spatial_hist is a class instance

[josh0-jrg]

When using nestSpatialRateNRSource or standard flamedisx SpatialRateNRSource in the tutorial spatial_hist is a class attribute and not an instance attribute.

This is an issue for if we want multiple spatially varying sources with different distributions later on.
@robertsjames suggested you take a look @JelleAalbers

Cheers,
Josh

[robertsjames]

Actually, I've thought about this a bit more, and I think it might be okay...

Because we direct the block attributes to the source attributes (see here), provided each class correctly sets its own spatial_hist, we should be good...

@josh0-jrg, I'd suggest testing this: create a detector-specific (i.e. LZ) SpatialRateNRSource, which can contain the map-checking feature we discussed. The create 2 specific spatially-dependent NR background sources inheriting from this, with different maps. Then, check those maps don't override each other!

[robertsjames]

I think you will have to make sure that the super().__getattribute__ method resolution order does direct to the spatial_hist in the 'bottom' derived class (in my example above, the guys inheriting from the detector-specific SpatialRateNRSource). @JelleAalbers , would be useful to hear what you think

[plt109]

@Ashley-Joy, did you bump into the same issue as well? You mentioned having some troubles fitting NR with multiple sources to take care of the multiple-scatter contamination, but is it related to the issue reported here?

[robertsjames]

@plt109 did you just publicly admit to your MS woes? ;)

[josh0-jrg]

Works fine, have to assign call self.spatial_hist before the super().__init__(*args, **kwargs) when creating a source in background_sources.py - This is unlike the rate_vs_energy and energies attributes, why?

[JelleAalbers]

Sorry, I missed this issue earlier. I can definitely imagine this is confusing, the documentation is not so great on this. If it helps:

spatial_hist from SpatialRateEnergySpectrum is a model attribute. You can set this (and other) attributes either as a class attribute (when you make a custom source class) or as an instance attribute through any of the methods described in #140.

However, background_sources.py (assuming we're looking at the same one from Rob's unmerged branches) sets model attributes in the custom __init__ of a source. (Maybe because the NEST sources predate #140 ?).

I suppose that works too, if indeed it's done before the super().__init__ call. If you do it afterwards, the rest of the source setup has already happened, which e.g. calculates a converted/normalized histogram from spatial_hist. The setup does nothing for attributes like energies of FixedShapeEnergySpectrum, so that one you can even set after the init without problems.