Does svtyper calculate read pairs which have long overlap twice?

[xyw1]

Hi Dave, I read you script, and noticed that RO is calculated by ref_seq+ref_span. and ref_seq is calculated by the following code.

#  svtyper/svtyper/singlesample.py line 255 to 259 
is_ref_seq_A = sam_fragment.is_ref_seq(read, None, chromA, posA, ciA, min_aligned)
is_ref_seq_B = sam_fragment.is_ref_seq(read, None, chromB, posB, ciB, min_aligned)
if (is_ref_seq_A or is_ref_seq_B):
    p_reference = prob_mapq(read)
    ref_seq += p_reference

I'm wondering that if this read has a paired end read that overlaps with it a lot (shown as the illustration below), will svtyper take read1 and read2 as two independent observations?