Enhance: Laser, Species Orbit Integration, Defaults

[tsung1029]

Hi:

I made some PICMI files in the past week and I noticed the following things about the PICMI specifications, I am wondering if these can be implemented.

(1) for the laser make it possible to use I0 as a way to specify the amplitude

(2) for the Gaussian laser “duration” is not well defined. How is it related to FWHM of the laser? I understand that in the time domain the laser needs to be completely off at some point in z (or time) so maybe we need both a variable called duration to specify the cutoff in time, and another variable for the FWHM of the laser?

(3) it would be useful to add a “method” attribute for the species as well, to specify the orbit integration method, similar to the fields.

(4) for some attributes we can have defaults, for example
method = "Yee" for fields
"Boris" for particles

cfl=.99 for the CFL condition

and so forth. Please let me know and I am happy to contribute to this effort.

[ax3l]

Discussed today in CAMPA meeting:

• consensus on the suggestion
• @RemiLehe will take on the laser definition question (2) (doc update); the definition is as in WarpX/FBPIC, but should also be documented in PICMI itself: https://fbpic.github.io/api_reference/lpa_utilities/laser_profiles/gaussian.html
• @tsung1029 will propose a PR for the suggestions

[RemiLehe]

Sorry for the late reply. Here are a few answers:

1. Agreed. Could you go ahead a create corresponding PR? I would suggest that the different picmi laser classes take I0 as an optional argument, and then convert it internally as self.E0 and self.a0, as is done here: https://github.com/picmi-standard/picmi/blob/master/PICMI_Python/lasers.py#L53 I think that we should probably not add a new attribute self.I0, otherwise this would require each of the codes that support PICMI to update their interface to the picmistandard.

2. Agreed. @tsung1029 Could you have a look at this PR: Add more documentation for the laser parameters #43 and let me know whether this now makes the definition of these parameters clear.

3./4. Sounds good. Should we actually pass this as an overall argument of the picmi.Simulation object? To my knowledge many codes do not currently have the flexibility to specify a different particle pusher for each species. Note that, in this case, the name of this new argument could be particle_pusher instead of method.

[ax3l]

Note: we use a default of cfl=0.999 now in WarpX, which I find quite reasonable to use for PICMI as well:
ECP-WarpX/WarpX#2940 (comment)