Component Copy (CAD Instances)

Change directly flows on from discussion: #90 (also relates closely to #96)

The target is for the following code-fragments to evaluate as they’re shown

Implicit Part behaviour

>>> c1 = Cube(size=10)
>>> c2 = Cube(size=10)
>>> c1.local_obj is c2.local_obj
False
>>> c1 == c2
True
>>> hash(c1) == hash(c2)
True
>>> c1.local_obj = c1.local_obj.faces('>Z').hole(1)
>>> c1 == c2
False

• done

Explicit cqpart.Part instance modifier

>>> from copy import copy
>>> c1 = Cube(size=10)
>>> c2 = copy(c1)
>>> c1.local_obj is c2.local_obj
True
>>> c1.local_obj = c1.local_obj.faces('>Z').hole(1)
>>> # OR
>>> #c2.local_obj = c2.local_obj.faces('>Z').hole(1)
>>> c1.local_obj is c2.local_obj
True

• done

>>> c1 = Cube(size=10)
>>> c2 = copy(c1)
>>> c1.world_coords = CoordSystem(origin=(10,0,0))
>>> c2.world_coords = CoordSystem(origin=(-10,0,0))
>>> c1.world_obj = c1.world_obj.faces('>X').hole(1)
>>> c2.world_obj = c2.world_obj.faces('>Y').hole(1)
>>> c1.local_obj is c2.local_obj
True
>>> c1.world_obj is c2.world_obj  # would never be True
False
>>> display(c1.local_obj)  # cube
>>> display(c2.local_obj)  # cube
>>> display(c1.world_obj)  # cube with hole along X axis
>>> display(c2.world_obj)  # cube with hole along Y axis

• done

Documentation

• review part build cycle docs
• document copy behaviour

PR

• ping @gntech @zignig @dcowden @zwn for review

[gntech]

I dont understand the workings of cqparts fully but what would be wrong with this:

What if c1 = Cube() only creates the local obj and c1 is then wrapped in a = WorldPart(c1) class that has the woorld coordinates and makes a world obj from the local obj? b = copy(a) would have the same reference to c1 but its own world coordinates and own world_obj.

Cube() only creates the actual geometry. WorldPart() wrapps a geometry and define world coordinates.

[gntech]

On second thought, I agree as well, keep it as pythonic as possible