Made Disk and SplineRoundDisk robust to transformations

[LasseHolch]

Disks

• Corrected diagonal_ratio
• Redefined spline curves, such they are stable, when core points are moved.

SplineRoundDisk

• Rebased the on top of Disks, such they have the same quad_map
• Redefined spline curves, such they are stable, when core points are moved.

Annulus

• added sweep angle alowing for half rings

Cylinder:

• created Half and Quater Cylinders, where symmetri patches can be named easily

[FranzBangar]

Excellent work, I'll take a look ASAP

[FranzBangar]

I tried to merge your changes but there are some issues (failing tests, type checking, etc.). Do you have your IDE prepared as instructed in https://github.com/damogranlabs/classy_blocks/blob/development/CONTRIBUTING.md? Pre-commit hook should run tests, analyze typing, sort imports, format code, ... Let me know if you need any help setting that up.

It seems your code contains a lot of lines like (self.side_1 + r_1 * np.cos(np.pi / 4)) * self.u_1, is there a way to un-repeat those? For instance, you could create a specialized parametric Curve object that would calculate that and re-use it in all round sketches. I can't tell you exactly what to do, I didn't have time to thoroughly check what your code does. I'd gladly help but I've been terribly busy last few weeks and it will stay so for the next few...

[LasseHolch]

Hi, I somehow missed your reply and didn't notice it before I was impatient and started looking for a reply. I did not run those precommit checks. Im still not used to doing things the proper way. I have now done it. python -m pytest       gave 2 errors, but nothing related to my changes.       ERROR tests/test_items/test_block.py       ERROR tests/test_items/test_wire.py mypy src tests       Gave a couple of two long line and I had to add Optional to som variables. This is corrected.              src\classy_blocks\construct\flat\sketches\disk.py:166: error: "ElementBaseT" has no attribute "origo" [attr-defined]       I needed to add the attribute origo as I wanted to be able to move the core, but still have the outer pooints defined from the original center.       That meant I had to remimplement the transformations, which gives me the proplem. Alternatively it could be added to self.parts, but I was not sure that was a good idea.       I have some problems as i base the the SplineRound on top of the DiskBase, as I need additional variables for the core_spline method.       I'm not sure why it is a problem that there are more variables in the reimplemated method. How do I best update my merge request with my small changes? Regarding the repeating code in the correct_disk method. You might have a point. Maybe I could make methrod, that worked for all 3 sketches based on the size of the core. But I fear it will be rather hard to figure out exactly what it does. I don't think I can parameterize it easily, but I might be wrong. Best regards Lasse

…

________________________________ Fra: Nejc Jurkovič ***@***.***> Sendt: 14. december 2024 18:13 Til: damogranlabs/classy_blocks ***@***.***> Cc: LasseHolch ***@***.***>; Author ***@***.***> Emne: Re: [damogranlabs/classy_blocks] Made Disk and SplineRoundDisk robust to transformations (PR #66) I tried to merge your changes but there are some issues (failing tests, type checking, etc.). Do you have your IDE prepared as instructed in https://github.com/damogranlabs/classy_blocks/blob/development/CONTRIBUTING.md? Pre-commit hook should run tests, analyze typing, sort imports, format code, ... Let me know if you need any help setting that up. It seems your code contains a lot of lines like (self.side_1 + r_1 * np.cos(np.pi / 4)) * self.u_1, is there a way to un-repeat those? For instance, you could create a specialized parametric Curve object that would calculate that and re-use it in all round sketches. I can't tell you exactly what to do, I didn't have time to thoroughly check what your code does. I'd gladly help but I've been terribly busy last few weeks and it will stay so for the next few... — Reply to this email directly, view it on GitHub<#66 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AOHXKFK3CKEJZDL567OX2OT2FRRNFAVCNFSM6AAAAABTKZGF7GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNBTGE4DQNBYHE>. You are receiving this because you authored the thread.

[FranzBangar]

Okay, if you're not in a hurry, you can just commit your changes on your local branch and I'll pull them and see what I can do. First I have to finish my W.I.P. though - but it looks promising so I might even get it done! In a while, of course. :)

[LasseHolch]

After a bit of googling it seems like i could just commit as usual and the merge request is updated. At least that is what I have tried. Let me know if there is anything I can do to help the request on the way... Also, can you merge the shape_optimization feature to the main branch. I can add an example where a cylinder decreases in diameter. That should both show illustrate my improvement to the disk, and the usage of shape optimization. I have a bit of a deadline as I will be on parental leave for four months from the 20th. Though I in principle could do a bit of coding, I expect it to be very limited...

…

________________________________ Fra: Nejc Jurkovič ***@***.***> Sendt: 2. januar 2025 20:42 Til: damogranlabs/classy_blocks ***@***.***> Cc: LasseHolch ***@***.***>; Author ***@***.***> Emne: Re: [damogranlabs/classy_blocks] Made Disk and SplineRoundDisk robust to transformations (PR #66) Okay, if you're not in a hurry, you can just commit your changes on your local branch and I'll pull them and see what I can do. First I have to finish my W.I.P. though - but it looks promising so I might even get it done! In a while, of course. :) — Reply to this email directly, view it on GitHub<#66 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AOHXKFPRJ2MX4H4SJKGTQHL2IWJCBAVCNFSM6AAAAABTKZGF7GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNRYGI4DEMZQGA>. You are receiving this because you authored the thread.

[FranzBangar]

I modified your code a bit and merged your PR.
Here's what I did:

• Formatting: there were some single quotes in examples; double quotes are default and should be replaced on save or commit (depends on your IDE). Since your other files are consistent I guess those examples were saved before you set your environment up.
• Formatting: import sorting: the same as above but could be that isort/black/whatever tool gets an update, changes style then ruff whines about it. If that's the case, a simple pip update ... will do.
• You defined an abstract orig_point attribute, then handled its transforms separately. That's some duplicated code and a high risk of errors (some tests were failing already). I put that abstract attribute into __init__ as a Point, then added that to the parts property. Now ElementBase that handles transforms of all parts (of sketch, shape, ...) also transforms that Point. The origo is now a property that simply exposes that Point's position. No more duplicate code!
• There was a problem with DiskBase.core_spline method; you created SplineRound.core_spline with different arguments (type checker shouldn't let you do that though!). I renamed both so that those are now separate methods (which actually are) but I'm not sure I did everything correctly. Please take a look at combined_example, I suspect there's still a problem somewhere.
• I'll create a separate example for a QuarterCylinder with symmetry planes.

Of course you won't have time on your parental leave, I've just been on one and wrote only somewhere around zero lines. But if you - by some weird coincidence - do find yourself programming, I strongly recommend you write tests for your splined shapes. If you cover your code with tests, I'll be able to refactor it with greater confidence I don't mess your code up.

pip install coverage
coverage run -m unittest discover tests
coverage html --skip-covered
<your_browser> htmlcov/index.html (focus on your source files)

I merged your updates into development branch but I don't think I'll be able to push all that into master just yet. The development branch is currently terribly messy with semi-finished features I still have to finish and that is not ready for the master yet. Is this a problem? Can you pip install directly from development branch?

Anyway, thank you for your contribution!