Compute embeddings with 2 or less images error

[PaulHax]

Steps

1. Drag Dataset Selection images slider to 2
2. Click Embeddings Compute button
3. Python Error: ValueError: n_components=3 must be between 0 and min(n_samples, n_features)=2 with svd_solver='full'

[vicentebolea]

This is a limitation of the pca|umap algorithm, the need a number of sample >= of the number of dimensions. We might want to limit the min number of elements in the sliders.

[vicentebolea]

Closing this since this is a limitation of umap/pca

[PaulHax]

Never good to have a Python traceback error, right? Should we just put a check in the "compute embedding" function that skips things if the parameters are degenerate? For extra credit, give message to user on why they are not seeing their embedding points update?

[vicentebolea]

Good point