Issue with Different Orbital Counts and Errors in BlockBandDiagonalizer

Dear Quanty developers,

I am trying to calculate the XAS using the wannier approach as explained in the tutorial. My crystal structure contains two iron sites. Both sites have the same coordination number and the transition-metal neighbors are located at almost the same distances for both sites.

After reading the clusters and printing out the RTB, however, I obtain different numbers of orbitals: around 20 orbitals for site 1 and 23 orbitals for site 2. I don’t understand why this happens.

During debugging I noticed an error arising from the function BlockBandDiagonalizer(). For site 1, the error is:

Error in CompactMatrixRemoveRow, tried to remove row 2, but there are only 2 rows (counting starts at zero)

For site 2, the errors are:

Error in CompactMatrixRemoveRow, tried to remove row 32764, but there are only 5 rows (counting starts at zero) Error in CompactMatrixRemoveRow, tried to remove row 32764, but there are only 5 rows (counting starts at zero)

This function is not documented on the website.

Do you have any comments on this? Is it normal to obtain different numbers of fermionic operators for two sites with the same coordination environment? If so, what causes the extra orbitals, and where do they come from?