Add radius arc

[gntech]

Add an function that draws an arc that is determined by startPoint, endPoint and radius.

[coveralls]

Coverage decreased (-0.1%) to 92.203% when pulling d75ed82 on gntech:add-radiusArc into 86f3b7c on dcowden:master.

[AppVeyorBot]

✅ Build cadquery 1.0.21 completed (commit f6c83eb355 by @gntech)

[codecov-io]

Codecov Report

Merging #254 into master will decrease coverage by 0.1%.
The diff coverage is 85%.

@@            Coverage Diff             @@
##           master     #254      +/-   ##
==========================================
- Coverage   92.85%   92.74%   -0.11%     
==========================================
  Files          10       10              
  Lines        2196     2206      +10     
==========================================
+ Hits         2039     2046       +7     
- Misses        157      160       +3

Impacted Files	Coverage Δ
cadquery/cq.py	94.12% <85%> (-0.35%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 86f3b7c...d75ed82. Read the comment docs.

[dcowden]

@gntech thanks for another contribution! this looks great!
Is it ready to merge in from your view?

[gntech]

Actually, I was thinking about some things.

1. For example: If the linear distance between startPoint and endPoint is more than 2 * radius the arc will fail since it it geometrically impossible. If this is the case I would like to raise some kind of Error and
   present it to the user in a logical way. Do you already do this kind of sanity checking of the parameters for other functions? Can you point me to some example?

if endPoint.sub(startPoint).Length > 2 * radius:
    raise ValueError

2. The tests I added for the radiusArc and sagittaArc functions are basically just a test that the arc functions dont throw any error. They don't actually test if the result is correct. Do you have any suggestion how I can improve the test cases for the arc functions?

[jmwright]

Do you already do this kind of sanity checking of the parameters for other functions? Can you point me to some example?

I think that rasing a ValueError is how we do it elsewhere. See here for a case that's inside the CQ core code, and here for a case that's in an example.

Do you have any suggestion how I can improve the test cases for the arc functions?

Not sure if this will do exactly what you're looking for, but you can check the length of edges. The last line in this sample code is the important one.

import cadquery as cq

(L, H, W, t) = (100.0, 20.0, 20.0, 1.0)

# Define the points that the polyline will be drawn to/thru
pts = [
    (W/2.0, H/2.0),
    (W/2.0, (H/2.0 - t)),
    (t/2.0, (H/2.0-t)),
    (t/2.0, (t - H/2.0)),
    (W/2.0, (t - H/2.0)),
    (W/2.0, H/-2.0),
    (0, H/-2.0)
]

result = cq.Workplane("front").moveTo(0, H/2.0) \
                              .polyline(pts)

# Do check on edge lengths here. May have to get the radius indirectly
print(result.edges('>Y').first().val().Length())

Notice the .val() call to get the underlying shape object. The higher level Workplane object doesn't have functions like Length() attached to it.

[jmwright]

@gntech Also, after looking at the website you have in your profile link, you might want to check out issue #173 @RustyVermeer has been working on DXF import, which may help you a lot, especially if the 2D profiles in your designs get complex.

[fragmuffin]

I would create a series of arcs and assert:

• start, end, and midpoint locations
• wire type (edge.geomType() == 'CIRCLE')

This apply_spline() function iterates along an arc or a spline getting vertices at different lengths along the wire.
It shows how to get the length of an arc.

[fragmuffin]

Suggestion for ValueError implementation

[fragmuffin]

As you mentioned, if the radius < length then the arc is not possible.
I think the best way to capture this is with a try block over the sag calculation

try:
    sag = abs(radius) - math.sqrt(radius**2 - length**2)
except ValueError:
    raise ValueError("Arc radius is not large enough to reach the given end-point")

[gntech]

@jmwright and @fragmuffin: Thanks! Good suggestions, I will try to get this implemented later this weekend.

[gntech]

See my latest commit, I discovered something strange with the toWorldCoords function. When runned in sequence it changes the coordinates. I dont think that this is the intent?

[AppVeyorBot]

✅ Build cadquery 1.0.22 completed (commit 3778358d0e by @gntech)

[gntech]

I realize that I made some wrong assumptions about what toWorldCoords do.

But this is my intention: sagittaArc, radiusArc and threePointArc all uses startPoint and endPoint but determines the arc between them in different ways: sag, radius or midPoint.

My idea is that sagittaArc and radiusArc should just calculate the corresponding midPoint and then hand that midpoint over to threePointArc for the actual arc creation.

Does this sound reasonable or is it prefferred to have separate functions that create the arcs by themself?

I will try to write some code later tonight to better explain.

[jmwright]

Does this sound reasonable or is it prefferred to have separate functions that create the arcs by themself?

I don't know what @dcowden thinks, but I think your approach is fine. I try to reuse existing building blocks whenever possible, which it sounds like you're doing.

[dcowden]

@gntech i like your approach.

[AppVeyorBot]

✅ Build cadquery 1.0.23 completed (commit e6768412b8 by @gntech)

[dcowden]

@gntech this looks good, +1 from me. Are you ready to merge?

@jmwright does this get +1 from you?

[gntech]

Yes, I consider this ready for merge

[jmwright]

+1 from me

[dcowden]

@gntech thanks for this contribution! You rock!

[gntech]

Thank you! And thanks for your work on cadquery which is awesome!

[dcowden]

@gntech absolutely! thanks for the kind words.

I think we can do really amazing things with cq once we get an OCC layer implemented, and we get proper feature trees and such.