preset.de

Lorenz Attractor OTL - a Python OP quick start

Abstract

This walk through explains the usage of the new python operator in Houdini 9. We will create a new SOP that modifies incoming geometry points to form a Lorenz Attractor [1].

Preparation

After starting Houdini 9 save the file to a directory of your choice.
Now create a geometry node with [TAB]-ge..

Press ENTER three times to select the node, lay it down and enter it.

*** Note: maybe the new Interface support another way to do this, but for now it serves just fine.

Delete the fileSOP and create a lineSOP and connect it to a resampleSOP. [TAB] - line ENTER ENTER lays down the line SOP. Shift-RMB on its output and type resa ENTER ENTER to add the resampleSOP.
Your network should look like this:

Resample the line to about 10000 points: disable the "Maximum Segment Length" option and switch to "Maximum Segments":

This is the base for our Lorenz Attractor. Save your file.

The Python Operator

To start we need to create a new Python Operator: File -> New Operator Type ...
Select Python Operator. I named my OP rdg-lorenz_attractor, labeled it as "RDG Lorenz Attraktor" and saved into a new OTL:

The window to edit the operator type properties appears. Create four float values and one float vector 3.
Name them: val_h, val_a, val_b, val_c and startpos. I labeled them as h, a, b, c and Start Position.

Switch to the Channels tab and set the default values for the parameters you just created:

val_h = 0.01, val_a = 10, val_b = 28, val_c= 2.6667 and startpos = (2, 1, 0)

Now press apply so save your operator. Leave the Edit Operator Type Properties open as we need it.

The operator is now installed and you can create an instance: In your network press [TAB] rd ENTER ENTER and lay the new node down. Connect the resampleSOP to the input of the RDG Lorenz Attraktor and set update the display/render flag by clicking on the blue side of the node:

Now it's time to enter the code! Switch to the Code tab and enter this snippet:

# This code is called when instances of this SOP cook.
geo = hou.pwd().geometry()

# Add code to modify the contents of geo.
# read the parameters of the node
 startpos = hou.Node.evalParmTuple(hou.pwd(), "startpos")
 h = hou.Node.evalParm(hou.pwd(), "val_h")
 a = hou.Node.evalParm(hou.pwd(), "val_a")
 b = hou.Node.evalParm(hou.pwd(), "val_b")
 c = hou.Node.evalParm(hou.pwd(), "val_c")
 # init some xyz
 x1 = 0.
 y1 = 0.
 z1 = 0.
 oldp = startpos
 # loop through the point of the SOP
 for pt in geo.points():
    # do the lorenz attraktor
    x1 = oldp[0] + h * a * (oldp[1] - oldp[0])
    y1 = oldp[1] + h * (oldp[0] * (b - oldp[2]) - oldp[1])
    z1 = oldp[2] + h * (oldp[0] * oldp[1] - c * oldp[2])
    newpos = (x1, y1, z1)
    # assign the new point position
    pt.setPosition(newpos)
    # save the position for recursive action
    oldp = newpos

This might cause some discussion. Jason Iversen kindly posted some different approaches to access parameters with python [2]. But I couldn't follow ... Using some print statements and the incomplete documentation of this first public Houdini 9 beta I kind of reverse engineered the access. But maybe this is just my impression.

hou.pwd() is the python version of oppwd() and returns the path to the current opperator,
hou.Node.evalParmTuple(node, parm) reads float3 value as tuple,
hou.Node.evalParm(node, parm) reads the float values as floats.

geo.points() returns all the points in the geometry as a list/tuple,
pt.setPosition(newpos) assigns the new position to the point currently processed.

Now it is time to press Apply again. Free the view to your Scene Viewer and press [SPACE] - a to home on the attractor. It should look like this:

Play with the parameters to modify the attractor and find the valid value ranges. You can change these in the parameters tab of Edit Operator Type Properties.

You can download a finished OTL and sample file via odforce[4]

Conclusion

This was just a quick start into modifying geometry with Python and the new Python Operator of Houdini 9. You can add icons, help like in any other digital asset and create some cool new SOPs. I have chosen the Lorenz Attractor because I really like it [3] and wasn't able to create one with H8, though I realized some time ago that VEX would have been the way.

If you have any suggestions or questions: info@preset.de

2007-07-11 Georg Dümlein

Footnotes

[1] http://local.wasp.uwa.edu.au/~pbourke/fractals/lorenz/index.html
[2] http://www.sidefx.com/index.php?option=com_forum&Itemid=172&page=viewtopic&t=8617
[3] http://www.preset.de/3dsMAX/lorenz/
[4] http://forums.odforce.net/index.php?s=&showtopic=5260&view=findpost&p=37695