[Plugin] RandomPushPull v2.5 20110709

[TIG]

This version is updated and tidied...
Range errors now corrected.
Negative pp values now allowed.
The max/min now works properly with +/-ve values.
The selection now passes to the newly pushpulled faces.
Area comparisons added for 're-selection' [thanks kyyu!']

RandomPushPull.rb

[JGA]

Thanks for sharing!

[derei]

There is a "judgement" error in the plugin code. That is why did not respect max/min.

The original plugin makes push-pull for this distance:

Code: Select all

selection.each{|s|s.pushpull((@ppmax-@ppmin) * rand)}

So, if I input Min(400); Max(600) it will make pushPull for a random value of 200.
In order to make pushPull for a random value between 400 and 600 i think, the code should look like this:

Code: Select all

selection.each{ |s| s.pushpull(@ppmin + (@ppmax-@ppmin) * rand) }

In this way, you set the minimum value (@ppmin) to pushpull, then add the random value of 200 (up to 600).
But, as rand method doesn't provide 0 or 1, there will not be pieces equal with 400 or 600.

[thomthom]

But, as rand method doesn't provide 0 or 1, there will not be pieces equal with 400 or 600.

rand does return 0

[derei]

But, as rand method doesn't provide 0 or 1, there will not be pieces equal with 400 or 600.

rand does return 0

I stand corrected , but surely it does not return 1, so max value of push-pull will be never returned, right?

[thomthom]

Correct

[TIG]

Here's v2.1 viewtopic.php?p=340452#p340452
It now uses the min/max range 'properly'... AND if the random value is actually ==min or ==max then it is used; the accuracy for using the actual max value is set at 1/1000001. On the incredibly rare times it matches exactly or it is exceeded very slightly then 'max' is used, as is 'min' when it's met exactly [rarely].

[derei]

Maybe you could add option in input box for push-pulling on negative values too, or force positive.

EDIT: - i'm not sure (as I'm not a programmer), but this code doesn't pushpull between min and max.

mx can have values much less than @ppmin, so pp can have very easy a negative value (if rn=10, mx = @ppmax * 0.0001; so what are the odds that between max and min to be a difference of 4 zeroes? In consequence pp = mx(very small) - @ppmin(highly probable much larger) ==> negative value for pp ).

Code: Select all

rn=rand(1000001)
     if rn!=0
       mx=@ppmax*(rn/1000000.0)
      if mx > @ppmax
         mx=@ppmax
       end
      else
       mx=0
     end
     pp=mx-@ppmin
     face.pushpull(pp) if pp!=0


[TIG]

Here's v2.2 viewtopic.php?p=340452#p340452
I think I've sorted the range AND allowed negative values too...

[derei]

Here's v2.2 viewtopic.php?p=340452#p340452
I think I've sorted the range AND allowed negative values too...

I'm not sure what you are trying to do there, but wouldn't be easier to force the pp value in v.2.1 to .abs? And...it still doesn't limit between max/min

[TIG]

Here's v2.3 viewtopic.php?p=340452#p340452
It now works properly between the max/min range and allows +/-ve values...
I hope!

[derei]

Well, good news! It works like a charm
Thank you very much!

[TIG]

My brain really hurt with the logic of it... but now it's too obvious

[derei]

My brain really hurt with the logic of it... but now it's too obvious

My brain hurt too trying to "discover" the reason why you tried such complicated ways to do it. Especially because my programming skills are like the result of rand method (yes, between 0 and almost 1 ) and I try to learn something from what you do.

[pilou]

So it was a true random result!

[pilou]

Always funny and robust!

Randompushpull.jpg

re-apply on result with the all faces (top and lateral)
result is not too regular? (I took 1 and .1)

Does it possible to keep the first selection? (top faces)

Randompushpul1.jpg

[TIG]

Unfortunately when you PushPull a face it disappears - the face that is at the new location isn't the same face - that's been deleted by Sketchup...

So... it's not easy to get a reference to the new face...

BUT you could do it... you know the plane of the picked face, you know its pt0=vertices[0].position, you know the pushpull vector, so you can offset the pt0 and then test for the new faces added since the pushpull was done and find the one with the same/reversed normal, parallel plane and coincident vertex at the offset pt0 location...

BUT I'm not going to code it

[pilou]

In this particular case same "top square"
I can use the Selected by aera by Sdmitch

[TIG]

Except there might be faces with the same area that were not the ones you want - so then things like face.normal / plane becomes relevant too - BUT comparing areas IS a good 'outside-of-the-box' test idea

[pilou]

Except there might be faces with the same area that were not the ones you want

As I had started from a sphere and took a minimum measure more large than the big edges on the surface's sphere
that can't be arrived
It was a very particular case

[TIG]

Here's v2.4 viewtopic.php?p=340452#p340452
The selection now passes to the newly pushpulled faces even though the original faces have been deleted by Sketchup...

[pilou]

Seems selection are not keeping all the initial faces selected!
(tryed with some adjacents facets on a thick sphere )

Does exist a Scale multiple ? (for each facets selected)

[TIG]

As I explained, the original face is lost with a pushpull so I need to find the equivalent new face... At the moment it matches area, face normal and face plane [offset by the pushpull distance] - even the slightest variance in the figures that Sketchup calculates will result in a mismatch - so I suspect that's why some faces are being ignored
Any ideas on a better way to spot newly pushpulled faces is invited...

[jorge2011]

you can add a random move tool?

[kyyu]

As I explained, the original face is lost with a pushpull so I need to find the equivalent new face... At the moment it matches area, face normal and face plane [offset by the pushpull distance] - even the slightest variance in the figures that Sketchup calculates will result in a mismatch - so I suspect that's why some faces are being ignored
Any ideas on a better way to spot newly pushpulled faces is invited...

TIG,

Seems to work 100%, if you account for the tiny area difference of the new face. In line #64, I used "if face.area-e[0].abs < 0.000001" instead of "if face.area==e[0]".

Well, at least this worked on a 7 inch diameter sphere. Don't know if it will fail for other scales.

[Addition]
Ok, I had a chance to check if scaling would make the new area test fail. Looks ok so far. My sphere had 288 faces. Initially 7 inches in diameter. I've scaled it up to 70, 700, and 7000 inch diameter. And even with the larger areas, 0.000001 difference still worked to compare if the faces have the same area. The plugin still kept the selection on all 288 faces, in all the tests.

RandomPushPull (v2.4 sel faces passed,my fix).rb

[sonnycds]

Muchas gracias, me viene al pelo

[TIG]

Here's v2.5 viewtopic.php?p=340452#p340452
Kyyu beat me to it with an 'area accuracy check', which now [hopefully] ensures that all of the selected faces that are pushpulled [and thereby get erased] now get their selection passed onto the the newly added face every time...

[pilou]

Works like a charm
Again some requests
- does it possible to keep the selection but separated like when you make a normal PushPpull with CTRL ?
- make a Scale on each face

This last must be asked also for Fredo Scale