Charles

@ericpre Thanks for the ping: I'll look at it more today. I definitely won't be making changes that affect my current user base, so I'm working to just enable its...

@francisco-dlp @ericpre @AndrewHerzing @jat255 255 See PR #33 * I wasn't sure if opening a PR was the usual way of counter-offering a PR Per https://github.com/hyperspy/hyperspy/pull/2172#issuecomment-625770320 ```python from pymcr.mcr import...

I admit: I had no idea that was an issue. I will make those changes as soon as possible. Thank you!

Hi @ClarkAH thanks for the effort! I somehow missed the notification. What's the status of this? I see it's failing tests but I will look into that. Thanks.

Hi Nicolas, Yes, you can do unimodality with a little bit of coding. In your case, does the mode need to follow an analytical shape or is it just enforcing...

@colapili Hi Nicolas, That's a great question and one that I don't think has a singular answer. I have seen it result from preprocessing steps, I've seen it when some...

@colapili Here is an example function that can be used to make a unimodal constraint (see if you can get it going using this example: https://pages.nist.gov/pyMCR/usage.html#single-gaussian-distribution-spectral-fitting-with-lmfit as a template) ```...

@AndrewHerzing @jat255 In general, all 0's is a no-no (in all sorts of optimization methods) because there is no instability to cause the algorithm to move/evolve. Better is to seed...

@AndrewHerzing @jat255 I may have misinterpreted "spectral channel". Do you mean portions of spectra or do you mean, e.g., you seeded in 3 initial guess spectra and 2 of them...

@jat255 So by channels you mean a particular energy bin on the CCD? So across the entire image, with (e.g.) 1024 energy/frequencies channels/bins, it is identically zero across the whole...