Eigen - reshape a vector to matrix - c++

I am trying to reshape a vector to a matrix, but getting the following error
unsigned int Nx = 8;
unsigned int Ny = 7;
Eigen::VectorXi G_temp = Eigen::VectorXi::LinSpaced((Nx + 2) * (Ny + 2),0,(Nx + 2) * (Ny + 2)-1);
Eigen::MatrixXd G = Eigen::Map<Eigen::MatrixXd>(G_temp.data(),Nx+2, Ny+2); // error: no matching constructor for initialization of 'Eigen::Map<Eigen::MatrixXd>'
I followed what is written here, but I do not understand the way I am doing wrong.

There is no implicit conversion from integer-valued to double-valued expressions in Eigen. Either just use VectorXd for G_temp (and the LinSpaced expression):
Eigen::VectorXd G_temp = Eigen::VectorXd::LinSpaced((Nx + 2) * (Ny + 2),0,(Nx + 2) * (Ny + 2)-1);
Or use a MatrixXi-Map and .cast<double>() the result before assigning it to G.
Eigen::MatrixXd G = Eigen::Map<Eigen::MatrixXi>(G_temp.data(),Nx+2, Ny+2).cast<double>();
To avoid any temporary, you can also allocate a MatrixXd and directly assign the proper values inside:
Eigen::MatrixXd G(Nx+2, Ny+2); // allocate matrix
// set values in-place:
Eigen::VectorXd::Map(G.data(), (Nx + 2) * (Ny + 2)).setLinSpaced(0,(Nx + 2) * (Ny + 2)-1);
Or with the master/3.4 branch:
G.reshaped().setLinSpaced(0,(Nx + 2) * (Ny + 2)-1);

Related

How do I generate Bezier Curves and NURBS in C++ and import it as an igs?

I am new to C++ NURBS libary. I learnt generating line (by CLine, from nurbs.h ) and save it as igs. But in case of
multiple control points, how to generate a curve ? Every other tutorial using graphics.h
(putpixel), but couldnt find anything about igs.
This should be a simple problem. But I have no idea which function can help me here.
Thanks in advance.
We have 4 control points here to begin with.
for (float t = 0.0; t <= 1.0; t += 0.2) {
double xt = 0.0, yt = 0.0;
xt = pow(1 - t, 3) * x[0] + 3 * t * pow(1 - t, 2) * x[1] + 3 * pow(t, 2) * (1 - t) * x[2]
+ pow(t, 3) * x[3];
yt = pow(1 - t, 3) * y[0] + 3 * t * pow(1 - t, 2) * y[1] + 3 * pow(t, 2) * (1 - t) * y[2]
+ pow(t, 3) * y[3];
count = count + 1;
//Math::Vector4f c(xt, yt, 0);
for (int i = 1; i < 3; i++) {
listt[i][0]= xt;
listt[i][1]= yt;
Math::Vector4f a(listt[i][0], listt[i][1],0);
myvector.push_back (&a);
}
}
......
.....
igs.Write("test.igs");
--- This is to create the points, but after that I dont know how to use the points to create a Bezier curve .

C++: Get value of variable after performing divide and assignment operator:

I am trying to get a value of a variable "pesoil" before its converted to to mm month^-1
pesoil = beta * rnsoil + CPAIR * RHOAIR * exp_h * vpdo / r_as;
pesoil /= (beta + PSY * exp_h * (1.0 + r_ss / r_as))); // W m^-2
pesoil *= (etimes * ndays[dm]) / LAMBDA; // convert to mm month^-1
can I do it like this? or will this mess up value stored in "pesoil"
epotential_W = (pesoil /= (beta + PSY * exp_h * (1.0 + r_ss / r_as))); // W m^-2

Converting 1-d array to 2d

I am trying to understand this code:
void stencil(const int nx, const int ny, const int width, const int height,
double* image, double* tmp_image)
{
for (int j = 1; j < ny + 1; ++j) {
for (int i = 1; i < nx + 1; ++i) {
tmp_image[j + i * height] = image[j + i * height] * 3.0 / 5.0;
tmp_image[j + i * height] += image[j + (i - 1) * height] * 0.5 / 5.0;
tmp_image[j + i * height] += image[j + (i + 1) * height] * 0.5 / 5.0;
tmp_image[j + i * height] += image[j - 1 + i * height] * 0.5 / 5.0;
tmp_image[j + i * height] += image[j + 1 + i * height] * 0.5 / 5.0;
}
}
}
The 1-d array notation is very confusing. I am trying to convert it to a 2-d notation (which I find easier to read). Could someone point me in the right direction as to how I can accomplish this?
All this code is doing is creating a new image from an original image by taking 60% from the corresponding pixel and 10% from each neighboring pixel.
When you see tmp_image[j + i * height], read it as tmp_image[i][j].
Changing the code to literally use 2D syntax may require knowing at least one of the dimensions at compile time, whereas now it is a runtime argument. So that might be a non-starter, unless you're using C++ and want to write or use a matrix class instead of plain arrays.

C++ and Eigen: How do I handle this 1x1 matrix?

Consider this excerpt:
for(int i = 0; i < 600*100*100; i++) {
( 1 / 2 * (1 - a) / a * x.transpose() * y * (z + (1 - a) *
z.transpose() * y(i) / z.sum() ) * x.transpose() * z );
}
In the code above, x, y, z are objects of the class MatrixXd in Eigen and a is a double. Through these multiplications, eventually the outcome is a scalar. The entire forloop took less than a second.
However, if I change my code:
for(int i = 0; i < 600*100*100; i++) {
F(i) = F(i) + ( 1 / 2 * (1 - a) / a * x.transpose() * y * (z + (1 - a) *
z.transpose() * y(i) / z.sum() ) * x.transpose() * z );
}
The forloop then takes 6 seconds. F is an ArrayXd. I'm trying to update each element of F through a loop and in each iteration I would do a series of simple matrix multiplications (which would result in a scalar).
I'm not sure what's wrong. How can I speed it up? I tried to use .noalias() but that didn't help. This could have to do with the fact that the outcome of the series of matrix multiplication results in a 1x1 MatrixXd and Eigen is having issues adding a MatrixXd to a number.
Update
Per #mars, I tried eval():
for(int i = 0; i < 600*100*100; i++) {
( 1 / 2 * (1 - a) / a * x.transpose() * y * (z + (1 - a) *
z.transpose() * y(i) / z.sum() ) * x.transpose() * z ).eval();
}
And it takes ~6 seconds as well. Does that mean there's no way to optimize?
Also, I used -O3to compile.

C++ Data Structure to Find Neighbouring Values in Multidimensional Array

I have a project where I read in an array that has 1 or more dimensions, and for this project I need to be able to determine a given element's neighbours quickly. I do not know the dimensionality ahead of time, and I likewise do not know the size of the dimensions ahead of time. What would be the best C++ data structure to store this data in? A colleague recommended a vector of vectors of vectors of . . ., but that seems incredibly unwieldy.
If you know the address of which element you need the neighbors for, could you just do pointer arithmetic to find out the neighbors. For example, if p is the location of the element, then p-- is the left neighbor and p++ is the right neighbor.
Think your multidimensional array as a 1D array. Let the dimension of the array is d1 * d2 * ....* dn
Then allocate memory for a 1D array, say A of size d1 * d2 * ....* dn. For example,
int *A = new int[d1 * d2 * ....* dn];
If you need to store data in the [i1][i2]...[in] th index, then store in the following index:
A[i1 * (d2*d3*d4.. *dn) + i2 * (d3*d4*....dn) + ..... + in]
Neighboring elements will be:
A[(i1 + 1) * (d2*d3*d4.. *dn) + i2 * (d3*d4*....dn) + ..... + in]
A[(i1 - 1) * (d2*d3*d4.. *dn) + i2 * (d3*d4*....dn) + ..... + in]
A[i1 * (d2*d3*d4.. *dn) + (i2 + 1) * (d3*d4*....dn) + ..... + in]
A[i1 * (d2*d3*d4.. *dn) + (i2 - 1) * (d3*d4*....dn) + ..... + in]
.............................
A[i1 * (d2*d3*d4.. *dn) + i2 * (d3*d4*....dn) + ..... + (in + 1)]
A[i1 * (d2*d3*d4.. *dn) + i2 * (d3*d4*....dn) + ..... + (in - 1)]