D. Benesch

@DominiqueMakowski > i think it's fine, perhaps to make it clearer we can have an internal function called _signal_interpolate_nan() and if y_values is None and x_values contains NaN, we directly...

> We need to find a way to deal with super long signals, do we display it all? do we display a bit to keep the scale reasonable? @DominiqueMakowski if...

@DominiqueMakowski > We'd need to start with something small that has only a handful of methods / figures How about we start with `ppg_process()`, with raw/cleaned/peaks + heart rate plots...

Is this along the lines of what you had in mind? https://danibene.github.io/neurokit2-ppg-report-example/myreport.html

> this way, users can also get this info in the console / in a re-useable programmatical form That would be good! It's currently only possible to use the `elgendi`...

I'm not sure how **kwargs could be used in case you want to customize both `ppg_clean()` and `ppg_findpeaks()` (without some hack: e.g. using inspect https://stackoverflow.com/a/13131673) How about using 2 separate...

This reference uses consistency between ECG & PPG heart rates in determining signal quality: https://academic.oup.com/jamia/article/13/3/309/699677

I personally think that would be a great feature! Would you be able to/want to implement it?

> Fixing peaks is a tricky issue Indeed! I've opened the PR for now with the 3 changes I mentioned above, it's a work in progress but I thought I...

Hi @DominiqueMakowski & @JanCBrammer , I've noticed a couple of problematic parts of the code from [the previous PR](https://github.com/neuropsychology/NeuroKit/pull/651) 1. If the intervals are standardized, the calculation of the number...