Create notebook equivalent to dist_thresh_vs_acc.ipynb but for seqScan

[paco-barreras]

• read the paper definining seqScan (https://dl.acm.org/doi/abs/10.1145/2666310.2666417)
• read on the experiment in the overleaf
• code based on pseudocode, or rely on sklearn dbscan
• branch off of exp_hdbscan_paper

[paco-barreras]

Sliding window W = [start, end] , initialized at [t_0, t_0].

Perform DBSCAN on the points falling in W and:

• If you get cluster labels of -1 and 0 only (just one or no clusters), then grow W by augmenting end.
• else, meaning you got -1, 0, and 1 labels, we want to perform a "cut" and "log" the first cluster (just means a permanent incremental label for the points in 0). The cut is the time of the largest ping in cluster 0 that is also a core point. AND then start becomes that cut point. And you permanently label those points in 0 with current_label and do current_label += 1

[paco-barreras]

part of #234