Accept input from 'LASY' laser python module

[Nkehoe-QUB]

The problem and the context

Accept laser initialisation from the LASY Python module.

https://lasydoc.readthedocs.io/en/stable/index.html

The solution

Failed attempts

Trying to implement into existing laser blocks isn't straight forward or, possibly, doable.

[beck-llr]

Hi.
It is actually doable with python calls in the namelist (I have already done it) but it is not automated and have to be adjusted to each case.
A full automated support would be great but unfortunately none of us has the time to work on that in the short term at least.

Here are some snippet that I used that might help.
Recover the laser envelope and mesh properties:

#Read values from propgation and decomposition
f = h5py.File('lasy_laser_He_05_000002.h5', 'r')
dset = f['data']['0']['meshes']['laserEnvelope']
data = dset[0,:,:] 
#Time and space steps
datadt,datadr = f['data']['0']['meshes']['laserEnvelope'].attrs.get("gridSpacing")
#Number of points in time and r
datant,datanr = data.shape

#According to the polarization, the data stores Ex in the openPMD standard. That is Ey in the Smilei standard.
#The corresponding B component is Bz = Ey/c
dataB   = data/cst.c # B envelope

Reconstruction of the field:

### Data interpolator ###
#Time and radius axis of the data
datat = np.linspace(0, datant*datadt, datant)
datar = np.linspace(0, datanr*datadr, datanr)
#Complex envelope interpolator
B_interp = RegularGridInterpolator((datat, datar), dataB)
#Bz component reconstruction from the complex envelope: Bz = Re( B exp(-iw_0t) )
nprect = np.vectorize(cmath.rect)
def Bz(t,r):
    return (B_interp((t,r))*nprect(1, -t)).real.astype(complex)

def br_mode0(r,t):
    return 0.j
def br_mode1(r,t):
    return 1j * Bz(t,r)
def bt_mode0(r,t):
    return 0.j
def bt_mode1(r,t):
    return Bz(t,r)

And the laser block becomes:

Laser(
    box_side         = "xmin",
    space_time_profile_AM = [br_mode0, bt_mode0, br_mode1, bt_mode1],
)

Hope this helps. It is not directly using the Lasy python module but it is a lazy workaround :-) Just be very careful about the normalizations.