Best Of

How often have I been searching for an existing FS which I remembered being discussed here or elsewhere and got real frustrated because none of my search terms revealed the slightest hint as to which one it might be. User guidance is an important thing in any design job and why should we play hide and seek with each other when it comes to tools?

Praeterea, credo bibliothecam bene ordinatam omnium scriptorum singularium tandem Carthagini aperiendam esse!

Wow this is really cool, Ramon! It will definitely come in handy. I love that I've been able to be at least part of the inspo here.

Here's what comes to mind for improvement ideas. It's already a very useful feature as-is and this is all just my opinion, so do whatever feels right to you and what you have the time and will for:

1. it would be nice if the guides were set automatically to tangent if the surrounding geometry is also tangent. You could find those edges with some kind of adjacency query, and check with evEdgeConvexity(), then automatically match that, at least for tangent/positional. If you want to get even more extreme you could also check the curvature of relevant geometry with evCurvature() to compare it and know if you should also use curvature. I agree with others a g2 option is important for a feature like this.
2. "Make loft" and "Merge surfaces" should be true by default. In the param map you can just add "Default" : true and if you think some users might not like that you could also add "UIHint" : UIHint.REMEMBER_PREVIOUS_VALUE to keep them set to whatever that user used last.
3. I think "Edge debugging" might have more mass appeal and match the native featureset rebranded to "Preview edges". It should be false default, but remember the value. Btw I love that you added a tooltip to this one. Made it very clear.
4. The Trim and Fill options might be automatable. Am I right to think that where the edge meets a sidewall a convex one should always trim and a concave one should always fill? That way you could even handle a mismatch case like this. Currently the trim works, but the fill doesn't happen, making the part a surface body.

Amazing. I've had thoughts of such a feature many a time, but never even gotten started.

In theory a Face Blend with width control should make this feature unnecessary, but it seems to fail if the dihedral goes to zero. In your test cases, Face blend could only handle test case 4, everything else would take multiple extra features.

Any chance that you can/will make the bridging curves G2 or G3?

Thanks for checking it out!

I figured this feature is a pretty niche, but I find myself going through the workflow way too often. Some of the test cases are covered by the vanilla Fillet, but I needed some quick geometry to test trimming away end caps and filling gaps.

I can add implementation for G2/G3. It's a pretty simple change in the code. I wanted the feature to be up and running before introducing more complexity to it as I'm still really new to FeatureScript and haven't dealt too deep into programming, really.

Because of @EvanReese's series on equations, variables, etc. in Onshape, I ended up getting sucked into the FeatureScript rabbit hole. Needless to say, I'm hooked. I'm not a real programmer, but I'm still pretty happy with coming up with something that'll automate a pain point in my CAD workflows. Plus, then I get to share this with the rest of my team.

Link to FeatureScript: Loft Fillet

I created a feature to help avoid the very common (but also very niche) occurrence when that arise from creating edges that terminate on smooth edges, tangent faces, you name it. They generally result in fillets that taper to a point or else just failing altogether. Usually, the workflow to get around this is to:

1. Sweep a circle
2. Extend sweep faces
3. Trim away main body faces
4. Create bridging curves
5. Loft between the open edges
6. Remaining cleanup with booleans, fills, and additional splits

Needless to say, it's a very involved process that requires a few steps and a ton of clicking. Loft Fillet does all of that in one feature. This feature was inspired by @MichaelPascoe's Analyze Fillet. I saw he had a Split Sweep feature, but I couldn't quite get it to work as I intended. I tried to handle as many use cases as I could think of, but obviously, I can't think of everything. Appreciate any feedback that you guys might have. If anything, feel free to peak under the hood to see if there are any optimizations I could implement.