Bug fix in SimpleRate

[soso128]

Hi, Sorry I inadvertently removed the Snowglobes inheritance in the SimpleRate class. This commit adds it back.

[soso128]

I also fixed a bug in the Convert_to_ROOT.py script: the script worked fine for a single detector but didn't update interaction channels properly when processing multiple detectors.

[JostMigenda]

Hi @soso128, thanks for fixing this so quickly! Two things:

1. Can you take a look at https://github.com/soso128/snewpy/blob/soso128_test/python/snewpy/snowglobes_interface.py#L300 – I think a few words accidentally got deleted there; it should probably say “Path to the SNOwGLoBES installation”, just like in the base class?
2. Could we add a few tests for this, to ensure we immediately notice if this part of the code ever breaks again? (See python/snewpy/test/test_snowglobes.py and …/snowglobes_integrationtest.py for similar tests that could be adapted.)

[soso128]

Hi, thank you for your feedback! I fixed the docstring and added a code to the test suite (is it automatically accounted for? I didn't see any test file list so I assumed so)

[soso128]

Ok, after a few issues the tests are now incorporated and seem to run fine.

[JostMigenda]

Excellent, thanks for the changes!

Regarding the test suite, SNEWPY has two separate sets of tests: unit tests (which run automatically on every commit) and integration tests (which only run when manually triggered, e.g. by requesting a review as I did just now).

Both unit and integration tests automatically include all tests in files which follow the conventional test_*.py file name pattern. The simplerate_integrationtest.py which you added doesn’t match that (like the previously added snowglobes_integrationtest.py), so you need to manually add it in https://github.com/SNEWS2/snewpy/blob/main/.github/workflows/integration.yml#L58. (Note that GitHub is a bit restrictive when it comes to editing workflows; if you run into issues, please try it in the web editor.)

One final question: I notice that the number of events in the two crosscheck tables—one used with SNOwGLoBES, one with SimpleRate—is within rounding error for most detectors, but differs by 3× for HALO and over 20× for IceCube. What’s going on there? Is that simply because the difference between smeared and unsmeared event numbers is unusually large for these detectors? Or is there something more complicated going on?

[soso128]

Hi Jost,

Thanks for your reply, I edited the workflow (which got associated with its own pull request it seems) and added the import for test_snowglobes.py. I was also very surprised to find such large discrepancies between the smeared and unsmeared rates for some experiments. I verified that, using my scripts, the smeared rates corresponded to the ones quoted in the test code, which is why I ended up trusting the unsmeared rates. But yes, I am still very surprised by such a large discrepancy. For IceCube it might be due to the fact that the energy smearing is actually huge.

[soso128]

Sorry for having given you more work and thanks for fixing the remaining issues!

[JostMigenda]

Don’t worry, it happens. (Maybe that’s a good reminder to run tests locally, too—the GitHub Actions are an important safety net, but running tests locally makes it much quicker to discover and fix these tiny bugs.)