Add other cameras to toy waveforms generator

[HealthyPear]

In testing new image extractors, it is possible that some of them at the beginning are optimized for a limited number of cameras (like the extractor in PR #1215, which is optimized for the moment for LSTCam and NectarCam).

The function camera_waveforms in ctapipe.image.tests.test_extractors.py seems to be fixed to work with SST-ASTRI/CHEC.

It would be useful to add also the other cameras for optional testing.

[kosack]

it would be good to have at least one hexagonal camera in there to make sure they work. Best would just be to use all cameras, or at least all CTA cameras (those that have a reference pulse at least).

[watsonjj]

At some point doesn't this become more of a benchmarking exercise than a unit test? The current tests ensures the methods run, and produce outputs you expect. The proposed changes are asking to make sure that camera-specific values (e.g. the tailcut thresholds used in #1215) are optimised.

I don't get the point about hexagonal cameras. Nothing about the ImageExtractor is specific to the pixel shape.

[kosack]

I was more thinking about less regular neighbor lists (for hex cameras) since the neighbors are used, but I guess it's not a problem

[kosack]

@HealthyPear if you want to test your new extractor using multiple cameras, you can of course do that. Otherwise I'll close this.

[HealthyPear]

I was asking because CHEC has only one channel, so I was thinking that testing with e.g LSTCam or NectarCam, which have 2 channels, would provide a more general case.

Nonethless, with SST-ASTRI/CHEC it seems that I managed to make the new ImageExtractor work within the precision set in the unit-test.

[watsonjj]

I was asking because CHEC has only one channel, so I was thinking that testing with e.g LSTCam or NectarCam, which have 2 channels, would provide a more general case.

It is now presumed that gain channel selection has occurred before image extraction. This was a change made in #1095

[maxnoe]

But you are producing waveforms, right? Not images.