absorption.cplx.x created files with special chars in names when using extended_kernel etc

[mike marchywka]

I realized jdftx was going to take some effort and I wanted some

initial results so I took the CO example and an old serial install of

espresso ( I copied it over from old machine without rebuild lol )

and duplicated the water example although geometry is not optimized.

After some initial apparent success ( plausible output without crashes)

I tried to include the coupling beyond TDA and remove the hdf5 output.

Leaving in full_bse but removing the hdf5 thing and extended_kernel

appears to run.  I changed the polarization  as an initial run claimed something

about zero oscillator strength and wasn't sure about geometry :) 

fwiw, I found this recent work for comparison in 10ev range for H2O,

https://pmc.ncbi.nlm.nih.gov/articles/PMC11371026/

Right now anyway I have the eh and noeh output to play with will try

to check cross section at some point :) 

cat absorption.inp
diagonalization

number_val_bands_coarse 4
number_val_bands_fine 4
number_cond_bands_coarse 32
number_cond_bands_fine 32

no_symmetries_coarse_grid
no_symmetries_fine_grid
no_symmetries_shifted_grid

screening_semiconductor

cell_box_truncation

use_momentum
polarization 1.0 0.0 0.0

gaussian_broadening
energy_resolution 0.15

skip_interpolation

full_bse

#extended_kernel

#dont_use_hdf5_output

eqp_corrections
write_eigenvectors -1