ParaView/Python/Tracking an object during animation

Here is a script that I wrote for someone on the mailing list. Add it as a macro from the Python shell for easy access. Essentially, this script sets up tracking for the active object in the active view. Look at the source file for instructions on how to use it. Note that the tracking is pretty basic: the camera moves with the center of the spatial bounds of the object. It would be easy to change the code to do fancier things. If you want to extract a block/cell/subset/whatever, first extract whatever you want tracked using the selection mechanism + extract selection filter and add tracking for that.

<source lang="python">

1. This script demonstrates how to track a moving object.
3. Each time it is executed, it marks the active view
4. and object. It also adds a track to the animation
5. that moves the camera with the center of the bounds
6. of that object.
8. To use, setup the view like you want the camera to
9. look, run this script and then run the animation.
11. To disable tracking, select the server object in the
12. pipeline browser and rerun this script.

from paraview.simple import *

ans = None try:

   ans = GetAnimationScene()

except NameError:

   for i in servermanager.ProxyManager().GetProxiesInGroup("animation").values():
       if i.IsA("vtkSMAnimationSceneProxy"):
           ans = i
           break

if ans:

   # Turn of caching so that we can get the bounds for the
   # data for each time step
   ans.Caching = False
   scene = ans.GetClientSideObject()
   # Remove the previous observer.
   try:
       scene.RemoveObserver(callbackid)
   except NameError:
       pass
   def callback(caller, *args):
       global rv, source, first_bounds, first_focal, first_pos, view_up
       if source:
           source.UpdatePipeline(rv.GetViewUpdateTime())
           bds = source.GetDataInformation().GetBounds()
           # Center of bounds
           bds = map(lambda x,y : (x+y)/2, bds[0:6:2], bds[1:6:2])
           # Difference between initial bounds and current bounds
           bds_diff = map(lambda x,y : x-y, bds, first_bounds)
           # Move the focal point as much as the center of the bounds
           # moved
           rv.CameraFocalPoint = map(lambda x,y: x+y, first_focal, bds_diff)
           # Move the camera as much as the center of the bounds
           # moved
           rv.CameraPosition = map(lambda x,y: x+y, first_pos, bds_diff)
           # Keep the original view up
           rv.CameraViewUp = view_up

   rv = GetActiveView()

   if rv:
       # Add the observer to the VTK scene with priority 1 so that it happens before
       # anything else
       callbackid = scene.AddObserver("AnimationCueTickEvent", callback, 1.0)

       source = GetActiveSource()
       bds = source.GetDataInformation().GetBounds()
       # Center of bounds
       first_bounds = map(lambda x,y : (x+y)/2, bds[0:6:2], bds[1:6:2])
       first_focal = rv.CameraFocalPoint[:]
       first_pos = rv.CameraPosition[:]
       view_up = rv.CameraViewUp[:]

def removeCameraObserver():

   global callbackid, scene
   scene.RemoveObserver(callbackid)

</source>