You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When zooming in deeply and doing a polygonal selection it seems that the decision if a point belongs to a cluster (contains()) becomes problematic: The polygon clearly contains more than 2 sequences.
This seems to happen only when zooming in very deeply so it may not be a frequent problem. Nevertheless, this needs to be checked/fixed.
The text was updated successfully, but these errors were encountered:
This seems to be related to the rounding which is necessary for Awt.Polygon.addPoint(int x, int y). Since we need to provide int but the mouse clicks produce double, we have a significant loss of precision here. The attached figure illustrates this nicely: two of the y-values get rounded to 64 (see debug message in background terminal) and thus only one point is left. The upper points have y > 64 and are thus ignored.
While the scale-factor (scale) in DataSetFactory.createDataSetFromPointFile(new FileInputStream(inpointsfile), labelsIS, scale); allows to overcome this to a certain extent (simply use a "large enough" scaling factor) there might always be situations where the zooming is so deep that the rounding will cause problems.
When zooming in deeply and doing a polygonal selection it seems that the decision if a point belongs to a cluster (
contains()
) becomes problematic:The polygon clearly contains more than 2 sequences.
This seems to happen only when zooming in very deeply so it may not be a frequent problem. Nevertheless, this needs to be checked/fixed.
The text was updated successfully, but these errors were encountered: