Add a possibility to load chronics from a numpy array

[BDonnot]

Is your feature request related to a problem? Please describe.

Describe the solution you'd like

Ex. This is how i would like it to be done:

import grid2op
from grid2op.Chronics import FromNPY

my_chronics = ...  # Something with FromNPY
my_chronics .init(load_p, load_q, gen_p, gen_v)
# load_p, load_q, gen_p and gen_v are computed somewhere else and compatible with the environment
# and the backend (right number of elements, right order etc.)

env_name = ...
env = grid2op.make(env_name, ..., chronics_class=FromNPY)

# etc.
...

This would make things easier for

Describe alternatives you've considered

No real alternative at the moment, the closest thing is to compute everything from the beginning.

Additional context

Be carefull: grid2op is probably initialized with class at the moment, and this makes it a bit more complex (especially when used
in parrallel environment)