reduce the voracity of the directional clustering: given the much lower occupancy, make the selection for this much tighter (higher accuracy, poly order up to 2, lower thickness, etc)
raise the gathering parameters of the unbiased dbscan for the rest
tune the noise thresholds a bit
Miscellanea:
new field in the LNGS runlog
random changes in several scripts for plotting stuff of the LIME overground performance paper
Main tuning of reconstruction:
Miscellanea: