make interaction counter conform with output format

three points:

for multiple showers the interaction counter writes two numpy files directly in the main corsika output directory. Should go in subdir interaction-counter or something. Otherwise making a large number of showers gets messy.
does it really have to be numpy binary files? can't we just write yaml? Making it convenient to read is the job of the IO Library..
shouldn't it be optional? Or at least configurable. if I want 100k proton showers at 1 TeV I don't want 200k numpy files..

To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information