Relevant things about the neig* class
Now that you are finishing with the k-nearest neighbors part. Some very important things:
- A model should be saved both for the reference and the full dataset
- We should keep the keys of the rows and the keys of the columns
- We should be able to choose the type of distance (metric) that is used
- We should be able to compare out-of-sample molecules without saving a model
- We should harmonize the matrix so that no negative values are given
Let's talk about it tomorrow. This is a top priority now.