Niklas Z

Hello, Since I took over from my account @IruNikZe, I want to give an update here. So far I - unified the implementations of `WhittakerSmooth`, `ArPls`, and `AirPls` - replaced...

I'd already like to add this tiny teaser for the automated smoothing as a result of the tests. I tried to mimic two peaks with Poisson-noise sitting on a baseline....

@paucablop In case you already want to see the new functionalities of `WhittakerSmooth`, you can try it out in this notebook: [whittaker_show.txt](https://github.com/paucablop/chemotools/files/15299563/whittaker_show.txt) After you pulled the [current version of the...

I timed the implementation against [this Rust-based package](https://github.com/AnBowell/whittaker-eilers/tree/main/whittaker-eilers-py). I found that: - on a single smooth level, we are slower (450 µs in the notebook vs. 60 µs) - however,...

Hi @paucablop 👋 You said you had many use cases, but the automatic smoothing requries weights and I was not sure whether you can get weights easily. Usually one uses...

@paucablop After I've seen #56, I renamed the `window_size` of the function to `window_length` as for SciPy's `savgol_filter`. Here is an updated notebook that works with the current branch version:...

@paucablop You're welcome! 😸 I will quickly rename it back to `window_size` then 💪

@paucablop Renamed and tested 👍

Hi, @acmoudleysa > The trick is to self._validate_data(..., reset=True) inside the fit method, which sets the n_features_in_ and to self._validate_data(..., reset=False) inside transform which makes sure the feature length matches...

Hi, I tested the new version of `ArPLS` refactored for #44 and it works like a charm for `dtype=int` without any prior conversion outside the class:  One can see...