matlantis_features.features.common.opt.FireASEOptFeature

class matlantis_features.features.common.opt.FireASEOptFeature(n_run: int = 200, fmax: float = 0.01, attach_methods: Optional[List[Tuple[Callable[[], None], int]]] = None, show_progress_bar: bool = False, tqdm_options: Optional[Dict[str, Any]] = None, show_logger: bool = False, logger_interval: int = 10, filter: Union[bool, MatlantisFilter] = False, trajectory: Optional[str] = None, estimator_fn: Optional[Callable[[], Estimator]] = None, maxstep: float = 0.2)#

Bases: ASEOptFeature

The matlantis-feature that uses the Fast Inertial Relaxation Engine (FIRE), as implemented within ASE.

For details of the FIRE algorithm, please see:
Guénolé et al., Comput. Mater. Sci. 175 (2020) 109584.

Methods

`__init__`([n_run, fmax, attach_methods, …])	Initialize an instance.
`__call__`(atoms)	Call function.
`attach`(extn, interval)	Used for attaching a function to the ASE optimizer.
`attach_ctx`([ctx])	Attach the feature to matlantis_features.utils.Context.
`check_estimator_fn`(estimator_fn)	Checks if the given estimator function is None and output a warning if so.
`cost_estimate`([atoms])	Estimate the cost of the feature.
`from_dict`(res)	Construct a FeatureBase object from serialized dict.
`get_savedir_from_ctx`()	Get the temporary save directory from the context.
`init_scope`()	Context manager that enable to set attribution of the feature.
`repeat`(n_repeat)	Set the maximum number of times that allowed to run the __call__ function.
`to_dict`()	Dictionary representation of the FireASEOptFeature.

__init__(n_run: int = 200, fmax: float = 0.01, attach_methods: Optional[List[Tuple[Callable[[], None], int]]] = None, show_progress_bar: bool = False, tqdm_options: Optional[Dict[str, Any]] = None, show_logger: bool = False, logger_interval: int = 10, filter: Union[bool, MatlantisFilter] = False, trajectory: Optional[str] = None, estimator_fn: Optional[Callable[[], Estimator]] = None, maxstep: float = 0.2)#

Initialize an instance.

Parameters

n_run (int, optional) – The maximum number of optimization steps performed to try to
reach force convergence (specified by fmax). Defaults to 200.
fmax (float, optional) – The maximum force (in eV/Angstrom) for convergence of optimization.
Defaults to 0.01.
attach_methods (list[tuple[ -> None, int]] or None, optional) – Functions to be called by the optimizer. A list of tuples must be specified,
where the first element of each tuple is a callable function, and the second element
is an integer N. The optimizer will call the function every N steps.
If None, no functions will be attached. Defaults to None.
show_progress_bar (bool, optional) – Show progress bar. Defaults to False.
tqdm_options (dict[str, Any] or None, optional) – Options for tqdm.
show_logger (bool, optional) – Show log information. Defaults to False.
logger_interval (int, optional) – The interval of when to print out logger information.
Defaults to 10.
filter (bool or MatlantisFilter, optional) – The MatlantisFilter used in the optimization. If True, the default filter, i.e.
UnitCellASEFilter, will be used for the cell shape optimization. If False, no
filter is used.
Note: The filter is only applicable in the optimization of ASEAtoms and MatlantisAtoms,
not NEB class.
Defaults to False.
trajectory (str or None, optional) – Filename for trajectory file. If None, no trajectory file will be saved. Defaults
to None.
estimator_fn (EstimatorFnType or None, optional) – A factory method to create a custom estimator.
Please refer Customizing estimator used in matlantis-features for detail.
maxstep (float, optional) – The maximum distance an atom can move per iteration. The unit is angstrom.
Defaults to 0.2.

__call__(atoms: Union[Atoms, MatlantisAtoms, NEB, Filter]) → OptFeatureResult#

Call function.

Parameters: atoms (ASEAtoms or MatlantisAtoms or NEB or Filter) – The system to be optimized. In addition to ASEAtoms and MatlantisAtoms,
NEB (Matlantis) and Filter (ASE) objects are also supported.
Note: the support of ASE Filter class will be deprecated in the future.
Please use the ‘filter’ parameter in the __init__ function.
Returns: A dataclass containing information about the result of optimization.
Return type: OptFeatureResult

attach(extn: Callable[[], None], interval: int) → None#

Used for attaching a function to the ASE optimizer.

Parameters

extn (-> None) – The function to be attached.
interval (int) – The interval (in optimization steps) of when the optimizer will call the function.

attach_ctx(ctx: Optional[Context] = None) → None#

Attach the feature to matlantis_features.utils.Context.

Parameters: ctx (Context or None, optional) – The matlantis_features.utils.Context object. Defaults to None.

check_estimator_fn(estimator_fn: Optional[Callable[[], Estimator]]) → None#

Checks if the given estimator function is None and output a warning if so.

Parameters: estimator_fn (EstimatorFnType or None, optional) – A factory method to create a custom estimator.
Please refer Customizing estimator used in matlantis-features for detail. Defaults to None.

cost_estimate(atoms: Optional[Union[Atoms, MatlantisAtoms]] = None) → FeatureCost#

Estimate the cost of the feature.

Parameters: atoms (ASEAtoms or MatlantisAtoms or None, optional) – The input atoms. Defaults to None.
Returns: The cost of the feature.
Return type: FeatureCost

classmethod from_dict(res: Dict[str, Any]) → FeatureBase#

Construct a FeatureBase object from serialized dict.

Parameters: res (dict[str, Any]) – A dict containing a serialized FeatureBase from to_dict().
Returns: The deserialized object from provided dict.
Return type: FeatureBase