Sklearn wrapper

[koloiyolo]

I already created pip package that fulfills the functionality I aimed to achieve with this PR. I just thought it would be more convenient if this was integrated directly into the main package.

This implementation depends on Scikit-learn and SciPy. Given their widespread use and the fact that most DS and ML setups (e.g., Anaconda) include these dependencies by default, I believe this should not be a significant issue. Algorithm under the hood is still Yours lightweight implementation.

I know that You aimed for as little bloat as possible so if that's an issue, my package is there to fill the gap.

Key changes:

• tweak the file structure to make imports more intuitive. As demonstrated here, it shouldn't break any compatibility.
• add an easy-to-use API that feels familiar to Scikit-learn users, making it simpler for developers and researchers to work with.
• add support for Scikit-learn Pipeline and GridSearchCV(and probably other similar estimators).
• put the unit tests into their own file to keep things organized.
• changed how pycodestyle and pytest are run—check out the README for the details.

Also, I named the main class in my package MiniSOM beacause I think that capitalized "SOM" makes more sense.

[JustGlowing]

hi @koloiyolo, thanks for submitting this work. I'll have a look as soon as I can.