Checkpoints with external beams about fbpic HOT 2 CLOSED

Hi Soeren,

Sorry for my late reply, and thanks for the feedback.

It is true that it would be nice for restart_from_checkpoint to reinitialize a fresh Particle object for every species in the checkpoint openPMD file. However, in practice the openPMD file is missing a few information that are needed to initialize a fresh Particles in a robust manner. For instance, the openPMD file does not tell whether the particles should be continuously injected or not, and if they should, it does not contain the parameters of the injection (density profile, etc.).
For the case of the electron bunch, it would be fine, as this species is not continuously injected and thus (I think) the openPMD file would contain all the information needed for initialization of a new Particles object. However, this would not generalize to other species.

Therefore, for now, I would prefer the solution in which we put add_laser/add_elec_bunch outside of the if block. It is not ideal (e.g. since we are doing extra work to calculate the space-charge field which is then overwritten by restart_from_checkpoint), but at least it is robust.

Does the above make sense, and if yes, would you be willing to do a corresponding pull request?

from fbpic.

I see the problem. I've created a pull request with the changes I suggested above. Again as this such a minor issue, if you are not happy with the change feel free to reject the PR.

from fbpic.