[Feature] Exploring other clustering approaches

[HahaBill]

[Harsh16gupta]

My findings while exploring different clustering approaches:

Currently, HDBSCAN and K-Medoids perform the best on our note embeddings. I did try X-Means as well, but it turned out to be very bad at note clustering, so I've removed it completely from the codebase.

Regarding K-Means, I was a bit confused initially and thought about removing it to focus only on HDBSCAN and k-medoids. But after doing some more research, I realized that K-Means is much faster and scales way better than K-Medoids at scale (when a user has a large number of notes). So to make sure the plugin scales well for larger vaults, I am planning to keep these three algorithms (HDBSCAN, K-Medoids, and K-Means) for now.