@cdeil FYI: For a comparison to HAP see https://hess-confluence.desy.de/confluence/display/~jking/Spectral+Fitting#SpectralFitting-Onerun-noenergydispersion-FluxPoints
@joleroi - Thanks!
In the tutorial, I don't get why you basically show the same example twice.
We have to think about what a good code organisation and API is for 1D spectral analysis.
My first guess would be that something like
The code to compute one flux point will get complex (e.g. root finding to get Rolke limits, or calling into
I didn't have a close look at this PR, and I'm not sure my suggestions are good ones or not.
Feel free to merge any time!
@cdeil the basic idea was to make the flux point computation independent of the spectral fit. That's why it's on
The thing that annoys me the most with FitSpectrum is that's is setup to do exactly one thing and it's impossible (or very hard) to get access to the individual components. So I am +10 to leave the separation between analysis steps (extraction, global fit, flux points) as it is.