def construct_datasets(self, tag_xyz_map):
dc = self.dataset_config
mc = self.model_config
pc = self.prediction_config
preprocesses = []
for (name, args, kwargs) in dc.preprocesses:
preprocess = PREPROCESS[name](*args, **kwargs)
preprocess.load(pc.load_dir / 'preprocess' / f'{name}.npz',
verbose=self.verbose)
preprocesses.append(preprocess)
datasets = []
for pattern in pc.tags:
for tag in fnmatch.filter(tag_xyz_map, pattern):
if self.verbose:
pprint(f'Construct sub dataset tagged as "{tag}"')
tagged_xyz = tag_xyz_map.pop(tag)
structures = AtomicStructure.read_xyz(tagged_xyz)
# prepare descriptor dataset
descriptor = DESCRIPTOR_DATASET[dc.descriptor](pc.order,
structures,
**dc.parameters)
descriptor.make(verbose=self.verbose)
##Save descriptor (symmetory_function)
descriptor_npz = pc.load_dir / f'{dc.descriptor}-pred.npz'
descriptor.save(descriptor_npz, verbose=self.verbose)
##End of saving descriptor (symmetory_function)
# prepare empty property dataset
property_ = PROPERTY_DATASET[dc.property_](pc.order,
structures)
# construct test dataset from descriptor & property datasets
dataset = HDNNPDataset(descriptor, property_)
dataset.construct(all_elements=pc.elements,
preprocesses=preprocesses,
shuffle=False,
verbose=self.verbose)
datasets.append(dataset)
dc.n_sample += dataset.total_size
mc.n_input = dataset.n_input
mc.n_output = dataset.n_label
return datasets
def construct_test_datasets(self, tag_xyz_map, load_descriptor):
dc = self.dataset_config
mc = self.model_config
tc = self.training_config
preprocess_dir = tc.out_dir / 'preprocess'
preprocesses = []
for (name, args, kwargs) in dc.preprocesses:
preprocess = PREPROCESS[name](*args, **kwargs)
preprocess.load(preprocess_dir / f'{preprocess.name}.npz',
verbose=self.verbose)
preprocesses.append(preprocess)
datasets = []
for pattern in tc.tags:
for tag in fnmatch.filter(tag_xyz_map, pattern):
if self.verbose:
pprint(f'Construct sub test dataset tagged as "{tag}"')
tagged_xyz = tag_xyz_map.pop(tag)
structures = AtomicStructure.read_xyz(tagged_xyz)
# prepare descriptor dataset
descriptor = DESCRIPTOR_DATASET[dc.descriptor](
self.loss_function.order['descriptor'], structures,
**dc.parameters)
descriptor_npz = tagged_xyz.with_name(
f'{dc.descriptor}-test.npz')
if load_descriptor:
descriptor.load(descriptor_npz,
verbose=self.verbose,
remake=dc.remake)
else:
descriptor.make(verbose=self.verbose)
descriptor.save(descriptor_npz, verbose=self.verbose)
# prepare property dataset
property_ = PROPERTY_DATASET[dc.property_](
self.loss_function.order['property'], structures)
property_npz = tagged_xyz.with_name(f'{dc.property_}-test.npz')
if property_npz.exists() and load_descriptor:
property_.load(property_npz,
verbose=self.verbose,
remake=dc.remake)
else:
property_.make(verbose=self.verbose)
property_.save(property_npz, verbose=self.verbose)
# construct test dataset from descriptor & property datasets
dataset = HDNNPDataset(descriptor, property_)
dataset.construct(all_elements=tc.elements,
preprocesses=preprocesses,
shuffle=False,
verbose=self.verbose)
dataset.scatter()
datasets.append(dataset)
dc.n_sample += dataset.total_size
mc.n_input = dataset.n_input
mc.n_output = dataset.n_label
return datasets