Aizatulin

The ramp is just a function, which takes a float value. In my version it is the u-value of the position on the curve, where the point has its closest distance to. For example, if this is the center of a line (where the point has its closest distance) the u-value is 0.5, The rotational effect comes from the input value (which comes from the curve) and not from the ramp. Another approach can be just rotating the box (outside) and after this applying the ramp function and then rotate the box back. You can just use the vop from your example. I'm strongly biased to VEX (it is much faster and easier to use - but this just my personal opinion). Here is another file (1 line VEX to capture input and set detail attribute) test_VOP_rotate.hipnc

You can do the same in VOP, it is only about the xyzdist() function, which is doing all the work in the background. The only thing is, that you need a second input curve. If interested, here is a VOP version. Probably there simpler more straightforward solutions. test_VOP_xyzdist.hipnc

Hi, one possible way might be using a curve to capture the u-attribute (with xyzdist() in a volume wrangle for example). Here is an example, using a float ramp (for density control) and an external curve for rotational control. test_capture_curve.hipnc

Hi, acos(dot(v1,v2)) is always getting the smallest angle between two vectors. If you want to get orange angles, you have to add an orientation information like a point normal. With the normal and the first edge you can construct a plane, to determine, if the second vector is on the positive or negative side. getHoleAngles_mod.hipnc @edit: I am not sure, if this solution works in every case, because it probably depends if the hole is on the left or right side <- but this should work because the angle will change aswell ....

Hi, you can also use intersect_all function, to detect the number of intersections, which can help you to find the inner prims, but this function works better with 3d geometry, so it might be better to extrude your shapes before you use the intersect function. 2D_shape_intersect_all.hipnc

Hi, boolean seems to work for me, you can also group->dissolve edges with small angle and fuse close points (optional). Polybevel_bool.hipnc

Hi, there is also a straight skeleton 2D beta version in SideFX Labs.

Hi, if interested, here is an older example (a bit modified) where I've tried to create a circular bevel for polylines. It is experimental so it will probably fail in some situation, but you can use custom radii and numbers of division. Polyline_Circular_Bevel.hipnc

Hi, it works for me, if I turn on "Minimum Distance" method instead of "Project Rays" in the ray sop. btw: if you take the idea using the closest distance as radius, you can calculate the greatest radius of each circle, which has the prim centroid as center an touching at least one line of the primitive (but without overlap). This radius can be used to set pscale for copy to points node aswell. ee_circlepack_ctp.hipnc

Hi, you can use birail, here is a modification of my birail example. You can use half circles as cross section (quite similar to sweep, the only difference is, that you have two guide curves instead of one). It is not necessarily nurbs (it can be poly aswell). test nurbs_birail.hipnc

Sure , if you want to avoid the foreach (and fuse), you can also use a single detail wrangle (using VEX), but the results are the same. Advantage of VEX: it is much easier to add custom functions, which can do specific stuff, which is harder to achieve with inbuild nodes. But this is of course my personal biased opinion. hexa_growthA.hipnc

Hi, here are some ideas: start with a single point (id = 0) give each new point an id based on iteration (in the loop) add 6 new points for each old point, which are on a circle with equal distance etc... -> problem: there are more points than needed fuse seems to work (smaller ids seems to dominate with least target point) hexa_growth.hipnc

Hi, if you have a string "s" you use len(s) to get its length and s to get access for a single char. string_to_array.hipnc

Hi, you can define an array, which contains all number you want generate. Now fit a 0-1 random float value to (0,len(array)) and use the truncated value as index for the array. random_array.hipnc

Hi, you can capture every curve independently for example and store all attributes into arrays, but I'm not sure, if this working well. I've substituted the nearpoint values of your example by the xyzdist() values, because it can use the primgroup restriction directly. All attributes are used as arrays in a for loop. wavelike_deformA.hipnc