[Paraview] What is the "numpy way"?

[Cory Quammen]

Adam,

Excellent! Thanks for sharing your modifications to make it work with your data.

Best regards,
Cory

On Fri, Jun 19, 2015 at 3:23 PM, Adam Lyon <lyon at fnal.gov> wrote:
> Thanks Cory! That was very helpful. One subtlety - my np.genfromtxt has the
> names argument set, so that returns a structured numpy array. So I had to
> do,
>
> bfieldVtkArray = ns.numpy_to_vtk( d[ ["Bx", "By", "Bz"] ].view( (np.float,3)
> ), 1, vtk.VTK_FLOAT )
> bfieldVtkArray.SetName("BField")
>
> That seems to work.
>
> The one-column arrays are a little easier. E.g.,
> timeVtkArray  = ns.numpy_to_vtk( d["time"].ravel(), 1, vtk.VTK_FLOAT)
> timeVtkArray.SetName("time")
>
> I have to do the .ravel() or else I get the "non-contiguous" error.
>
> Avoiding that old-fashioned loops seems to make it quite a bit faster too.
>
> Thanks again for the help! -- Adam
>
>
>
>
>
> ------
>
> Adam L. Lyon
> Scientist; Associate Division Head for Systems for Scientific Applications
>
> Scientific Computing Division & Muon g-2 Experiment
> Fermi National Accelerator Laboratory
> 630 840 5522 office
> www.fnal.gov
> lyon at fnal.gov
>
> Connect with us!
> Newsletter  |  Facebook  |  Twitter
>
> On Fri, Jun 19, 2015 at 12:36 PM, Cory Quammen <cory.quammen at kitware.com>
> wrote:
>>
>> Hi Adam,
>>
>> There are some utilities in VTK for converting back and forth from VTK
>> arrays to numpy arrays. Here is an example of usage from within
>> pvpython that should work:
>>
>> >>> import numpy as np
>> >>> d = np.genfromtxt('/home/cory/sample.csv', skip_header=1,
>> >>> delimiter=',')
>> >>> d
>> array([[ 1.,  2.,  3.],
>>        [ 4.,  5.,  6.],
>>        [ 7.,  8.,  9.]])
>> >>> import vtk
>> >>> import vtk.util.numpy_support as ns
>> >>> vtk_array = ns.numpy_to_vtk(d, 1, vtk.VTK_FLOAT)
>> >>> vtk_array
>> (vtkFloatArray)0x7f76a9b92410
>> >>> vtk_array.GetNumberOfComponents()
>> 3
>> >>> vtk_array.GetNumberOfTuples()
>> 3L
>> >>> vtk_array.GetTuple(0)
>> (1.0, 2.0, 3.0)
>>
>> vtk_array is a vtkDoubleArray holding the values in the numpy array d.
>> The 1 in the second argument means that the data should be deep copied
>> into the VTK array - this is usually what you want unless you hold a
>> reference to the numpy array d somewhere else in your code.
>>
>> I hope that helps.
>>
>> Best regards,
>> Cory
>>
>> On Fri, Jun 19, 2015 at 12:26 PM, Adam Lyon <lyon at fnal.gov> wrote:
>> > Hi all, I've been debugging a simulation of particle physics experiment
>> > -- I
>> > usually have the program make a csv file of some data to visualize (for
>> > example, x,y,z,time,Bx,By,Bz) where, e.g., Bx is a component of a
>> > magnetic
>> > field. I can read that file straight into ParaView with its text file
>> > reader, but I find it more convenient to run the csv file through a
>> > python
>> > script that makes a vtp file and I load that into ParaView. These files
>> > tend
>> > to be big, and having a compressed binary vtp file means that ParaView
>> > can
>> > read it very quickly (e.g. I parse csv file once with my script instead
>> > of
>> > loading it into ParaView to parse each time I open it).
>> >
>> > The examples in the documentation show how to use the vtk-numpy
>> > interface to
>> > make a vector of points and "dump" them into a vtkPoints object. For
>> > example,
>> >
>> > d = np.genfromtxt(fname, ...)  # more arguments not shown for brevity
>> >
>> > coord = algs.make_vector(d["x"], d["y"], d["z"])
>> > pts = vtk.vtkPoints()
>> > pts.SetData(dsa.numpyTovtkDataArray(coord, "Points"))
>> >
>> > That looks very nice and must be very efficient.
>> >
>> > I don't see an obvious way to do the same thing with point or cell data.
>> > So,
>> > for example, I have...
>> >
>> > numPts = pts.GetNumberOfPoints()
>> > bfield = vtk.vtkFloatArray()
>> > bfield.SetNumberOfComponents(3)
>> > bfield.SetNumberOfTuples(numPts)
>> > bfield.SetName("BField")
>> >
>> > for a in xrange(numPts):
>> >    bfield.InsertTuple(a, [ d["Bx"][a], d["By"][a], d["Bz"][a] ])
>> >
>> > polyData.GetPointData().AddArray(bfield)  # polyData is a vtkPolyData
>> > object
>> >
>> > That loop and explicitly decomposing the "d" array to get at the
>> > magnetic
>> > field components looks "old fashioned" to me.
>> >
>> > What would be the "numpy way" to make that work? Thanks! -- Adam
>> >
>> > ------
>> >
>> > Adam L. Lyon
>> > Scientist; Associate Division Head for Systems for Scientific
>> > Applications
>> >
>> > Scientific Computing Division & Muon g-2 Experiment
>> > Fermi National Accelerator Laboratory
>> > 630 840 5522 office
>> > www.fnal.gov
>> > lyon at fnal.gov
>> >
>> > Connect with us!
>> > Newsletter  |  Facebook  |  Twitter
>> >
>> > _______________________________________________
>> > 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
>> >
>>
>>
>>
>> --
>> Cory Quammen
>> R&D Engineer
>> Kitware, Inc.
>
>



-- 
Cory Quammen
R&D Engineer
Kitware, Inc.