Add offset-dependent table PSF class #556
@dltiziani - Thanks for the pull request!
It's great that you're pushing forward with the HESS PSF investigations and this table PSF format specifically.
So far our strategy has been to add one class per IRF format that does I/O and other methods:
I would like to stick with that pattern also for this PSF format you're adding in this PR:
So my suggestion is that you introduce a new class. I don't know a good name at the moment, maybe just call it
It's also important to have tests that execute the functionality once and assert that it's basically working,
@dltiziani - I realise that this is quite a bit of work, especially when you're new to Gammapy development. Please let me know if you want to and have time to do this soon, or if I or some other Gammapy dev should take over this PR and continue with it.
I don't understand why you don't encourage people to use the
@dltiziani - Thank you for implementing this!
As @joleroi mentioned, he recently introduced a base class NDDataArray base class, which should be able to handle I/O and evaluation for any IRF. Currently it's only used for one IRF, EffectiveAreaTable2D, but the idea is that we use it for the other IRFs as well to avoid duplicated similar boilerplate code.
My suggestion for this PR would be that you add one or two basic tests to this PR first to establish and show that it works:
You could also add a script
@dltiziani - It would be great to have this in
The main thing before it can go in would be to fix the travis-ci issue with the Sphinx reference.
The class now appears in the online docs:
@dltiziani - I've added a changelog entry for this PR: