sponge.colvar.function.combine 源代码

# Copyright 2021-2023 @ Shenzhen Bay Laboratory &
#                       Peking University &
#                       Huawei Technologies Co., Ltd
#
# This code is a part of MindSPONGE:
# MindSpore Simulation Package tOwards Next Generation molecular modelling.
#
# MindSPONGE is open-source software based on the AI-framework:
# MindSpore (https://www.mindspore.cn/)
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
"""
Combine Colvar
"""

from typing import Union, List, Tuple
from numpy import ndarray
import mindspore as ms
import mindspore.numpy as msnp
from mindspore.ops import functional as F
from mindspore import Tensor
from mindspore.nn import CellList

from ..colvar import Colvar
from ...function import get_ms_array, any_none, any_not_none, check_broadcast


[文档]class CombineCV(Colvar): r""" Polynomial combination of a set of Colvar :math:`{s_i}` with shape (S_1, S_2, ..., S_n). `{S_i}` means dimensions of collective variables. .. math:: S = \sum_i^n{w_i (s_i - o_i)^{p_i}} Args: colvar (Union[List[Colvar], Tuple[Colvar]]): Array of `Colvar` to be combined :math:`{s_i}`. weights (Union[List[float], Tuple[Float], float, Tensor]): Weights :math:`{w_i}` for each Colvar. If a list or tuple is given, the number of the elements should be equal to the number of CVs. If a float or Tensor is given, the value will be used for all Colvar. Default: 1 offsets (Union[List[float], Tuple[Float], float, Tensor]): Offsets :math:`{o_i}` for each Colvar. If a list or tuple is given, the number of the elements should be equal to the number of CVs. If a float or Tensor is given, the value will be used for all Colvar. Default: 0 exponents (Union[List[float], Tuple[Float], float, Tensor]): Exponents :math:`{p_i}` for each Colvar. If a list or tuple is given, the number of the elements should be equal to the number of CVs. If a float or Tensor is given, the value will be used for all Colvar. Default: 1 normal (bool): Whether to normalize all weights to 1. Default: ``False``. periodic_min (Union[float, ndarray, Tensor]): The periodic minimum of the output of the combination of the CVs. If the output is not periodic, it should be None. Default: ``None``. periodic_max (Union[float, ndarray, Tensor]): The periodic maximum of the output of the combination of the CVs. If the output is not periodic, it should be None. Default: ``None``. periodic_mask (Union[Tensor, ndarray]): Mask for the periodicity of the outputs. The shape of the tensor should be as the same as the outputs, i.e. `(S_1, S_2, ..., S_n)`. Default: ``None``. 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: 'combine' Supported Platforms: ``Ascend`` ``GPU`` """ def __init__(self, colvar: Union[List[Colvar], Tuple[Colvar]], weights: Union[float, List[float], Tuple[float], Tensor] = 1, offsets: Union[float, List[float], Tuple[float], Tensor] = 0, exponents: Union[float, List[float], Tuple[float], Tensor] = 1, normal: bool = False, periodic_min: Union[float, ndarray, Tensor] = None, periodic_max: Union[float, ndarray, Tensor] = None, periodic_mask: Union[Tensor, ndarray] = None, use_pbc: bool = None, name: str = 'combine', ): super().__init__( periodic=(periodic_min is not None), name=name ) if any_none([periodic_min, periodic_max]) and any_not_none([periodic_min, periodic_max]): raise ValueError('The "periodic_min" and "periodic_max" must both be None, or neither.') if isinstance(colvar, Colvar): colvar = [colvar] elif not isinstance(colvar, (list, tuple)): raise TypeError(f'The type of "colvar" must be list of Colvar but got: {type(colvar)}') self.num_colvar = len(colvar) def _check_parameters(parameters: Union[list, tuple, float, Tensor], name: str): """check parameters for combination""" if isinstance(parameters, (list, tuple)): num_ = len(parameters) if num_ == self.num_colvar: return [get_ms_array(p, ms.float32) for p in parameters] if num_ != 1: raise ValueError(f'The number of {name} ({num_}) does not match ' f'the number of colvar {self.num_colvar}') return [get_ms_array(parameters[0], ms.float32)] * self.num_colvar return [get_ms_array(parameters, ms.float32)] * self.num_colvar self.weights = _check_parameters(weights, 'weights') self.offsets = _check_parameters(offsets, 'offsets') self.exponents = [(None if (e == 1).all() else e) for e in _check_parameters(exponents, 'exponents')] if normal: norm_factor = 0 for w in self.weights: norm_factor += w self.weights = [w / norm_factor for w in self.weights] shape = None colvar_ = [] for i, cv in enumerate(colvar): try: shape = check_broadcast(shape, cv.shape) except ValueError: raise ValueError(f'The shape of the {i}-th colvar {cv.shape} cannot be ' f'broadcast to the shape of the output: {shape}') wshape = self.weights[i].shape try: check_broadcast(wshape, cv.shape) except ValueError: raise ValueError(f'The shape of the {i}-th weight {wshape} cannot be broadcast to ' f'the shape of the corresponding colvar: {cv.shape}') oshape = self.offsets[i].shape try: check_broadcast(oshape, cv.shape) except ValueError: raise ValueError(f'The shape of the {i}-th offset {oshape} cannot be broadcast to ' f'the shape of the corresponding colvar: {cv.shape}') if self.exponents[i] is not None: eshape = self.exponents[i].shape try: check_broadcast(eshape, cv.shape) except ValueError: raise ValueError(f'The shape of the {i}-th exponent {eshape} cannot be broadcast to ' f'the shape of the corresponding colvar: {cv.shape}') if use_pbc is not None: cv.set_pbc(use_pbc) colvar_.append(cv) self.colvar: List[Colvar] = CellList(colvar_) self._shape = shape self._ndim = len(self._shape) self.periodic_min = None self.periodic_max = None self.periodic_range = None self.periodic_mask = None if self._periodic: self.periodic_min = msnp.broadcast_to(get_ms_array(periodic_min, ms.float32), self._shape) self.periodic_max = msnp.broadcast_to(get_ms_array(periodic_max, ms.float32), self._shape) self.periodic_range = self.periodic_max - self.periodic_min if (self.periodic_range <= 0).any(): raise ValueError(f'periodic_max {self.periodic_max} must be greater than' f'periodic_min {self.periodic_min}!') if periodic_mask is not None: self.periodic_mask = msnp.broadcast_to(get_ms_array(periodic_mask, ms.bool_), self._shape) self._periodic = msnp.broadcast_to(get_ms_array(self._periodic, ms.bool_), self._shape)
[文档] def set_pbc(self, use_pbc: bool): """set whether to use periodic boundary condition""" self._use_pbc = use_pbc self.get_vector.set_pbc(use_pbc) for i in range(self.num_colvar): self.colvar[i].set_pbc(use_pbc) return self
def construct(self, coordinate: Tensor, pbc_box: Tensor = None): r"""get position coordinates of colvar group Args: coordinate (Tensor): Tensor of shape `(B, A, D)`. Data type is float. Position coordinate of colvar in system. `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 pbc_box (Tensor): Tensor of shape `(B, D)`. Data type is float. Tensor of PBC box. Default: ``None``. Returns: combine (Tensor): Tensor of shape `(B, S_1, S_2, ..., S_n)`. Data type is float. """ colvar = 0 for i in range(self.num_colvar): colvar_ = self.colvar[i](coordinate, pbc_box) - self.offsets[i] if self.exponents[i] is not None: colvar_ = msnp.power(colvar_, self.exponents[i]) colvar += colvar_ * self.weights[i] if self.periodic_range is None: return colvar period = F.floor((colvar - self.periodic_min) / self.periodic_range) period_colvar = colvar - self.periodic_range * period if self.periodic_mask is None: return period_colvar return F.select(self.periodic_mask, period_colvar, colvar)
[文档]class ColvarCombine(CombineCV): r""" See `CombineCV`. NOTE: This module will be removed in a future release, please use `CombineCV` instead. .. math:: S = \sum_i^n{w_i (s_i - o_i)^{p_i}} Args: colvar (list or tuple): Array of `Colvar` to be combined :math:`{s_i}`. weights (list, tuple, float, Tensor): Weights :math:`{w_i}` for each Colvar. If a list or tuple is given, the number of the elements should be equal to the number of CVs. If a float or Tensor is given, the value will be used for all Colvar. Default: 1 offsets (list, tuple, float, Tensor): Offsets :math:`{o_i}` for each Colvar. If a list or tuple is given, the number of the elements should be equal to the number of CVs. If a float or Tensor is given, the value will be used for all Colvar. Default: 0 exponents (list, tuple, float, Tensor): Exponents :math:`{p_i}` for each Colvar. If a list or tuple is given, the number of the elements should be equal to the number of CVs. If a float or Tensor is given, the value will be used for all Colvar. Default: 1 normal (bool): Whether to normalize all weights to 1. Default: ``False``. periodic_min (float, ndarray, Tensor): The periodic minimum of the output of the combination of the CVs. If the output is not periodic, it should be None. Default: ``None``. periodic_max (float, ndarray, Tensor): The periodic maximum of the output of the combination of the CVs. If the output is not periodic, it should be None. Default: ``None``. periodic_mask (Tensor, ndarray): Mask for the periodicity of the outputs. The shape of the tensor should be as the same as the outputs, i.e. `(S_1, S_2, ..., S_n)`. Default: ``None``. 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_combination' Supported Platforms: ``Ascend`` ``GPU`` """ def __init__(self, colvar: Union[List[Colvar], Tuple[Colvar]], weights: Union[float, List[float], Tuple[float], Tensor] = 1, offsets: Union[float, List[float], Tuple[float], Tensor] = 0, exponents: Union[float, List[float], Tuple[float], Tensor] = 1, normal: bool = False, periodic_min: Union[float, ndarray, Tensor] = None, periodic_max: Union[float, ndarray, Tensor] = None, periodic_mask: Union[Tensor, ndarray] = None, use_pbc: bool = None, name: str = 'colvar_combination', ): super().__init__( colvar=colvar, weights=weights, offsets=offsets, exponents=exponents, normal=normal, periodic_min=periodic_min, periodic_max=periodic_max, periodic_mask=periodic_mask, use_pbc=use_pbc, name=name ) print('[WARNING] This module will be removed in a future release, please use `CombineCV` instead.')