MO-coefficients get mangled when written and read again about trexio

[eljost]

Thanks for getting back to me so quickly. I seem to have misread the shape of the coordinate array ... Good to know the library is fine and that the error is on my side.

[scemama]

Hello,
I think that you are confused by the column-major ordering :-)

In the documentation, the nuclear coordinates are:

coord	float	(3,nucleus.num)	Coordinates of the atoms

So in Python, if you use the default ordering (row major), you are supposed to make a numpy array of shape (nucleus_num, 3). This is indeed what you do:

coords3d = np.zeros((2, 3))

and everything is fine.

For the MO coefficients, the documentation states that it is:

coefficient	float	(ao.num, mo.num)

So in Python, you should use a numpy array of shape (mo_num, ao_num) where mo_num = 2*ao_num in your particular case. But your MO coefficients array in Python has shape (ao_num, mo_num), which is wrong.

Thanks for giving feedback. As you are the 2nd person getting confused with this, we will try to make things clearer in the documentation.

Sorry for the confusion...