Free Function Operations Leaving Edges Behind

[jmwright]

I've been trying to use the free function API more to get a better feel for it. I've noticed edge artifacts in a few cases, and so I put together an MRE.

That is produced by the following code, and I see the same visual glitch using cadquery.vis.show and CQ-editor.

import cadquery as cq
from cadquery.func import *
from cadquery.vis import show

part = extrude(face(rect(20.0, 20.0)), (0.0, 0.0, 3.0))
part += extrude(face(rect(10.0, 10.0).moved(z=3.0)), (0.0, 0.0, 3.0))

# This notch leaves the edges, no matter where I place it or how deep I cut
part -= extrude(face(rect(2.5, 2.5).moved(x=5.0, y=5.0, z=3.0)), (0.0, 0.0, 3.0))

show(part)

Is this a bug, or am I using the API incorrectly somehow?

[lorenzncode]

The Free function API provides a clean function. clean removes superfluous edges and faces. clean requires an explicit call. It is not automatically applied by default.

Workplane handle this differently. By default, "clean" is applied automatically after a boolean operation.

Say I have two Workplane objects where a boolean op is applied:

import cadquery as cq

wp1 = cq.Workplane().box(20, 20, 3).faces(">Z").rect(10, 10).extrude(3)
wp2 = cq.Workplane(origin=(5, 5, 3)).box(2.5, 2.5, 3)

result_clean = wp1.cut(wp2)  # clean is applied by default
result_without_clean = wp1.cut(wp2, clean=False)  # disable clean

Why not always automatically apply the clean operation?

The underlying OCCT functionality is from ShapeUpgrade_UnifySameDomain. Eliminating extraneous calls this functionality can improve performance. You can choose where to apply it instead of on every boolean op. It can also be useful to disable cleaning for debugging or for cases where there is a CAD kernel issue with a specific shape.

Try applying clean explicitly after the free function cut:

part = clean(part)

It's not a bug - you are probably more accustomed to Workplane with default clean=True.

Note Sketch also defines a clean method. It is an explicit call similar to free function.

[jmwright]

Thanks @lorenzncode that makes sense, I guess my mind was assuming an implicit call of clean. We should keep your answer in mind when this question comes up again later.