Fusion360 and OpenSCAD
Thank you, Torsten, but this "tool" is not documented sufficiently for
me to even understand where to start.
Jon
On 10/6/2016 1:05 PM, Torsten Paul wrote:
On 10/06/2016 01:49 PM, jon wrote:
Or a tool that allows you to position Bezier handles
which are then converted to smoothed lines and then
imported into OpenSCAD
Not integrated, but still nice way to just draw polygon:
source: https://github.com/daid/OpenSCAD-polygon-editor
ciao,
Torsten.
Interesting idea, Richard.
A couple of thoughts....
OpenSCAD would need to monitor the "precompiler" source code, not the
OpenSCAD source file generated by the precompiler, for your idea of
automatically running a script beforehand. Or would the script generate a
temp file to be read by OpenSCAD itself?
It's already common for people to write Python scripts to generate OpenSCAD
source. What's the benefit of yet another super-OpenSCAD language? Or were
you thinking of something like the way PHP embeds chunks of HTML?
This can be done now by using an IDE or something like NotePad++ to manage
the source editing, and on command run the precompiler script to generate
the .scad file which in turn would trigger OpenSCAD to reload it. There's
no need to see the ugly generated code... you can tell OpenSCAD to hide its
editor window.
Frank
On Fri, Oct 7, 2016 at 6:30 AM, jon jon@jonbondy.com wrote:
Thank you, Torsten, but this "tool" is not documented sufficiently for me
to even understand where to start.
Jon
On 10/6/2016 1:05 PM, Torsten Paul wrote:
On 10/06/2016 01:49 PM, jon wrote:
Or a tool that allows you to position Bezier handles
which are then converted to smoothed lines and then
imported into OpenSCAD
Not integrated, but still nice way to just draw polygon:
On 06. okt. 2016 18:58, Torsten Paul wrote:
On 10/06/2016 01:33 PM, arnholm@arnholm.org wrote:
I have been experimenting with extracting 2d profiles from
hand drawings, so they can be used to obtain 3d-printable
models.
See for example http://www.printrbottalk.com/forum/download/file.php?id=11168
Nice, what was used for the image->vector conversion?
The "some experimental code" sounds like it's not using
any existing tool or library?
No, it is a new library + application I have been playing with lately to
do this. It is experimental as noted but solves some difficult examples
already. It does not use any pre-existing tool or library (just
wxWidgets for the application GUI).
Another example, this time a famous logo:
The starting point, a B/W image
https://gyazo.com/5a256235cf21905c5946a4971c50e2ed
The white areas are triangulated into a mesh (home-grown triangulator).
Rather dense triangles in this case:
https://gyazo.com/79195935193fa14ff73e725e0807f4e8
The contours are identified from the mesh, note the colours of the inner
and outer contours, it is automatic
https://gyazo.com/c048ff34f3a1599ddc9234f9775ec3ac
The contours are exported as Angelscript source code and extruded in my
CSG application (using carve for 3d booleans). Below is OpenSCAD used as
a viewer of the resulting polyhedron:
Linear extrude
https://gyazo.com/bae667e63d29c7526d59b99693f7f7af
Rotate extrude (20 degrees)
https://gyazo.com/52a0cef818160c151b9b9b4c86762ef5
It would no doubt be possible to use the exact same approach to export
to .scad format and let OpenSCAD do the extrusions, but I have not done it.
I had a quick look at the potrace algorithm now, since you mentioned it.
http://potrace.sourceforge.net/potrace.pdf
Nice paper, the approach is quite different from mine, but some very
good ideas there.
Originally, I had a similar idea as your trace(...) command of
integrating image->vector conversion into my CSG application, but now I
think it really is impractical because the conversion needs to be
inspected and possibly redone with other parameters before it can be
accepted and used in a 3d model. So I found it easier to handle it
separately as shown above. The result is easy to integrate with other
modeling code since the profiles become source code functions returning
2d shapes (equivalent to an OpenSCAD module), you just copy/paste the
generated functions.
The potrace integration idea is nice but I suspect it might be subject
to difficult user issues if integrated too tightly.
I have briefly used potrace with OpenSCAD in the past
http://arnholm.org/a-reverse-3d-printing-experiment/, but I didn't know
the algorithm used by potrace.
Just my thoughts.
Carsten Arnholm
On 10/06/2016 09:17 PM, Carsten Arnholm wrote:
I had a quick look at the potrace algorithm now, since you
mentioned it. http://potrace.sourceforge.net/potrace.pdf
Nice paper, the approach is quite different from mine, but
some very good ideas there.
Yeah, looks like both approaches give quite good results.
The potrace integration idea is nice but I suspect it might
be subject to difficult user issues if integrated too tightly.
I think it's more a matter of scope. It's not intended to
cover all possible cases giving thousands of parameters. It's
supposed to "just work" for the simple/normal case. For
everything more advanced, I agree that a separate, specialized
application is a better choice.
I do expect it to work well for simple scanned sketches and
could reduce the effort to base a design on that quite a bit.
ciao,
Torsten.
frankv wrote
OpenSCAD would need to monitor the "precompiler" source code, not the
OpenSCAD source file generated by the precompiler, for your idea of
automatically running a script beforehand. Or would the script generate a
temp file to be read by OpenSCAD itself?
Yes.
It's already common for people to write Python scripts to generate
OpenSCAD
source. What's the benefit of yet another super-OpenSCAD language? Or were
you thinking of something like the way PHP embeds chunks of HTML?
In this case, no; this would be a complete compiler, at least initially, for
my own delictation. Although I do have an application already written that
would do exactly that.
This can be done now by using an IDE or something like NotePad++ to manage
the source editing, and on command run the precompiler script to generate
the .scad file which in turn would trigger OpenSCAD to reload it.
And I do elsewhere; save-make-run is a normal part of programming. It would
just be nice not to have to do it.
There's no need to see the ugly generated code... you can tell OpenSCAD
to hide its
editor window.
Doesn't everyone? ;) I was just pointing out how transparent it would be for
very little effort.
tp3 wrote
The other option is to help making OpenSCAD itself better
which will then can benefit everyone who uses it.
Yes, it's more complicated to work together and find
solutions that don't break existing stuff, and yes that
can be frustrating at times, but there's also lots of
things to gain.
Granted, and I might. But there are other gains in working alone on a
private project, such as not having to convince everyone you have a good
idea before you try it out and find it is rubbish.
--
View this message in context: http://forum.openscad.org/Fusion360-and-OpenSCAD-tp18493p18581.html
Sent from the OpenSCAD mailing list archive at Nabble.com.
On 10/06/2016 10:04 PM, Richard Urwin wrote:
tp3 wrote
The other option is to help making OpenSCAD itself better
which will then can benefit everyone who uses it.
Yes, it's more complicated to work together and find
solutions that don't break existing stuff, and yes that
can be frustrating at times, but there's also lots of
things to gain.
Granted, and I might. But there are other gains in working
alone on a private project, such as not having to convince
everyone you have a good idea before you try it out and
find it is rubbish.
Just trying something is actually very simple when using
OpenSCAD as base. Creating a new node that does something
new is quite easy, e.g. I had a python() node that could
generate geometry working in a couple of hours and I mostly
had to fight with the actual python interpreter calls.
The neat thing is that it's possible to reuse all the
existing parser and viewer code without even having to
look at it.
It's only getting a bit more complicated when integrating
new modules as it's hard to change language syntax once
it's out in the wild.
(e.g. https://github.com/openscad/openscad/pull/1380)
But it pretty much all the cases the result of the
discussions is a better solution, even if it takes some
time.
Anyway, I'm just asking because it's a bit sad seeing
all those dead pyscad projects (just an example, but I
feel there are at least 10 different ones with that name)
with description "Make better OpenSCAD and work around
limitations" with maybe 100 commits from some years ago.
Yep, there's lots of things that need improvement, and
at least some of those don't need a separate project.
As said before, it's not wrong to start a project like
that, it's very likely there's lots of things to learn
and that a good thing. But it's also very likely the
project is dead a couple of month later.
ciao,
Torsten.
potrace is an excellent tool for this and you can see a good example of
its integration into inkscape code.
Which is where I use it to generate openscad files direct for export
from vector linework. Of which inkscape is an excellent tool
On 10/7/2016 8:46 AM, Torsten Paul wrote:
On 10/06/2016 09:17 PM, Carsten Arnholm wrote:
I had a quick look at the potrace algorithm now, since you
mentioned it. http://potrace.sourceforge.net/potrace.pdf
Nice paper, the approach is quite different from mine, but
some very good ideas there.
Yeah, looks like both approaches give quite good results.
The potrace integration idea is nice but I suspect it might
be subject to difficult user issues if integrated too tightly.
I think it's more a matter of scope. It's not intended to
cover all possible cases giving thousands of parameters. It's
supposed to "just work" for the simple/normal case. For
everything more advanced, I agree that a separate, specialized
application is a better choice.
I do expect it to work well for simple scanned sketches and
could reduce the effort to base a design on that quite a bit.
ciao,
Torsten.
There's no need to see the ugly generated code... you can tell OpenSCAD
to hide its
editor window.
Doesn't everyone? ;) I was just pointing out how transparent it would be for
very little effort.
Or you allow it to run as a command line in a pipe (which for some of the
other tools is the only way to do it)
I actually go to the trouble of generating nice code with proper
indentation from my tools. In the longer term I've found it pays off
repeatedly when you need to take tool output and debug it, or when you
want to take tool output and hand tweak it for a special case.
It would also not be hard to write an "indent" for openscad 8)
Alan
On Oct 7, 2016, at 01:54, Mark Schafer mschafer@wireframe.biz wrote:
Which is where I use it to generate openscad files direct for export from vector linework. Of which inkscape is an excellent tool
FYI: The OpenSCAD development snapshot has experimental support for direct SVG import. It’s still experimental - we need more people testing this and commenting on how it fits their use-cases.
-Marius
So how would it be if OpenSCAD had an ability to watch the source file and
then run a configurable script on it before loading the result of that
script. There's various levels of doing that, from a script that is blindly
run on all files, to something that can treat files differently based on
their extensions with macros for the filenames involved. A modular first
hack should be simple to produce and then others with more interest could
extend it as required and send pull requests.
You could solve your problem by running a separate tool that watches your
source file, and invokes your language->OpenSCAD compiler when the source
file changes.
There are lots of tools like this available on the internet. Here's one
option: http://entrproject.org/
There's no need to build this tool into OpenSCAD.