The best way is to first run a Telemac-2d simulation on the same grid and then you can give the results file to Tomawac, without programming anything.
Otherwise you can code what you want directly in couuti.f by filling the arrays UC1,UC2,VC1,VC2 for the NPOIN of your mesh at each time step AT. You can take inspiration from meteo.f, but I think spatial interpolations are not done... If you plan to do it by yourself, you can use the subroutine fasp.f for instance, but be aware that your resulting currents may no longer have physical coherence.
Regards,
Laurent