[Paraview] Filter for interactive spreadsheet?

Tue Jan 17 12:09:08 EST 2017

Hi Cornelis,

I'm confused a bit. With PV 5.2 when I do Extract Selection with preserve
topology on I only get the vtkInsidedness array out from that -- do you
also get your other field arrays? I had to use the Append Attributes filter
to get all of the arrays back into a single filter output.

Also, I don't see an option for "if(1=vtkInsidedness, ...)" in the
Calculator filter. Did you get that to work for you and if you did, what
was the full expression? Using the Python Calculator I can do something
like "numpy.where(vtkInsidedness>0, -RTData, RTData)" to negate values in
the RTData array based on the vtkInsidedness array. Note that I had to use
the Calculator fitler to create a double type of array from the original
signed char vtkInsidedness array. This is because of the error that you
noticed which is that we don't have a way to create a numpy array from a
signed char array and thus the Python Calculator fails.

To answer your question on how to speed up this process, you can try doing
a Python trace and saving that as a Macro button or maybe create a custom
filter that combines all of those filters. Details on both of those should
be in the ParaView Users Guide.

Best,
Andy

On Tue, Jan 17, 2017 at 5:12 AM, Cornelis Bockemühl <
cornelis.bockemuehl at gmail.com> wrote:

> Once again: I just see that I received more information than just the
> assertion error:
>
> ""
>
> Traceback (most recent call last):
>
> File "<string>", line 4, in <module>
>
> File "D:\Users\cbockenm\Documents\BTest\ParaView\OwnBinaries\ParaView
> 5.1.2\lib\site-packages\paraview\calculator.py", line 122, in execute
>
> variables = get_arrays(inputs[0].GetAttributes(self.
> GetArrayAssociation()))
>
> File "D:\Users\cbockenm\Documents\BTest\ParaView\OwnBinaries\ParaView
> 5.1.2\lib\site-packages\paraview\calculator.py", line 34, in get_arrays
>
> arrays[varname] = attribs[key]
>
> File "D:\Users\cbockenm\Documents\BTest\ParaView\OwnBinaries\ParaView
> 5.1.2\lib\site-packages\paraview\vtk\numpy_interface\dataset_adapter.py",
> line 587, in __getitem__
>
> return self.GetArray(idx)
>
> File "D:\Users\cbockenm\Documents\BTest\ParaView\OwnBinaries\ParaView
> 5.1.2\lib\site-packages\paraview\vtk\numpy_interface\dataset_adapter.py",
> line 599, in GetArray
>
> array = vtkDataArrayToVTKArray(vtkarray, self.DataSet)
>
> File "D:\Users\cbockenm\Documents\BTest\ParaView\OwnBinaries\ParaView
> 5.1.2\lib\site-packages\paraview\vtk\numpy_interface\dataset_adapter.py",
> line 138, in vtkDataArrayToVTKArray
>
> narray = numpy_support.vtk_to_numpy(array)
>
> File "D:\Users\cbockenm\Documents\BTest\ParaView\OwnBinaries\ParaView
> 5.1.2\lib\site-packages\vtk\util\numpy_support.py", line 214, in
> vtk_to_numpy
>
> "Unsupported array type %s"%typ
>
> AssertionError: Unsupported array type 15
>
>
> ""
>
> It does not help me a lot, but maybe it indicates to somebody else what
> might be the problem!?
>
> Regards,
> Cornelis
>
>
> 2017-01-17 11:08 GMT+01:00 Cornelis Bockemühl <
> cornelis.bockemuehl at gmail.com>:
>
>> Thanks again Andy!
>>
>> I really appreciate your hints: It has already pointed me into a good
>> direction!
>>
>> Actually what I tried for my purpose was:
>>
>> 1. do a selection, apply the "extract selection" filter, with the "keep
>> topology" option on. With this I am getting just one additional column -
>> "vtkInsidedness" - and this tells me whether in the initial data (actually
>> cells of an "unstructured grid") an item was selected (value 1) or not
>> (value -1).
>>
>> 2. on top of this I applied the "calculator" filter like you suggested,
>> and with just an if(1=vtkInsidedness, ...) I could get exactly what I
>> wanted!
>>
>> Bottom line: these two filters are actually my friends for the purpose!
>>
>> Only two things I am still struggling:
>>
>> 1. if I want this procedure to be a bit more simple I should maybe still
>> write some kind of "wrapper filter" that does the above steps for me just
>> in one. Like derive one filter from the calculator, including the nice
>> calculator panel on the left, and just automatically do the "extract
>> selection" before applying the calculator
>>
>> 2. with the "python calculator" I was not successful, and I wonder why!?
>> Because if I applied that python calculator to my original "unstructured
>> grid" data set everything worked fine (so it is not about missing python
>> support!), but if I did the procedure as above, applying the "python
>> calculator" on the extracted data, I always got a message: "AssertionError:
>> Unsupported array type 15". This was independent on what was the expression
>> I tried to apply, even if it was just a constant value! Looking into the
>> source code it looks indeed like a mistake if some column has "type 15": I
>> found only about 10 type definitions at all...
>>
>> Regards,
>> Cornelis
>>
>> 2017-01-17 10:51 GMT+01:00 Andy Bauer <andy.bauer at kitware.com>:
>>
>>> Hi Cornelis,
>>>
>>> How would you like to select the single or group of values to change?
>>> Also, how would you like to change them (e.g. the same transformation for
>>> each or each with its own transformation)? These two questions are meant to
>>> guide you to do what you want in ParaView.
>>>
>>> You probably want to look at the Python Programmable Filter (
>>> http://www.paraview.org/Wiki/Python_Programmable_Filter) for changing
>>> individual array values. If you are working in serial you may also be able
>>> to bring in the entities to change based on a selection. A better solution
>>> would be to use the global ids filter, then do a selection on what you want
>>> and finally pass that in as a secondary input to the Python Programmable
>>> filter.
>>>
>>> Developing your own filter is also a possibility.
>>>
>>> cheers,
>>> Andy
>>>
>>> On Tue, Jan 17, 2017 at 1:38 AM, Cornelis Bockemühl <
>>> cornelis.bockemuehl at gmail.com> wrote:
>>>
>>>> Dear Andy,
>>>>
>>>> Thanks for this hint: Indeed I was not fully aware of the way how you
>>>> can easily add new calculated columns with these two filters! What they do
>>>> is some kind of "vector math", resulting in a new vector that is appended
>>>> to the input data set.
>>>>
>>>> However my question went in the direction of directly editing single
>>>> values, or groups of values.
>>>>
>>>> Example: Assume that we have a data array "Density" in some data set,
>>>> then I want to be able to directly change the "density" value for just a
>>>> few cells, or even better: for a selected set of data rows.
>>>>
>>>> Because: If such a thing does not exist yet I would consider trying to
>>>> write it myself - but I do not insist on reinventing any wheel! And I am
>>>> sometimes astonished how many "wheels" already exist in the ParaView
>>>> "world"...
>>>>
>>>> Regards,
>>>> Cornelis
>>>>
>>>> PS: Why do I never see my own posts to the paraview mailing list? Is
>>>> this a bug or a feature or something I am doing wrong?
>>>>
>>>> 2017-01-17 0:57 GMT+01:00 Andy Bauer <andy.bauer at kitware.com>:
>>>>
>>>>> This can be done through either the calculator or python calculator
>>>>> filter. If you're in the spreadsheet view it should switch over to the
>>>>> calculator/python calculator output when the filter is applied.
>>>>>
>>>>> On Mon, Jan 16, 2017 at 5:14 PM, Cornelis Bockemühl <
>>>>> cornelis.bockemuehl at gmail.com> wrote:
>>>>>
>>>>>> Has somebody already written and contributed a filter that allows
>>>>>> some simple "spreadsheet like" user manipulation of the data?
>>>>>>
>>>>>> I mean something that accepts some data input, then allows user
>>>>>> changes, and the output would be the changed data set.
>>>>>>
>>>>>> Regards,
>>>>>> Cornelis
>>>>>>
>>>>>> --
>>>>>> Cornelis Bockemühl
>>>>>> Basel, Schweiz
>>>>>>
>>>>>> _______________________________________________
>>>>>> Powered by www.kitware.com
>>>>>>
>>>>>> Visit other Kitware open-source projects at
>>>>>> http://www.kitware.com/opensource/opensource.html
>>>>>>
>>>>>> Please keep messages on-topic and check the ParaView Wiki at:
>>>>>> http://paraview.org/Wiki/ParaView
>>>>>>
>>>>>> Search the list archives at: http://markmail.org/search/?q=ParaView
>>>>>>
>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>> http://public.kitware.com/mailman/listinfo/paraview
>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>> Cornelis Bockemühl
>>>> Basel, Schweiz
>>>>
>>>
>>>
>>
>>
>> --
>> Cornelis Bockemühl
>> Basel, Schweiz
>>
>
>
>
> --
> Cornelis Bockemühl
> Basel, Schweiz
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/paraview/attachments/20170117/f0e7a922/attachment.html>