CadQuery Varibles Editor doesn't set variables

[robodan]

In b2, a simple cadquery script can set the global variables through the CadQuery Variable Editor. This works fine for me in FreeCad.
b2.txt

[ edited: uploaded bm2.txt was the wrong file. See my reply below with correct cqpart code]

Ubuntu Linux 18.04 amd64
FreeCad 0.17
Cadquery 1.2.0 pip install
cqparts 0.2.0 pip install

[jmwright]

@robodan This may be more of a CadQuery/FreeCAD Module issue than a cqparts one. The variables in your second script are made up of objects rather than being simple int, float, string, etc variables. At this time, CQGI (which the Variables Editor in the FreeCAD module relies on) only understands simple variables.

[robodan]

Oops, I included the wrong file. It should be this (which I can't seem to upload as a file).
I only expect config and scale to be configurable. They are simple variables (same as b2.txt).

from __future__ import print_function
from __future__ import division
from __future__ import absolute_import

import cadquery
import cqparts
from cqparts.params import *

import logging
logging.basicConfig(level=logging.DEBUG)

config = "FFF"                            # 3D printing
scale = 1.0                               # another config test


class MyPart(cqparts.Part):
    width = PositiveFloat(5, doc="just that big")
    length = PositiveFloat(5, doc="how far we'll go")

    def make(self):
        logging.info("config: %r scale: %f" % (config, scale))
        obj = cadquery.Workplane('XY').circle(self.width / 2)
        obj = obj.extrude(self.length)
        return obj

if __name__ == '__main__':
    mypart = MyPart()
    mypart.exporter("stl")("example_part.stl")
else:
    from cqparts.display import display
    mypart = MyPart()
    display(mypart)

[jmwright]

@robodan Can you wrap that in triple back ticks so that I can copy it? GitHub is trying to format it like markdown.

[robodan]

Should be fixed now.

…

On Thu, May 24, 2018 at 11:50 AM, Jeremy Wright ***@***.***> wrote: @robodan <https://github.com/robodan> Can you wrap that in triple back ticks so that I can copy it? GitHub is trying to format it like markdown. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#95 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AA4ZVTLCUlwM1-etimpxHktLmN95dub6ks5t1wD_gaJpZM4UMvuQ> .

[jmwright]

Your first script uses the show_object function which taps into CQGI, and your second script uses cqpart's display function which doesn't hook into CQGI. That's why the second script's variables aren't dynamically editable. There have been recent discussions on portability issues like this. I think that everyone would like things to work the same whether you're using stock CadQuery or cqparts. For now, I was able to make your second script work as expected by commenting out the display line and putting a show_object line right below it.

[robodan]

show_object of what?
This errors for me: show_object(mypart)

[jmwright]

Sorry, you have to pull local_obj out of mypart to show it through show_object. If it's an assembly I think you'd have to pull out global_obj, but @fragmuffin would know better than I on that.

show_object(mypart.local_obj)

[fragmuffin]

@robodan yep, what @jmwright should work.

I think to leverage the current jmwright/cadquery-freecad-module with cqparts, you'd need to pass globally defined variables to the part's constructor.

g_width = 5
g_length = 5

mypart = MyPart(width=g_width, length=g_length)
show_object(mypart.local_obj)

from cqparts.utils import CoordSystem
mypart.world_coords = CoordSystem(origin=(0,20,0))  # (see note below)
show_object(mypart.world_obj)

@jmwright is that wright correct?

note: requesting world_obj before setting world_coords will raise an exception.
read this for details.

[jmwright]

@fragmuffin Correct. The global variables will be accessible to CQGI, and passing them into MyPart serves the original intent.

[fragmuffin]

@robodan I'll close this issue with the above workaround.
This is also not strictly a problem with cqparts, but the cadquery.cqgi.

I've raised dcowden/cadquery#273 because this is not fixed, but the discussion needs to move to the right place.

Thank you for the feedback @robodan !
cqparts is moving fast, and I'm so glad people like yourself are critiquing it