Set occupations in yaml format

Dear Olivier,

the input.occ file syntax should be still active and working. See the test of the C atom in the tests/DFT/cubic/C directory of the bigdft directory.

Otherwise, you may put these lines in the input.yaml

 occupation:
   K point 1:
     Orbital 2 : 2/3
     Orbital 3 : 2/3
     Orbital 4 : 2/3

which I think are self-explanatory.

However, these occupation are important for particular purposes and in principle there should be no need to use them.
In particular for your case as you want to treat a spin-polarized system you might want to deal with traditional occupation filling, but with the inclusion of information for the input guess wavefunction. This shoudl correspond to the ig_occupation keyword of the yaml input file, which came from the input.occup file in the old format.
See the following page for details:

http://bigdft.org/Wiki/index.php?title=Input.occup

Do not hesitate

Many thanks

Luigi

Hi Luigi, thank you for the quick and precise answer.

It still didn't solve entirely my problem though.

Without your method I get this in the output (still for the isolated nitrogen atom):

   Atomic Input Orbital Generation:
   - {Atom Type: N, Electronic configuration: {
 s: [ 1.00, 1.00],
 p: [ 2/3, 2/3, 2/3, 1/3, 1/3, 1/3]}}

And with it, I get:

   Atomic Input Orbital Generation:
   - {Atom Type: N, Electronic configuration: {
 (s): [ 1.00, 1.00],
 p: [ 1.00, 1.00, 1.00, 0.00, 0.00, 0.00]}}

which is perfectly what I want.

But after the Input Guess Overlap Matrices are calculated and diagonalized, in both cases, it comes back to:

    #Eigenvalues and New Occupation Numbers
   Orbitals: [
 {e: -6.898215192636E-01, f: 1.0000, s: 1}, {e: -6.357830401437E-01, f: 1.0000, s: -1}, # 00001
 {e: -2.673313592728E-01, f: 1.0000, s: 1}, {e: -2.130941012303E-01, f: 1.0000, s: -1}, # 00002
 {e: -2.673313592728E-01, f: 1.0000, s: 1}, {e: -2.130941012303E-01, f: 0.0000, s: -1}, # 00003
 {e: -2.673299650038E-01, f: 0.0000, s: 1}, {e: -2.130926989709E-01, f: 0.0000, s: -1}] # 00004

which is clearly not the ground-state.

I don't know if it is something easy to fix or circumvent, but it would be very kind of you to help me again.

Olivier

Here is my input file if it can help you reproduce the problem:

 dft:
   hgrids: 0.31
   rmult: [5.0, 8.0]
   ixc: 11
   nspin: 2
   mpol: 1
   output_wf: 1
   gnrm_cv: 5.0e-5
 ig_occupation:
   N:
     (2s): 2
     2p:
       up: 3
       down: 0

Thanks Luigi Genovese, that solved my question.

Thank you.
I was pretty sure I did something stupid and there it was.
Have a good day!
Olivier