[WIP] Implemented estimation of particle size, peak picking and migrated functionality from 'xrd' package

[kjappelbaum]

• as discussed in should be peak picking a method of common-spectrum or method specific? common-spectrum#1, we might want to migrate the peak picking and annotation to common-spectrum as this is not specific to XRD and can be probably be reused by other spectra types, which should make maintaining easier
• I still need to test the annotations (hence WIP)
• For the UI, we might have some default values with explanations for the shape factor
• We should support and the annotation should be saved. If people use this, it would be a quite neat way to "mine" if a pattern is phase-pure or not
• For the profile fitting, we currently support only Gaussian and Lorentzian. Pseudo-Voigt would probably be a better choice in some cases but I didn't have the time yet to implement this in mljs
• I currently copy pasted some analyses to the tests. Maybe I should also factor this out