Only the Border Matters

Two groups of points can be split by a straight line. A support vector machine picks the line with the widest gap, and only the points touching the gap decide where it goes.

Write linear_svm(points, labels, new_points).

• Fit SVC(kernel="linear", C=10) on the points. The two columns share the same units, so no scaling is needed.
• Count the support vectors: add up the model's n_support_.
• Predict the label of every point in new_points, in order.

Return a dict {"support_vectors": ..., "predictions": [...]} with an int and a list of ints.