory(self, regex, oszicar=True):
"""returns the computed entry for a VASP directory matching the regex"""
# scan in reverse alpha-numeric order under the assumption that
# directories with the highest (local) index correspond to final VaspRun
for entry in reversed(self.data):
if fnmatch.fnmatch(entry.data['filename'], regex):
if oszicar and not entry.energy < 1e10: continue
return entry
@abc.abstractmethod
def compile(self):
"""compile the extracted data into a reduced dataset to be contributed"""
return
def get_energies(rootdir, reanalyze, verbose, detailed, sort):
"""
Doc string.
"""
if verbose:
FORMAT = "%(relativeCreated)d msecs : %(message)s"
logging.basicConfig(level=logging.INFO, format=FORMAT)
if not detailed:
drone = SimpleVaspToComputedEntryDrone(inc_structure=True)
else:
drone = VaspToComputedEntryDrone(inc_structure=True,
data=["filename",
"initial_structure"])
ncpus = multiprocessing.cpu_count()
logging.info("Detected {} cpus".format(ncpus))
queen = BorgQueen(drone, number_of_drones=ncpus)
if os.path.exists(SAVE_FILE) and not reanalyze:
msg = "Using previously assimilated data from {}.".format(SAVE_FILE) \
+ " Use -f to force re-analysis."
queen.load_data(SAVE_FILE)
else:
if ncpus > 1:
queen.parallel_assimilate(rootdir)
else:
queen.serial_assimilate(rootdir)
msg = "Analysis results saved to {} for faster ".format(SAVE_FILE) + \
"subsequent loading."
queen.save_data(SAVE_FILE)
entries = queen.get_data()
if sort == "energy_per_atom":
entries = sorted(entries, key=lambda x: x.energy_per_atom)
elif sort == "filename":
entries = sorted(entries, key=lambda x: x.data["filename"])
all_data = []
for e in entries:
if not detailed:
delta_vol = "{:.2f}".format(e.data["delta_volume"] * 100)
else:
delta_vol = e.structure.volume / \
e.data["initial_structure"].volume - 1
delta_vol = "{:.2f}".format(delta_vol * 100)
all_data.append((e.data["filename"].replace("./", ""),
re.sub("\s+", "", e.composition.formula),
"{:.5f}".format(e.energy),
"{:.5f}".format(e.energy_per_atom),
delta_vol))
if len(all_data) > 0:
headers = ("Directory", "Formula", "Energy", "E/Atom", "% vol chg")
t = PrettyTable(headers)
t.align["Directory"] = "l"
for d in all_data:
t.add_row(d)
print(t)
print(msg)
else:
print("No valid vasp run found.")
def get_energies(rootdir, reanalyze, verbose, pretty):
if verbose:
FORMAT = "%(relativeCreated)d msecs : %(message)s"
logging.basicConfig(level=logging.INFO, format=FORMAT)
drone = GaussianToComputedEntryDrone(inc_structure=True,
parameters=['filename'])
ncpus = multiprocessing.cpu_count()
logging.info('Detected {} cpus'.format(ncpus))
queen = BorgQueen(drone, number_of_drones=ncpus)
if os.path.exists(save_file) and not reanalyze:
msg = 'Using previously assimilated data from {}. ' + \
'Use -f to force re-analysis'.format(save_file)
queen.load_data(save_file)
else:
queen.parallel_assimilate(rootdir)
msg = 'Results saved to {} for faster reloading.'.format(save_file)
queen.save_data(save_file)
entries = queen.get_data()
entries = sorted(entries, key=lambda x: x.parameters['filename'])
all_data = [(e.parameters['filename'].replace("./", ""),
re.sub("\s+", "", e.composition.formula),
"{}".format(e.parameters['charge']),
"{}".format(e.parameters['spin_mult']),
"{:.5f}".format(e.energy), "{:.5f}".format(e.energy_per_atom),
) for e in entries]
headers = ("Directory", "Formula", "Charge", "Spin Mult.", "Energy",
"E/Atom")
print(tabulate(all_data, headers=headers))
print("")
print(msg)
def test_load_data(self):
with warnings.catch_warnings():
warnings.simplefilter("ignore")
drone = VaspToComputedEntryDrone()
queen = BorgQueen(drone)
queen.load_data(os.path.join(test_dir, "assimilated.json"))
self.assertEqual(len(queen.get_data()), 1)
def _extract_MP_data(self,MP_data_filename):
drone = VaspToComputedEntryDrone()
queen = BorgQueen(drone, "dummy", 1)
queen.load_data(MP_data_filename)
computed_entries = queen.get_data()
del drone
del queen
return computed_entries
class BorgQueenTest(unittest.TestCase):
def setUp(self):
drone = VaspToComputedEntryDrone()
self.queen = BorgQueen(drone, test_dir, 1)
def test_get_data(self):
data = self.queen.get_data()
self.assertEqual(len(data), 8)
def test_load_data(self):
drone = VaspToComputedEntryDrone()
queen = BorgQueen(drone)
queen.load_data(os.path.join(test_dir, "assimilated.json"))
self.assertEqual(len(queen.get_data()), 1)
def get_e_v(path):
"""
uses pymatgen drone and borgqueen classes to get energy and
volume data from the given directory path.
"""
volumes = []
energies = []
drone = MPINTVaspDrone(inc_structure=True)
bg = BorgQueen(drone)
# bg.parallel_assimilate(path)
bg.serial_assimilate(path)
allentries = bg.get_data()
for e in allentries:
if e:
energies.append(e.energy)
volumes.append(e.structure.lattice.volume)
return (volumes, energies)
def extract_json_data():
"""
Routine tries to read VASP data into pymatgen objects, and
then extracts only the relevant data. This is then written to json,
allowing the voluminous OUTCAR and vasprun.xml files to be discarded.
"""
try:
o = Outcar('OUTCAR')
found_outcar = True
except:
print('OUTCAR file missing or not readable')
found_outcar = False
try:
vr = Vasprun('vasprun.xml')
found_vasprun = True
except:
print('vasprun.xml file missing or not readable')
found_vasprun = False
dictionary_data = {}
if found_outcar:
dictionary_data['OUTCAR'] = o.as_dict()
if found_vasprun:
try:
# try to extract a Computed Entry object, using
# pymatgen technology
drone = VaspToComputedEntryDrone()
queen = BorgQueen(drone, './', 1)
entry = queen.get_data()[0]
dictionary_data['ComputedEntry'] = entry.as_dict()
except:
print('ComputedEntry COULD NOT BE EXTRACTED BY PYMATGEN...')
"""
# Do not extract DOS in run_data.json; this is too memory intensive!
try:
dictionary_data['DOS'] = vr.complete_dos.as_dict()
pymatgen_dos_success = True
except:
print('DOS COULD NOT BE EXTRACTED BY PYMATGEN...')
pymatgen_dos_success = False
"""
relaxation_data = []
for step in vr.ionic_steps:
data_dict = {}
for key in ['forces','structure','stress']:
if key in step:
data_dict[key] = step[key]
data_dict['electronic'] = step['electronic_steps'][-1]
relaxation_data.append(data_dict)
dictionary_data['relaxation'] = relaxation_data
if found_outcar or found_vasprun:
pmg_dump(dictionary_data, 'run_data.json')
return
def get_energies(rootdir, reanalyze, verbose, detailed,
sort, formulaunit, debug, hull, threshold, args, templatestructure):
ion_list = 'Novalue'
ave_key_list = 'Novalue'
threscount = 0
"""
Doc string.
"""
if (verbose and not debug):
FORMAT = "%(relativeCreated)d msecs : %(message)s"
logging.basicConfig(level=logging.INFO, format=FORMAT)
elif debug:
logging.basicConfig(level=logging.DEBUG)
if not detailed:
drone = SimpleVaspToComputedEntryDrone(inc_structure=True)
else:
drone = VaspToComputedEntryDrone(inc_structure=True,
data=["filename",
"initial_structure"])
ncpus = multiprocessing.cpu_count()
logging.info("Detected {} cpus".format(ncpus))
queen = BorgQueen(drone, number_of_drones=ncpus)
if os.path.exists(SAVE_FILE) and not reanalyze:
msg = "Using previously assimilated data from {}.".format(SAVE_FILE) \
+ " Use -f to force re-analysis."
queen.load_data(SAVE_FILE)
else:
if ncpus > 1:
queen.parallel_assimilate(rootdir)
else:
queen.serial_assimilate(rootdir)
msg = "Analysis results saved to {} for faster ".format(SAVE_FILE) + \
"subsequent loading."
queen.save_data(SAVE_FILE)
entries = queen.get_data()
if sort == "energy_per_atom":
entries = sorted(entries, key=lambda x: x.energy_per_atom)
elif sort == "filename":
entries = sorted(entries, key=lambda x: x.data["filename"])
# logging.debug('First Energy entry is {}'.format(entries[0]))
base_energy = entries[0].energy
logging.debug('Type of entries is: {}'.format(type(entries)))
logging.debug('First Element of Entries is:{}'.format(entries[0]))
# logging.debug('First Energy entry structure is {}'.format(entries[0].structure))
xy_direction = int(args.XYdirection)
tolerance = float(args.tolerance)
if args.template:
logging.debug('Temp Structure site info is: {}'.format(Na12(['Co','Mn'],['Na'],templatestructure,templatestructure,XY_Direction=xy_direction,tol=tolerance)))
template_site_info = Na12(['Co','Mn'],['Na'],templatestructure,templatestructure,XY_Direction=xy_direction,tol=tolerance)
all_data = []
energy_diff = []
threshold=float(threshold)
Structure_info_dict = {}
check_ion_seq = [args.dupion]
for e in entries:
if not detailed:
delta_vol = "{:.2f}".format(e.data["delta_volume"] * 100)
else:
delta_vol = e.structure.volume / \
e.data["initial_structure"].volume - 1
delta_vol = "{:.2f}".format(delta_vol * 100)
entry_path = e.data['filename'].rsplit('/',1)[0]
entry_site_info = Na12(['Co','Mn'],['Na'],e.structure,e.structure,XY_Direction=xy_direction,tol=tolerance)
logging.debug('Total Na site: {}'.format(entry_site_info['Total_Na_Site']))
#Coordination extraction part
# na_layer_site_fcoords = [site._fcoords for site in s if site.specie.symbol == "Na"]
# if 'Cif_Structure' in e.data.keys():
# na_sites_fcoords = [site._fcoords for site in e.data['Cif_Structure'] if site.specie.symbol == 'Na']
# na_sites_fcoords_list_tuple = [tuple(coord) for coord in na_sites_fcoords]
na_sites_fcoords = [site._fcoords for site in e.data['CONTCAR_Structure'] if site.specie.symbol == 'Na']
na_sites_fcoords_list_tuple = [tuple(coord) for coord in na_sites_fcoords]
if args.nupdown:
entry_data= [rootdir,e.data["filename"].replace("./", ""),
re.sub("\s+", "", e.composition.formula),
"{:.5f}".format(e.energy),
"{:.5f}".format(1000*(e.energy-base_energy)/int(formulaunit)),
"{:.5f}".format(e.energy_per_atom),
delta_vol,e.parameters['run_type'],
e.data['NUPDOWN'],e.data['ISMEAR'],na_sites_fcoords_list_tuple]
else:
entry_data= [rootdir,e.data["filename"].replace("./", ""),
re.sub("\s+", "", e.composition.formula),
"{:.5f}".format(e.energy),
"{:.5f}".format(1000*(e.energy-base_energy)/int(formulaunit)),
"{:.5f}".format(e.energy_per_atom),
delta_vol,e.parameters['run_type'],na_sites_fcoords_list_tuple]
if args.structure:
entry_data.extend([entry_site_info['Total_Na_Site'],entry_site_info['Na2_Site'],entry_site_info['Na1_Mn_Site'],
entry_site_info['Na1_Co_Site'],entry_site_info['Na1_Mn_Co_Site']])
if args.template:
entry_data.extend([template_site_info['Total_Na_Site'],template_site_info['Na2_Site'],template_site_info['Na1_Mn_Site'],
template_site_info['Na1_Co_Site'],template_site_info['Na1_Mn_Co_Site']])
# sitelist = ['Existed','Duplicate_Entry']
logging.debug(e.data)
if args.duplicate:
# filename.rsplit('/',2)[-2]
Duplicate, Duplicat_Entry, Structure_info_dict = check_ex(check_ion_seq,Structure_info_dict,
e,args.tolerance)
entry_data.extend([Duplicate,Duplicat_Entry])
if args.ion_list:
if args.ion_list[0] == "All":
ion_list = None
else:
(start, end) = [int(i) for i in re.split("-", args.ion_list[0])]
ion_list = list(range(start, end + 1))
for d in entry_path:
magdata = get_magnetization(d, ion_list)
entry_data.extend(magdata)
if args.ion_avg_list:
ave_mag_data, ave_key_list = get_ave_magnetization(entry_path,args.ion_avg_list)
entry_data.extend(ave_mag_data)
if threshold != 0:
all_data.append(entry_data)
if float(entry_data[4])<threshold:
threscount +=1
elif threshold == 0:
all_data.append(entry_data)
energy_diff.append("{:.5f}".format(1000*(e.energy-base_energy)/int(formulaunit)))
# if len(all_data) > 0:
# headers = ("Directory", "Formula", "Energy", "Energy Diff (meV)/F.U.","E/Atom", "% vol chg")
# t = PrettyTable(headers)
# t.align["Directory"] = "l"
# for d in all_data:
# logging.debug('data row in all data is: \n {}'.format(d))
# t.add_row(d)
# print(t)
# print(msg)
# else:
# print("No valid vasp run found.")
if hull:
print 'Analyzing group: {}\n'.format(rootdir)
print 'Energy above hull is: \n'
print map(lambda x: x.encode('ascii'), energy_diff)
logging.info('In group: {}, number of entries fall in threshold is {}'.format(rootdir,threscount))
all_data.append([])
return all_data
def test_load_data(self):
drone = VaspToComputedEntryDrone()
queen = BorgQueen(drone)
queen.load_data(os.path.join(test_dir, "assimilated.json"))
self.assertEqual(len(queen.get_data()), 1)
def test_assimilate(self):
"""Borg assimilation code.
This takes too long for a unit test!
"""
simulate = True if VaspToDbTaskDroneTest.conn is None else False
drone = VaspToDbTaskDrone(database="creator_unittest",
simulate_mode=simulate,
parse_dos=True, compress_dos=1)
queen = BorgQueen(drone)
queen.serial_assimilate(os.path.join(test_dir, 'db_test'))
data = queen.get_data()
self.assertEqual(len(data), 6)
if VaspToDbTaskDroneTest.conn:
db = VaspToDbTaskDroneTest.conn["creator_unittest"]
data = db.tasks.find()
self.assertEqual(data.count(), 6)
warnings.warn("Actual db insertion mode.")
for d in data:
dir_name = d['dir_name']
if dir_name.endswith("killed_mp_aflow"):
self.assertEqual(d['state'], "killed")
self.assertFalse(d['is_hubbard'])
self.assertEqual(d['pretty_formula'], "SiO2")
elif dir_name.endswith("stopped_mp_aflow"):
self.assertEqual(d['state'], "stopped")
self.assertEqual(d['pretty_formula'], "ThFe5P3")
elif dir_name.endswith("success_mp_aflow"):
self.assertEqual(d['state'], "successful")
self.assertEqual(d['pretty_formula'], "TbZn(BO2)5")
self.assertAlmostEqual(d['output']['final_energy'],
-526.66747274, 4)
elif dir_name.endswith("Li2O_aflow"):
self.assertEqual(d['state'], "successful")
self.assertEqual(d['pretty_formula'], "Li2O")
self.assertAlmostEqual(d['output']['final_energy'],
-14.31446494, 6)
self.assertEqual(len(d["calculations"]), 2)
self.assertEqual(d['input']['is_lasph'], False)
self.assertEqual(d['input']['xc_override'], None)
self.assertEqual(d["oxide_type"], "oxide")
elif dir_name.endswith("Li2O"):
self.assertEqual(d['state'], "successful")
self.assertEqual(d['pretty_formula'], "Li2O")
self.assertAlmostEqual(d['output']['final_energy'],
-14.31337758, 6)
self.assertEqual(len(d["calculations"]), 1)
self.assertEqual(len(d["custodian"]), 1)
self.assertEqual(len(d["custodian"][0]["corrections"]), 1)
elif dir_name.endswith("Li2O_aflow_lasph"):
self.assertEqual(d['state'], "successful")
self.assertEqual(d['pretty_formula'], "Li2O")
self.assertAlmostEqual(d['output']['final_energy'],
-13.998171, 6)
self.assertEqual(len(d["calculations"]), 2)
self.assertEqual(d['input']['is_lasph'], True)
self.assertEqual(d['input']['xc_override'], "PS")
if VaspToDbTaskDroneTest.conn:
warnings.warn("Testing query engine mode.")
qe = QueryEngine(database="creator_unittest")
self.assertEqual(qe.query().count(), 6)
#Test mappings by query engine.
for r in qe.query(criteria={"pretty_formula": "Li2O"},
properties=["dir_name", "energy",
"calculations", "input"]):
if r["dir_name"].endswith("Li2O_aflow"):
self.assertAlmostEqual(r['energy'], -14.31446494, 4)
self.assertEqual(len(r["calculations"]), 2)
self.assertEqual(r["input"]["is_lasph"], False)
self.assertEqual(r['input']['xc_override'], None)
self.assertEqual(d["oxide_type"], "oxide")
elif r["dir_name"].endswith("Li2O"):
self.assertAlmostEqual(r['energy'],
-14.31337758, 4)
self.assertEqual(len(r["calculations"]), 1)
self.assertEqual(r["input"]["is_lasph"], False)
self.assertEqual(r['input']['xc_override'], None)
# Test lasph
e = qe.get_entries({"dir_name":{"$regex":"lasph"}})
self.assertEqual(len(e), 1)
self.assertEqual(e[0].parameters["is_lasph"], True)
self.assertEqual(e[0].parameters["xc_override"], "PS")
# Test query one.
d = qe.query_one(criteria={"pretty_formula": "TbZn(BO2)5"},
properties=["energy"])
self.assertAlmostEqual(d['energy'], -526.66747274, 4)
d = qe.get_entries_in_system(["Li", "O"])
self.assertEqual(len(d), 3)
self.assertIsInstance(d[0], ComputedEntry)
self.assertEqual(d[0].data["oxide_type"], "oxide")
s = qe.get_structure_from_id(d[0].entry_id)
self.assertIsInstance(s, Structure)
self.assertEqual(s.formula, "Li2 O1")
self.assertIsInstance(qe.get_dos_from_id(d[0].entry_id), CompleteDos)
def submit_vasp_directory(self, rootdir, authors, projects=None,
references='', remarks=None, master_data=None,
master_history=None, created_at=None,
ncpus=None):
"""
Assimilates all vasp run directories beneath a particular
directory using BorgQueen to obtain structures, and then submits thhem
to the Materials Project as SNL files. VASP related meta data like
initial structure and final energies are automatically incorporated.
.. note::
As of now, this MP REST feature is open only to a select group of
users. Opening up submissions to all users is being planned for
the future.
Args:
rootdir:
Rootdir to start assimilating VASP runs from.
authors:
*List* of {"name":'', "email":''} dicts,
*list* of Strings as 'John Doe <*****@*****.**>',
or a single String with commas separating authors. The same
list of authors should apply to all runs.
projects:
List of Strings ['Project A', 'Project B']. This applies to
all structures.
references:
A String in BibTeX format. Again, this applies to all
structures.
remarks:
List of Strings ['Remark A', 'Remark B']
masterdata:
A free form dict. Namespaced at the root level with an
underscore, e.g. {"_materialsproject":<custom data>}. This
data is added to all structures detected in the directory,
in addition to other vasp data on a per structure basis.
created_at:
A datetime object
ncpus:
Number of cpus to use in using BorgQueen to assimilate
"""
drone = VaspToComputedEntryDrone(inc_structure=True,
data=["filename",
"initial_structure"])
queen = BorgQueen(drone, number_of_drones=ncpus)
queen.parallel_assimilate(rootdir)
structures = []
metadata = []
# TODO: Get histories from the data.
for e in queen.get_data():
structures.append(e.structure)
m = {
"_vasp": {
"parameters": e.parameters,
"final_energy": e.energy,
"final_energy_per_atom": e.energy_per_atom,
"initial_structure": e.data["initial_structure"].to_dict
}
}
if master_data is not None:
m.update(master_data)
metadata.append(m)
histories = None
if master_history is not None:
histories = master_history * len(structures)
return self.submit_structures(
structures, authors, projects=projects, references=references,
remarks=remarks, data=metadata, histories=histories,
created_at=created_at)
def test_assimilate(self):
simulate = True if VaspToDbTaskDroneTest.conn is None else False
drone = VaspToDbTaskDrone(database="creator_unittest",
simulate_mode=simulate)
queen = BorgQueen(drone)
queen.serial_assimilate(os.path.join(test_dir, 'db_test'))
data = queen.get_data()
self.assertEqual(len(data), 5)
if VaspToDbTaskDroneTest.conn:
db = VaspToDbTaskDroneTest.conn["creator_unittest"]
data = db.tasks.find()
self.assertEqual(data.count(), 5)
warnings.warn("Actual db insertion mode.")
for d in data:
dir_name = d['dir_name']
if dir_name.endswith("killed_mp_aflow"):
self.assertEqual(d['state'], "killed")
self.assertFalse(d['is_hubbard'])
self.assertEqual(d['pretty_formula'], "SiO2")
elif dir_name.endswith("stopped_mp_aflow"):
self.assertEqual(d['state'], "stopped")
self.assertEqual(d['pretty_formula'], "ThFe5P3")
elif dir_name.endswith("success_mp_aflow"):
self.assertEqual(d['state'], "successful")
self.assertEqual(d['pretty_formula'], "TbZn(BO2)5")
self.assertAlmostEqual(d['output']['final_energy'],
-526.66747274, 4)
elif dir_name.endswith("Li2O_aflow"):
self.assertEqual(d['state'], "successful")
self.assertEqual(d['pretty_formula'], "Li2O")
self.assertAlmostEqual(d['output']['final_energy'],
-14.31446494, 6)
self.assertEqual(len(d["calculations"]), 2)
elif dir_name.endswith("Li2O"):
self.assertEqual(d['state'], "successful")
self.assertEqual(d['pretty_formula'], "Li2O")
self.assertAlmostEqual(d['output']['final_energy'],
-14.31337758, 6)
self.assertEqual(len(d["calculations"]), 1)
self.assertEqual(len(d["custodian"]), 1)
self.assertEqual(len(d["custodian"][0]["corrections"]), 1)
if VaspToDbTaskDroneTest.conn:
warnings.warn("Testing query engine mode.")
qe = QueryEngine(database="creator_unittest")
self.assertEqual(qe.query().count(), 5)
#Test mappings by query engine.
for r in qe.query(criteria={"pretty_formula": "Li2O"},
properties=["dir_name", "energy",
"calculations"]):
if r["dir_name"].endswith("Li2O_aflow"):
self.assertAlmostEqual(r['energy'], -14.31446494, 4)
self.assertEqual(len(r["calculations"]), 2)
elif r["dir_name"].endswith("Li2O"):
self.assertAlmostEqual(r['energy'],
-14.31337758, 4)
self.assertEqual(len(r["calculations"]), 1)
# Test query one.
d = qe.query_one(criteria={"pretty_formula": "TbZn(BO2)5"},
properties=["energy"])
self.assertAlmostEqual(d['energy'], -526.66747274, 4)
d = qe.get_entries_in_system(["Li", "O"])
self.assertEqual(len(d), 2)
self.assertIsInstance(d[0], ComputedEntry)
s = qe.get_structure_from_id(d[0].entry_id)
self.assertIsInstance(s, Structure)
self.assertEqual(s.formula, "Li2 O1")
def test_assimilate(self):
"""Borg assimilation code.
This takes too long for a unit test!
"""
simulate = True if VaspToDbTaskDroneTest.conn is None else False
drone = VaspToDbTaskDrone(database="creator_unittest",
simulate_mode=simulate,
parse_dos=True,
compress_dos=1)
queen = BorgQueen(drone)
queen.serial_assimilate(os.path.join(test_dir, 'db_test'))
data = queen.get_data()
self.assertEqual(len(data), 6)
if VaspToDbTaskDroneTest.conn:
db = VaspToDbTaskDroneTest.conn["creator_unittest"]
data = db.tasks.find()
self.assertEqual(data.count(), 6)
warnings.warn("Actual db insertion mode.")
for d in data:
dir_name = d['dir_name']
if dir_name.endswith("killed_mp_aflow"):
self.assertEqual(d['state'], "killed")
self.assertFalse(d['is_hubbard'])
self.assertEqual(d['pretty_formula'], "SiO2")
elif dir_name.endswith("stopped_mp_aflow"):
self.assertEqual(d['state'], "stopped")
self.assertEqual(d['pretty_formula'], "ThFe5P3")
elif dir_name.endswith("success_mp_aflow"):
self.assertEqual(d['state'], "successful")
self.assertEqual(d['pretty_formula'], "TbZn(BO2)5")
self.assertAlmostEqual(d['output']['final_energy'],
-526.66747274, 4)
elif dir_name.endswith("Li2O_aflow"):
self.assertEqual(d['state'], "successful")
self.assertEqual(d['pretty_formula'], "Li2O")
self.assertAlmostEqual(d['output']['final_energy'],
-14.31446494, 6)
self.assertEqual(len(d["calculations"]), 2)
self.assertEqual(d['input']['is_lasph'], False)
self.assertEqual(d['input']['xc_override'], None)
elif dir_name.endswith("Li2O"):
self.assertEqual(d['state'], "successful")
self.assertEqual(d['pretty_formula'], "Li2O")
self.assertAlmostEqual(d['output']['final_energy'],
-14.31337758, 6)
self.assertEqual(len(d["calculations"]), 1)
self.assertEqual(len(d["custodian"]), 1)
self.assertEqual(len(d["custodian"][0]["corrections"]), 1)
elif dir_name.endswith("Li2O_aflow_lasph"):
self.assertEqual(d['state'], "successful")
self.assertEqual(d['pretty_formula'], "Li2O")
self.assertAlmostEqual(d['output']['final_energy'], -13.998171,
6)
self.assertEqual(len(d["calculations"]), 2)
self.assertEqual(d['input']['is_lasph'], True)
self.assertEqual(d['input']['xc_override'], "PS")
if VaspToDbTaskDroneTest.conn:
warnings.warn("Testing query engine mode.")
qe = QueryEngine(database="creator_unittest")
self.assertEqual(qe.query().count(), 6)
#Test mappings by query engine.
for r in qe.query(
criteria={"pretty_formula": "Li2O"},
properties=["dir_name", "energy", "calculations",
"input"]):
if r["dir_name"].endswith("Li2O_aflow"):
self.assertAlmostEqual(r['energy'], -14.31446494, 4)
self.assertEqual(len(r["calculations"]), 2)
self.assertEqual(r["input"]["is_lasph"], False)
self.assertEqual(r['input']['xc_override'], None)
elif r["dir_name"].endswith("Li2O"):
self.assertAlmostEqual(r['energy'], -14.31337758, 4)
self.assertEqual(len(r["calculations"]), 1)
self.assertEqual(r["input"]["is_lasph"], False)
self.assertEqual(r['input']['xc_override'], None)
#Test lasph
e = qe.get_entries({"dir_name": {"$regex": "lasph"}})
self.assertEqual(len(e), 1)
self.assertEqual(e[0].parameters["is_lasph"], True)
self.assertEqual(e[0].parameters["xc_override"], "PS")
# Test query one.
d = qe.query_one(criteria={"pretty_formula": "TbZn(BO2)5"},
properties=["energy"])
self.assertAlmostEqual(d['energy'], -526.66747274, 4)
d = qe.get_entries_in_system(["Li", "O"])
self.assertEqual(len(d), 3)
self.assertIsInstance(d[0], ComputedEntry)
s = qe.get_structure_from_id(d[0].entry_id)
self.assertIsInstance(s, Structure)
self.assertEqual(s.formula, "Li2 O1")
self.assertIsInstance(qe.get_dos_from_id(d[0].entry_id),
CompleteDos)
from pymatgen.ext.matproj import MPRester
from pymatgen.apps.borg.hive import VaspToComputedEntryDrone
from pymatgen.apps.borg.queen import BorgQueen
from pymatgen.entries.compatibility import MaterialsProjectCompatibility
from pymatgen.analysis.phase_diagram import PhaseDiagram, PDPlotte
# Assimilate VASP calculations into ComputedEntry object. Let's assume that
# the calculations are for a series of new LixSnySz phases that we want to
# know the phase stability.
drone = VaspToComputedEntryDrone()
queen = BorgQueen(drone, rootpath=".")
entries = queen.get_data()
# Obtain all existing Li-Sn-S phases using the Materials Project REST API
with MPRester("Key") as m:
mp_entries = m.get_entries_in_chemsys(["Li", "Sn", "S"])
# Combined entry from calculated run with Materials Project entries
entries.extend(mp_entries)
# Process entries using the MaterialsProjectCompatibility
compat = MaterialsProjectCompatibility()
entries = compat.process_entries(entries)
# Generate and plot Li-Sn-S phase diagram
pd = PhaseDiagram(entries)
plotter = PDPlotter(pd,
color='black',
markerfacecolor='c',
markersize=15,
def test_load_data(self):
drone = VaspToComputedEntryDrone()
queen = BorgQueen(drone)
queen.load_data(os.path.join(test_dir, "assimilated.json"))
self.assertEqual(len(queen.get_data()), 1)
def get_phase_diagram_data(self):
"""
Returns grand potential phase diagram data to external plot
Assumes openelement specific element equals None
:return: Data to external plot
"""
open_elements_specific = None
open_element_all = Element(self.open_element)
mpr = MPRester("sMnWB7h8Lf4NKmzo")
# import do dados dos arquivos tipo vasp
drone = VaspToComputedEntryDrone()
queen = BorgQueen(drone, rootpath=".")
entries = queen.get_data()
# Get data to make phase diagram
mp_entries = mpr.get_entries_in_chemsys(self.system,
compatible_only=True)
entries.extend(mp_entries)
compat = MaterialsProjectCompatibility()
entries = compat.process_entries(entries)
#explanation_output = open("explain.txt",'w')
entries_output = open("entries.txt", 'w')
compat.explain(entries[0])
print(entries, file=entries_output)
#print(entries)
if open_elements_specific:
gcpd = GrandPotentialPhaseDiagram(entries, open_elements_specific)
self.plot_phase_diagram(gcpd, False)
self.analyze_phase_diagram(gcpd)
if open_element_all:
pd = PhaseDiagram(entries)
chempots = pd.get_transition_chempots(open_element_all)
#print(chempots)
#all_gcpds = list()
toplot = []
# dic = {}
for idx in range(len(chempots)):
if idx == len(chempots) - 1:
avgchempot = chempots[idx] - 0.1
else:
avgchempot = 0.5 * (chempots[idx] + chempots[idx + 1])
gcpd = GrandPotentialPhaseDiagram(
entries, {open_element_all: avgchempot}, pd.elements)
# toplot.append(self.get_grand_potential_phase_diagram(gcpd))
min_chempot = None if idx == len(chempots) - 1 else chempots[
idx + 1]
max_chempot = chempots[idx]
#gcpd = GrandPotentialPhaseDiagram(entries, {open_element_all: max_chempot}, pd.elements)
toplot.append(self.get_grand_potential_phase_diagram(gcpd))
#toplot.append(max_chempot)
#self.plot_phase_diagram(gcpd, False)
#print({open_element_all: max_chempot})
# Data to plot phase diagram
return toplot
def get_energies(rootdir, reanalyze, verbose, quick, sort, fmt):
"""
Get energies of all vaspruns in directory (nested).
Args:
rootdir (str): Root directory.
reanalyze (bool): Whether to ignore saved results and reanalyze
verbose (bool): Verbose mode or not.
quick (bool): Whether to perform a quick analysis (using OSZICAR instead
of vasprun.xml
sort (bool): Whether to sort the results in ascending order.
fmt (str): tablefmt passed to tabulate.
"""
if verbose:
logformat = "%(relativeCreated)d msecs : %(message)s"
logging.basicConfig(level=logging.INFO, format=logformat)
if quick:
drone = SimpleVaspToComputedEntryDrone(inc_structure=True)
else:
drone = VaspToComputedEntryDrone(
inc_structure=True, data=["filename", "initial_structure"])
ncpus = multiprocessing.cpu_count()
logging.info(f"Detected {ncpus} cpus")
queen = BorgQueen(drone, number_of_drones=ncpus)
if os.path.exists(SAVE_FILE) and not reanalyze:
msg = f"Using previously assimilated data from {SAVE_FILE}. Use -r to force re-analysis."
queen.load_data(SAVE_FILE)
else:
if ncpus > 1:
queen.parallel_assimilate(rootdir)
else:
queen.serial_assimilate(rootdir)
msg = f"Analysis results saved to {SAVE_FILE} for faster subsequent loading."
queen.save_data(SAVE_FILE)
entries = queen.get_data()
if sort == "energy_per_atom":
entries = sorted(entries, key=lambda x: x.energy_per_atom)
elif sort == "filename":
entries = sorted(entries, key=lambda x: x.data["filename"])
all_data = []
for e in entries:
if quick:
delta_vol = "NA"
else:
delta_vol = e.structure.volume / e.data[
"initial_structure"].volume - 1
delta_vol = f"{delta_vol * 100:.2f}"
all_data.append((
e.data["filename"].replace("./", ""),
re.sub(r"\s+", "", e.composition.formula),
f"{e.energy:.5f}",
f"{e.energy_per_atom:.5f}",
delta_vol,
))
if len(all_data) > 0:
headers = ("Directory", "Formula", "Energy", "E/Atom", "% vol chg")
print(tabulate(all_data, headers=headers, tablefmt=fmt))
print("")
print(msg)
else:
print("No valid vasp run found.")
os.unlink(SAVE_FILE)
return 0
def submit_vasp_directory(self,
rootdir,
authors,
projects=None,
references='',
remarks=None,
master_data=None,
master_history=None,
created_at=None,
ncpus=None):
"""
Assimilates all vasp run directories beneath a particular
directory using BorgQueen to obtain structures, and then submits thhem
to the Materials Project as SNL files. VASP related meta data like
initial structure and final energies are automatically incorporated.
.. note::
As of now, this MP REST feature is open only to a select group of
users. Opening up submissions to all users is being planned for
the future.
Args:
rootdir:
Rootdir to start assimilating VASP runs from.
authors:
*List* of {"name":'', "email":''} dicts,
*list* of Strings as 'John Doe <*****@*****.**>',
or a single String with commas separating authors. The same
list of authors should apply to all runs.
projects:
List of Strings ['Project A', 'Project B']. This applies to
all structures.
references:
A String in BibTeX format. Again, this applies to all
structures.
remarks:
List of Strings ['Remark A', 'Remark B']
masterdata:
A free form dict. Namespaced at the root level with an
underscore, e.g. {"_materialsproject":<custom data>}. This
data is added to all structures detected in the directory,
in addition to other vasp data on a per structure basis.
created_at:
A datetime object
ncpus:
Number of cpus to use in using BorgQueen to assimilate
"""
drone = VaspToComputedEntryDrone(
inc_structure=True, data=["filename", "initial_structure"])
queen = BorgQueen(drone, number_of_drones=ncpus)
queen.parallel_assimilate(rootdir)
structures = []
metadata = []
#TODO: Get histories from the data.
for e in queen.get_data():
structures.append(e.structure)
m = {
"_vasp": {
"parameters": e.parameters,
"final_energy": e.energy,
"final_energy_per_atom": e.energy_per_atom,
"initial_structure": e.data["initial_structure"].to_dict
}
}
if master_data is not None:
m.update(master_data)
metadata.append(m)
histories = None
if master_history is not None:
histories = master_history * len(structures)
return self.submit_snl(structures,
authors,
projects=projects,
references=references,
remarks=remarks,
data=metadata,
histories=histories,
created_at=created_at)