Too small n_particles in config causes particles to be omitted

[mortele]

Specifying the wrong n_particles (too small) in config.toml causes the remainder of the particles to be ignored because input_parser.check_n_particles is called after the "indices" dataset in the input .HDF5 file is read.

We should remove this keyword from the config and always read the number of particles from the length of in_file["indices"].

[mortele]

We should be able to fish out the number of particles in the input .HDF5 file cheaply by

n_particles = in_file["indices"].size

without having to read the entire dataset into memory or iterating over it.