Remove Cherenkov threshold for LS detectors

[JostMigenda]

Fixes #52. As a side effect (because the allowed eNu range for certain channels is now slightly larger), the event rate in es and c12e/c12eb channels increases by a few per cent for LS detectors.

(Not 100% happy with how I implemented this, but I think that’s about as good as it gets without a larger restructuring of the code base. At least I got to clean up the main() implementation slightly.)

[JostMigenda]

Hmm … the integration test is failing under Ubuntu; the only difference I can see from the logs is in the c12e channel:

[c12e] Generated 64 particles (expected: 66.86 particles) (macOS/Windows)

[c12e] Generated 64 particles (expected: 66.82 particles) (Ubuntu)

Not sure what's going on there; I'll look into it (but am on annual leave this afternoon, so this will need to wait until next week).

[JostMigenda]

Fixed the issue—turns out, ProcessPoolExecutor uses different ways to create new processes depending on OS and Python version. macOS and Windows use spawn by default, which creates a blank slate process; Linux uses fork which inherits existing module imports from the parent process (so updating sntools.interaction_channels.cherenkov_threshold inside gen_evts() was too late).

@mcolomerm Please go ahead and try this branch now; I’ll merge it sometime after the Easter holidays if you don’t discover any further issues.

[mcolomerm]

Fixed the issue—turns out, ProcessPoolExecutor uses different ways to create new processes depending on OS and Python version. macOS and Windows use spawn by default, which creates a blank slate process; Linux uses fork which inherits existing module imports from the parent process (so updating sntools.interaction_channels.cherenkov_threshold inside gen_evts() was too late).

@mcolomerm Please go ahead and try this branch now; I’ll merge it sometime after the Easter holidays if you don’t discover any further issues.

Thanks @JostMigenda for having a look at this. I will try to test it on the coming days.