Document clustering algorithm

[ilectra]

# Check if hit with limits [x1,x2] belongs to the index-th cluster:
# The cluster low and high limits are actually ranges,
# (x1cl_min, x1cl_max) for low, (x2cl_min, x2cl_max) for high.
# The hit belongs to the cluster if BOTH its ends overlap the respective cluster
# ends within the required percentage tolerance:
# x1 within (x1cl_min-overlap_min, x1cl_max+overlap_min) AND
# x2 within (x2cl_min-overlap_min, x2cl_max+overlap_min)
# Note that overlap_min is calculated differently for "inside" and "outside"
# the cluster (min and max cluster length).