[Paraview] Different Python startup error message

[Eric E. Monson]

Hey John,

I'm glad that helped.

What I've been doing for testing is to save a state file, then load it  
in during a straight Python session, and see if I can animate a time  
series from there. I want to use this eventually to generate batch  
animations using Python along with the Point Sprite representations of  
our data. So far, though, I've only been able to get it to show  
anything on the screen when using the Simple Point Mapper Render Mode.  
The view shows up blank when using a real Point Sprite, and I crashed  
with ShadingLanguage FastSphere. Have you tried anything like this, or  
have you just been using the Python shell within ParaView?

I'll try to figure out where it's looking for the point sprite bitmap  
when rendering from Python, in case that's the problem, but please let  
me know if you have any ideas, too.

Also (sorry that this is a bit off-thread), whenever I try to use  
ShadingLanguage Mapper Render Mode in pv-meshless, I get an error from  
vtkGLSLShader.cxx, line 293 "Shader not compiled!!!". I have the cmake  
flag VTK_USE_GLSL_SHADERS: ON, is there another one I should be  
turning on, too?

Thanks,
-Eric


On Aug 8, 2008, at 4:48 AM, John Biddiscombe wrote:

> Eric
>
> You big Star! - ten points for finding the problem. The  
> ColorChannelArray and ColorMode are two inherited properties from  
> the base mapper and I had renamed (and then back again) them in one  
> (or other) version, but negleged to correct the xml at some point.  
> They were dangling properties with no implementation. This also  
> explains a dodgy colour selection bug that might be fixed when the  
> xml is corrected.
>
> delete
>         <Property name="ColorMode" />
> it is the same as MapScalars,
>
>         <Property name="ColorChannelArray" />
> should be
>         <Property name="ColorArray" />
> but actually can be removed because we don't need to expose it here  
> (none of the other representations do, it is used directly from the  
> mapper property and not via the representation property) (actualkly,  
> most of the others don't need to be exposed either probably, never  
> mind).
>
> You have really saved me a lot of pain. Cheers. My python shell is  
> working now, woo-hoo!
>
> I'll check in stuff to the pv-m repository later today when I've  
> tested a bit more.
>
> JB
>
>
>> Hey John,
>>
>> I need to do some more testing, but I may have gotten around that  
>> 'GetDocumentation' error:
>>
>> First, I deleted the SPHProbePointOverTime from CSCS_SPH_Server.xml  
>> (since I am trying to fix the Point Sprites problem first). Then,  
>> in rendering.xml, <PointSpriteRepresentationProxy  
>> name="PointSpriteRepresentation">, comment out the subproxy exposed  
>> properties lines for ColorMode and ColorChannelArray, so it looks  
>> like this:
>>     ....
>>      <SubProxy>
>>        <!--
>>          Mapper for high-res geometry.
>>        -->
>>        <Proxy name="Mapper"
>>               proxygroup="mappers"
>>               proxyname="PointSpriteMapper" />
>>        <ExposedProperties>
>>          <Property name="LookupTable" />
>>          <!-- <Property name="ColorMode" /> -->
>>          <Property name="MapScalars" />
>>          <Property name="ImmediateModeRendering" />
>>          <Property name="InterpolateScalarsBeforeMapping" />
>>          <Property name="UseLookupTableScalarRange" />
>>          <Property name="ClippingPlanes" />
>>          <Property name="NumberOfSubPieces" />
>>
>>          <!-- <Property name="ColorChannelArray" /> -->
>>          <Property name="AlphaChannelArray" />
>>          <Property name="RadiusChannelArray" />
>>          ....
>>
>> (Maybe the PointSpriteMapper doesn't have these properties to  
>> expose?) This got me through the "from paraview import  
>> servermanager" stage using pvpython. Then, to be able to load a  
>> saved meshless state file I also needed to comment out the line in  
>> servermanager.py which Jean Favre had trouble with (http://www.paraview.org/pipermail/paraview/2008-May/008120.html 
>> )
>>
>>    rvname =  
>> vtkSMRenderViewProxy.GetSuggestedRenderViewType(connection.ID)
>>    if rvname:
>>        # loader.SetRenderViewXMLName(rvname)
>>        pm = ProxyManager()
>>        pm.LoadState(filename, ActiveConnection.ID, loader)
>>
>> After this I was able to load a state file which displayed points  
>> using simple point sprites from Python and do a still frame render.  
>> I'll try more testing tomorrow, but maybe this will allow me to do  
>> batch animation rendering using the point sprites!
>>
>> Talk to you later,
>> -Eric
>>
>>
>> On Jul 30, 2008, at 6:59 AM, John Biddiscombe wrote:
>>
>>> Eric
>>>
>>>> ... and gets to the "representations", crapping out with  
>>>> "AttributeError: 'NoneType' object has no attribute  
>>>> 'GetDocumentation'" when trying to process  
>>>> PointSpriteRepresentation (from rendering.xml).
>>>>
>>>
>>> Thanks veryy much for this information. I spent a bit of time on  
>>> it last week, desperately trying to find out why the point sprite  
>>> representation causes this dodgyness. I failed. I ended up by  
>>> having a single 2 line xml entry with nothing in it at all, and  
>>> this caused the error. I tried cloning other ones and everything -  
>>> but it seems like as soon as I add a new anything to the  
>>> representations this error comes back.
>>>
>>> Can anyone tell me what the error really means?
>>>
>>> I tried adding methods for GetDocumentation, tried adding xml  
>>> entries for doc, removing them, everything, but since I don't  
>>> really know what the rror message is trying to tell me, I am  
>>> clueless - what is the NoneType object referred to, and why is the  
>>> doc not found?
>>>
>>>
>>> JB
>>
>
>
> -- 
> John Biddiscombe,                            email:biddisco @ cscs.ch
> http://www.cscs.ch/about/BJohn.php
> CSCS, Swiss National Supercomputing Centre  | Tel:  +41 (91) 610.82.07
> Via Cantonale, 6928 Manno, Switzerland      | Fax:  +41 (91) 610.82.82
>