`bte.grad.distribution`#

Module Contents#

Classes#

`distributionBase`	The base class for gas distribution.
`HermitedistributionBase`
`HermiteDis`	HermiteDis: Gas distribution in Hermite form.
`DVDisMeta`	Base class for all neural network modules.
`DVDisMeta_Grid`	Base class for all neural network modules.
`DVDis`	The gas distribution for discrete velocity method.
`DVDis_Chu`	The base class for gas distribution.
`HermiteDisND`	HermiteDis: Gas distribution in Hermite form.

Functions#

`d3roll`(tensor, idx)
`Project_RK4_helper`(dis, u0, theta0, u, theta)
`Project_RK4_step`(dis, u0, theta0, u, theta, h)
`Project_RK4`(dis, u, theta[, STEP])	Project_RK4 使用R-K4格式对分布函数进行投影
`Project_RK4_helperT`(dis, u0, theta0, u, theta)
`Project_RK4_stepT`(dis, u0, theta0, u, theta[, h])
`Project_RK4T`(dis, u0, theta0, u, theta[, STEP])
`eval_hermitenorm`(n, x)
`eval_hermitenorm_seq`(n, x)
`uniform_velocity`(nv, vmin, vmax)	return a discrete of velocity with uniform partition.
`legendre_velocity`(nv, vmin, vmax)	return a discrete of velocity with points on legendre-gauss points.
`uniform_velocity_2d`(nv, vmin, vmax)
`legendre_velocity_2d`(nv, vmin, vmax)
`ND_velocity`(func)
`velocity_list`(func, nvT, vminT, vmaxT)
`product_meshgrid`(vL, wL)

Attributes#

`uniform_velocity_nd`
`legendre_velocity_nd`

class bte.grad.distribution.distributionBase[source]#

The base class for gas distribution.

abstract density()[source]#

abstract velocity()[source]#

abstract temperature()[source]#

abstract __add__(anotherDis)[source]#

abstract __radd__(anotherDis)[source]#

abstract __sub__(anotherDis)[source]#

abstract __mul__(multiplier)[source]#

abstract __rmul__(multiplier)[source]#

abstract __truediv__(divider)[source]#

class bte.grad.distribution.HermitedistributionBase[source]#

abstract down()[source]#

abstract new(*args, **kw)[source]#

class bte.grad.distribution.HermiteDis(u, theta, coef)[source]#

Bases: HermitedistributionBase

HermiteDis: Gas distribution in Hermite form.

Parameters:

u (bte.utils.indexs.torch.Tensor) –
theta (bte.utils.indexs.torch.Tensor) –
coef (bte.utils.indexs.torch.Tensor) –

property ORDER[source]#

property rho[source]#

property moments[source]#

property rho_u_theta[source]#

new(u, theta, coef, *args, **kw)[source]#

__str__()[source]#

Return str(self).

Return type:: str

static empty(order, device='cpu')[source]#

Parameters:: order (int) –
Return type:: HermiteDis

from_DVDis(dis, u=None, theta=None)[source]#

Parameters:: dis (DVDis) –
Return type:: HermiteDis

cpu()[source]#

cuda()[source]#

to(args)[source]#

to_tensor()[source]#

density()[source]#

velocity()[source]#

temperature()[source]#

as_tuple()[source]#

__add__(anotherDis)[source]#

__radd__(anotherDis)[source]#

__sub__(anotherDis)[source]#

__mul__(multiplier)[source]#

__rmul__(multiplier)[source]#

__truediv__(divider)[source]#

roll(idx)[source]#

clone()[source]#

mulvec()[source]#

down()[source]#

Project_to_STD(STEP=1)[source]#

Project_to_STD 将分布参数重设，重新投影到最佳参数上

Parameters:: STEP (int) –

Project_2B(dest, STEP=1)[source]#

Project_A2B 将分布投影到dest对应的参数上

Parameters:

dest (HermiteDis) –
STEP (int) –

Return type:

HermiteDis

bte.grad.distribution.d3roll(tensor, idx)[source]#

bte.grad.distribution.Project_RK4_helper(dis, u0, theta0, u, theta)[source]#

bte.grad.distribution.Project_RK4_step(dis, u0, theta0, u, theta, h)[source]#

Parameters:: h (float) –

bte.grad.distribution.Project_RK4(dis, u, theta, STEP=1)[source]#

Project_RK4 使用R-K4格式对分布函数进行投影

Parameters:

dis (torch.Tensor) – 分布函数
u (torch.Tensor) – 分布中心
theta (torch.Tensor) – 分布宽度
STEP (int, optional) – RK方法的步数, by default 1

Returns:

投影后的分布函数

Return type:

torch.Tensor

bte.grad.distribution.Project_RK4_helperT(dis, u0, theta0, u, theta)[source]#

bte.grad.distribution.Project_RK4_stepT(dis, u0, theta0, u, theta, h=1)[source]#

Parameters:: h (float) –

bte.grad.distribution.Project_RK4T(dis, u0, theta0, u, theta, STEP=1)[source]#

Parameters:: STEP (int) –

bte.grad.distribution.eval_hermitenorm(n, x)[source]#

Parameters:

n (int) –
x (bte.utils.indexs.torch.Tensor) –

bte.grad.distribution.eval_hermitenorm_seq(n, x)[source]#

Parameters:

n (int) –
x (bte.utils.indexs.torch.Tensor) –

class bte.grad.distribution.DVDisMeta(v, v_w, *args)[source]#

Bases: torch.nn.Module

Base class for all neural network modules.

Your models should also subclass this class.

Modules can also contain other Modules, allowing to nest them in a tree structure. You can assign the submodules as regular attributes:

import torch.nn as nn
import torch.nn.functional as F

class Model(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = nn.Conv2d(1, 20, 5)
        self.conv2 = nn.Conv2d(20, 20, 5)

    def forward(self, x):
        x = F.relu(self.conv1(x))
        return F.relu(self.conv2(x))

Submodules assigned in this way will be registered, and will have their parameters converted too when you call to(), etc.

Note

As per the example above, an __init__() call to the parent class must be made before assignment on the child.

Variables:: training (bool) – Boolean represents whether this module is in training or evaluation mode.

class bte.grad.distribution.DVDisMeta_Grid(vL, v_wL, *args)[source]#

Bases: DVDisMeta

Base class for all neural network modules.

Your models should also subclass this class.

Modules can also contain other Modules, allowing to nest them in a tree structure. You can assign the submodules as regular attributes:

import torch.nn as nn
import torch.nn.functional as F

class Model(nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = nn.Conv2d(1, 20, 5)
        self.conv2 = nn.Conv2d(20, 20, 5)

    def forward(self, x):
        x = F.relu(self.conv1(x))
        return F.relu(self.conv2(x))

Submodules assigned in this way will be registered, and will have their parameters converted too when you call to(), etc.

Note

As per the example above, an __init__() call to the parent class must be made before assignment on the child.

Variables:

training (bool) – Boolean represents whether this module is in training or evaluation mode.

Parameters:

vL (bte.utils.indexs.List) –
v_wL (bte.utils.indexs.List) –

class bte.grad.distribution.DVDis(v_meta, data)[source]#

Bases: distributionBase

The gas distribution for discrete velocity method.

Parameters:: v_meta (DVDisMeta) –

property v[source]#

property v2[source]#

property v3[source]#

property w[source]#

cpu()[source]#

cuda()[source]#

to(args)[source]#

__add__(anotherDis)[source]#

Parameters:: anotherDis (DVDis) –

__radd__(anotherDis)[source]#

Parameters:: anotherDis (DVDis) –

__sub__(anotherDis)[source]#

__mul__(multiplier)[source]#

__rmul__(multiplier)[source]#

__truediv__(divider)[source]#

static empty(nv, vmin, vmax, method='uni', device='cpu')[source]#

An empty new distribution. with velocity grid created by method ‘uni’ or ‘leg’.

Parameters:

nv (int) – Numbers of grid
vmin (float) – vmin
vmax (float) – vmax
method (str, optional) – ‘uni’ for uniform_velocity, ‘leg’ for legendre_velocity. Defaults to ‘uni’.

Raises:

ValueError – [description]

Returns:

[description]

Return type:

DVDis

sum(f)[source]#

Parameters:: f (bte.utils.indexs.torch.Tensor) –

density(f=None)[source]#: Compute the macro quantity: density

velocity(f=None)[source]#: Compute the macro quantity: velocity

temperature(f=None)[source]#: Compute the macro quantity: temperature

_m0(f=None)[source]#: 0-th order momentum of f

_m1(f=None)[source]#: 1st order momentum of f

_m2(f=None)[source]#: 2nd order momentum of f

_m3(f=None)[source]#: 3nd order momentum of f

heatflux(f=None)[source]#

rho_u_theta(f=None)[source]#

maxwellian(rho_u_theta=None)[source]#

maxwellian_half(rho_u_theta=None, sign=0)[source]#

one_side(f=None, sign=0)[source]#

reverse(f=None)[source]#

from_HermiteDis(dis)[source]#

Parameters:: dis (HermiteDis) –
Return type:: DVDis

bte.grad.distribution.uniform_velocity(nv, vmin, vmax)[source]#

return a discrete of velocity with uniform partition.

v_0=vmin+0.5*(vmax-vmin)/nv v_{nv-1}=vmax-0.5*(vmax-vmin)/nv

Parameters:

nv (int) – Numbers of grid
vmin (float) – vmin
vmax (float) – vmax

Returns:

v, weight of v

Return type:

Tuple[torch.Tensor,torch.Tensor]

bte.grad.distribution.legendre_velocity(nv, vmin, vmax)[source]#

return a discrete of velocity with points on legendre-gauss points.

Parameters:

nv (int) – Numbers of grid
vmin (float) – vmin
vmax (float) – vmax

Returns:

v, weight of v

Return type:

Tuple[torch.Tensor,torch.Tensor]

bte.grad.distribution.uniform_velocity_2d(nv, vmin, vmax)[source]#

bte.grad.distribution.legendre_velocity_2d(nv, vmin, vmax)[source]#

bte.grad.distribution.ND_velocity(func)[source]#

Parameters:: func (Callable[[int, int, int], bte.utils.indexs.Tuple]) –

bte.grad.distribution.uniform_velocity_nd[source]#

bte.grad.distribution.legendre_velocity_nd[source]#

bte.grad.distribution.velocity_list(func, nvT, vminT, vmaxT)[source]#

Parameters:

nvT (bte.utils.indexs.Tuple) –
vminT (bte.utils.indexs.Tuple) –
vmaxT (bte.utils.indexs.Tuple) –

bte.grad.distribution.product_meshgrid(vL, wL)[source]#

Parameters:

vL (bte.utils.indexs.List) –
wL (bte.utils.indexs.List) –

class bte.grad.distribution.DVDis_Chu(v_meta, data_g, data_h, ndim=3)[source]#

Bases: distributionBase

The base class for gas distribution.

property v[source]#

property v2[source]#

property v3[source]#

property w[source]#

__add__(another)[source]#

__radd__(another)[source]#

__sub__(another)[source]#

__mul__(multiplier)[source]#

__rmul__(multiplier)[source]#

__truediv__(divider)[source]#

density(g=None, h=None)[source]#

velocity(g=None, h=None)[source]#

temperature(g=None, h=None)[source]#

rho_u_theta(g=None, h=None)[source]#

maxwellian(g=None, h=None, rho_u_theta=None)[source]#

class bte.grad.distribution.HermiteDisND(u, theta, coef, indt)[source]#

Bases: HermitedistributionBase

HermiteDis: Gas distribution in Hermite form.

Parameters:

u (bte.utils.indexs.torch.Tensor) –
theta (bte.utils.indexs.torch.Tensor) –
coef (bte.utils.indexs.torch.Tensor) –
indt (bte.utils.indexs.index_table) –

property ORDER[source]#

property rho[source]#

property moments[source]#

property rho_u_theta[source]#

new(u, theta, coef, indt=None)[source]#

M()[source]#

__str__()[source]#

Return str(self).

Return type:: str

static empty(order, dim=3, device='cpu')[source]#

Parameters:: order (int) –
Return type:: HermiteDisND

from_DVDis(dis, u=None, theta=None)[source]#

Parameters:: dis (DVDis) –
Return type:: HermiteDis

cpu()[source]#

cuda()[source]#

to(args)[source]#

to_tensor()[source]#

density()[source]#

velocity()[source]#

temperature()[source]#

as_tuple()[source]#

__add__(anotherDis)[source]#

__radd__(anotherDis)[source]#

__sub__(anotherDis)[source]#

__mul__(multiplier)[source]#

__rmul__(multiplier)[source]#

__truediv__(divider)[source]#

roll(idx)[source]#

clone()[source]#

down()[source]#

Project_to_STD(STEP=1)[source]#

Project_to_STD 将分布参数重设，重新投影到最佳参数上

Parameters:: STEP (int) –

Project_2B(dest, STEP=1)[source]#

Project_A2B 将分布投影到dest对应的参数上

Parameters:

dest (HermiteDisND) –
STEP (int) –

Return type:

HermiteDis

mulvec(axis=0)[source]#: axis=0 for x, 1 for y, 2 for z.

Project_RK4_helper(dis, u0, theta0, u, theta)[source]#

Project_RK4_step(dis, u0, theta0, u, theta, h)[source]#

Parameters:: h (float) –

Project_RK4(u, theta, STEP=1)[source]#

Project_RK4 使用R-K 4格式对分布函数进行投影

Parameters:

dis (torch.Tensor) – 分布函数
u (torch.Tensor) – 分布中心
theta (torch.Tensor) – 分布宽度
STEP (int, optional) – RK方法的步数, by default 1

Returns:

投影后的分布函数

Return type:

torch.Tensor

bte.grad.distribution#

Module Contents#

Classes#

Functions#

Attributes#

`bte.grad.distribution`#