sponge.colvar.Colvar

class sponge.colvar.Colvar(shape: Tuple[int] = (), periodic: Union[bool, List[bool]] = False, use_pbc: bool = None, name: str = 'colvar', unit: str = None, dtype: type = ms.float32)

Base class for generalized collective variables (CVs) \(s(R)\).

In mathematics, CVs \(s(R)\) are defined as a low dimensional function of the atomistic coordinate \(R\) of the simulation system, which should refer to the variable describing the slow motion in the process of interest.

In MindSPONGE, Colvar Cell is the base class for "generalized" CVs. A narrow CV is generally a vector, i.e., its rank (ndim) is 1. For example, a CV of shape (S). Whereas a Colvar Cell can be of higher rank (ndim), for example, a Colvar of shape (S_1, S_2, …, S_n)

For a Colvar, multiple values can be calculated using multiple sets of coordinates. Therefore, for a Colvar Cell of shape (S_1, S_2, … , S_n), a calculation using the B set of atomic coordinates represented by a tensor with shape (B, A, D) yields a Tensor with shape (B, S_1, S_2, … , S_n). B means Batchsize, i.e. number of walkers in simulation. A means Number of colvar in system. D means Dimension of the simulation system. Usually is 3. {S_i} means Dimensions of the collective variables.

Reference:

Yang, Y. I.; Shao, Q.; Zhang, J.; Yang, L.; Gao, Y. Q. Enhanced Sampling in Molecular Dynamics [J]. The Journal of Chemical Physics, 2019, 151(7): 070902.

Parameters

• shape (Tuple) – Shape of collective variables. Default: ()

• periodic (bool) – Whether the collective variables is periodic. Default: False.

• use_pbc (bool) – Whether to use periodic boundary condition. If None is given, it will determine whether to use periodic boundary conditions based on whether the pbc_box is provided. Default: None.

• name (str) – Name of the collective variables. Default: 'colvar'

• unit (str) – Unit of the collective variables. NOTE: This is not the Units Cell that wraps length and energy. Default: None.

• dtype (type) – Data type of the collective variables. Default: float32

Supported Platforms:

Ascend GPU

property all_periodic: bool

whether all dimensions are periodic

property any_periodic: bool

whether any dimension is periodic

property dtype: type

data type of the collective variables.

Returns

type, data type of the Colvar

get_unit(units: Units = None)

return unit of the collective variables

property name: str

name of the collective variables

Returns

str, name of the CV

property ndim: int

rank (number of dimensions) of the collective variables

Returns

int, rank of the CV

property periodic: mindspore.common.tensor.Tensor

return a Tensor of data type bool to indicate whether the CV is periodic or not

reshape(input_shape: tuple)

rearranges the shape

set_name(name: str)

set the name of the collective variables

set_pbc(use_pbc: bool)

set whether to use periodic boundary condition

property shape: tuple

shape of the collective variables (S_1, S_2, …, S_n)

Returns

shape (tuple), Shape of the Colvar

property use_pbc: bool

whether to use periodic boundary condition

Returns

bool, whether to use periodic boundary condition.

classmethod vector_in_pbc(vector: Tensor, pbc_box: Tensor)

Make the difference of vectors at the range from -0.5 box to 0.5 box