[Paraview] Point Cloud in Ensight Case Format

[Holger Bauer]

sorry, if this pops up twice. Think I got the wrong sender email ...


I have an Ensight .case file in the following format:

----------------- dummy.case --------------------------------------
FORMAT
type: ensight

GEOMETRY
model: dummy_exp.GEO
measured: case_iterXX-30K.ensdat
VARIABLE
scalar per measured node: HTRAN case_iterXX-30K.ensdat_htran
scalar per measured node: TEMPERATURE case_iterXX-30K.ensdat_temp
----------------------------------------------------------------------

A dummy geometry file (basically containing just a boundary box with 6
cells)
was created as Ensight needs some underlaying geometry (dummy_exp.GEO).
What I need to display are actually scalar values on a point cloud.
So I was advised to put these scalars into three files:

case_iterXX-30K.ensdat contains just the xyz point locations
case_iterXX-30K.ensdat_htran contains first scalar value for the measured
points
case_iterXX-30K.ensdat_temp contains second scalar value for the measured
points

Ensight is able to display color according to the scalar values for the
points.
Paraview successfully reads in the dummy geometry, the points and the
scalar
values, allows me to select the two scalar values (HTRAN/Temperature) in
the
color section of the Display tab but the points are not colorized at all.

So I tried to set interpolation to 'flat' instead of 'Gouraud', changed
representation to point instead of surface and so on. But still no
success.

I had compiled paraview myself using these options:

BUILD_SHARED_LIBS OFF
CMAKE_BACKWARDS_COMPATIBILITY 2.2
CMAKE_BUILD_TYPE
CMAKE_INSTALL_PREFIX /home/xxxx/opt/paraview
VTK_USE_CG_SHADERS OFF
VTK_USE_GLSL_SHADERS ON

I had also tried 'VTK_USE_GLSL_SHADERS OFF'.

Is there anything I am missing?

As I said: Ensight has no trouble setting colors according to the scalars.

By the way the scalar range is read in correctly as can be seen by
displaying
the scalar bar and exploring the range there.

- Holger -

[Berk Geveci]

It looks like you need to glyph the dataset. ParaView does not display
points that are not associated by geometry. The glyph filters creates
geometry (in your cae, you probably want a vertex per point)
associated with points. Apply the glyph filter to your dataset.

-Berk

Post by Holger Bauer
sorry, if this pops up twice. Think I got the wrong sender email ...
----------------- dummy.case --------------------------------------
FORMAT
type: ensight
GEOMETRY
model: dummy_exp.GEO
measured: case_iterXX-30K.ensdat
VARIABLE
scalar per measured node: HTRAN case_iterXX-30K.ensdat_htran
scalar per measured node: TEMPERATURE case_iterXX-30K.ensdat_temp
----------------------------------------------------------------------
A dummy geometry file (basically containing just a boundary box with 6
cells)
was created as Ensight needs some underlaying geometry (dummy_exp.GEO).
What I need to display are actually scalar values on a point cloud.
case_iterXX-30K.ensdat contains just the xyz point locations
case_iterXX-30K.ensdat_htran contains first scalar value for the measured
points
case_iterXX-30K.ensdat_temp contains second scalar value for the measured
points
Ensight is able to display color according to the scalar values for the
points.
Paraview successfully reads in the dummy geometry, the points and the
scalar
values, allows me to select the two scalar values (HTRAN/Temperature) in
the
color section of the Display tab but the points are not colorized at all.
So I tried to set interpolation to 'flat' instead of 'Gouraud', changed
representation to point instead of surface and so on. But still no
success.
BUILD_SHARED_LIBS OFF
CMAKE_BACKWARDS_COMPATIBILITY 2.2
CMAKE_BUILD_TYPE
CMAKE_INSTALL_PREFIX /home/xxxx/opt/paraview
VTK_USE_CG_SHADERS OFF
VTK_USE_GLSL_SHADERS ON
I had also tried 'VTK_USE_GLSL_SHADERS OFF'.
Is there anything I am missing?
As I said: Ensight has no trouble setting colors according to the scalars.
By the way the scalar range is read in correctly as can be seen by
displaying
the scalar bar and exploring the range there.
- Holger -
_______________________________________________
ParaView mailing list
http://www.paraview.org/mailman/listinfo/paraview

[Holger Bauer]

Thanks for your answer. However, I am not really sure whether
I have understood what you are saying.

Yes, Paraview does show my points. But it does not colorize it using
the scalar
(data) values.
Applying the glyph filter adds arrows/cones/etc. while still not showing any
color. However, the arrows get propperly scaled by the scalar values ...
So the question remains why I cannot apply the color as a scalar scale to the
points?

-Holger-

Post by Berk Geveci
It looks like you need to glyph the dataset. ParaView does not display
points that are not associated by geometry. The glyph filters creates
geometry (in your cae, you probably want a vertex per point)
associated with points. Apply the glyph filter to your dataset.
-Berk

[Kent Eschenberg]

Holger,

You've probably already tried this ... go to the glyph property sheet, to
the Display tab, to the "Color by" pulldown menu, and select one of your
scalars. Does that work?

Kent
Pittsburgh Supercomputing Center

Yes, Paraview does show my points. But it does not colorize it using the
scalar (data) values.

[Holger Bauer]

Thanks for all the help. No, in the pull-down selection box of the
color field
there are then only the fields:

Color by:
Cell cellNormals (3)
and Point Normals (3) (+of course the normal uni-color "Property"
selection)
This is what drive me crazy ...

Any other ideas?

I have found a python script that basically converts the xyz data
into:

xyz + scalar .vtk data file. Using this just for the
point location information and setting simply the scalar value to the
z-coordinate (quick and dirty) the .vtk file is read and all the
colors are
displayed perfectly on the points. The file format is quite simple:

Therefore I decided to write a small program that parses the .case
file and
converts all the data into one single VTK file.

The converter (as far as I have understood) just takes all the points
and
defines single-vertex cells (whatever this is). However this can be
opened and
colorized in paraview without any problems.

Any further ideas (I think I still am doing something wrong) are
welcome ...

For those who are interested, this is the format of the .vtk file
that works:

- Holger -

-----------------------------------------------------------------------------
# vtk DataFile Version 2.0
loop
ASCII
DATASET UNSTRUCTURED_GRID
POINTS 1510578 float
1.59800e+02? 1.46900e+02? 1.38800e+02
....
CELLS 1510578 3021156
1 0
1 1
1 2
...
1 1510577
CELL_TYPES 1510578
1
1
...
1
1

POINT_DATA 1510578
SCALARS Z_Value float 1
LOOKUP_TABLE default
1.38800e+02
1.38800e+02
1.38800e+02
... -------------------------------------------------------------------------

After loading your Ensight data, apply the Glyph filter with the

sphere

glyph to create the particle effect in your points and choose your

scalar

field to scale and/or colorize your data. To colorize, go to

"display" tab

in the field "color by" choose your scalar field. I guess it'll

work...

Regards

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://public.kitware.com/pipermail/paraview/attachments/20060113/184e81ca/attachment.htm

[Dominik Szczerba]

did you try cellData2PointData filter?
--dsz

Post by Holger Bauer
Thanks for all the help. No, in the pull-down selection box of the
color field
Cell cellNormals (3)
and Point Normals (3) (+of course the normal uni-color "Property"
selection)
This is what drive me crazy ...
Any other ideas?
I have found a python script that basically converts the xyz data
xyz + scalar .vtk data file. Using this just for the
point location information and setting simply the scalar value to the
z-coordinate (quick and dirty) the .vtk file is read and all the
colors are
Therefore I decided to write a small program that parses the .case
file and
converts all the data into one single VTK file.
The converter (as far as I have understood) just takes all the points
and
defines single-vertex cells (whatever this is). However this can be
opened and
colorized in paraview without any problems.
Any further ideas (I think I still am doing something wrong) are
welcome ...
For those who are interested, this is the format of the .vtk file
- Holger -
-----------------------------------------------------------------------------
# vtk DataFile Version 2.0
loop
ASCII
DATASET UNSTRUCTURED_GRID
POINTS 1510578 float
1.59800e+02 1.46900e+02 1.38800e+02
....
CELLS 1510578 3021156
1 0
1 1
1 2
...
1 1510577
CELL_TYPES 1510578
1
1
...
1
1
POINT_DATA 1510578
SCALARS Z_Value float 1
LOOKUP_TABLE default
1.38800e+02
1.38800e+02
1.38800e+02
...
-------------------------------------------------------------------------

After loading your Ensight data, apply the Glyph filter with the

sphere

glyph to create the particle effect in your points and choose your

scalar

field to scale and/or colorize your data. To colorize, go to

"display" tab

in the field "color by" choose your scalar field. I guess it'll

work...

Regards

------------------------------------------------------------------------
_______________________________________________
ParaView mailing list
http://www.paraview.org/mailman/listinfo/paraview

[Holger Bauer]

O.k. folks,
I'll explore that next week when I am back at my work place.
For all the other comments: Thank you, but of course I have tried all possible
kinds of combinations as well as changing to point representation, changing
sizes of points, etc.

Post by Dominik Szczerba
did you try cellData2PointData filter?
--dsz

I'll try that next week. However, isn't it rather point data -> cell data??

- Holger -

[Holger Bauer]

o.k., I do not know whether sending images to the list are allowed,
so I try: This is what I get ...
It even shows me the "Point TEMPERATURE" value but the points still are white
...
The scalar bar is o.k.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: ddd.jpg
Type: image/jpeg
Size: 23545 bytes
Desc: not available
Url : Loading Image...

[Wylie, Brian]

Yes, and on the glyph filter you might want to change the glyph type
from arrow to sphere or whatever... :)

Brian Wylie - Org 9227
Sandia National Laboratories
MS 0822 - Building 880/A1-J
[phone] FAX[phone]
____ _ __
/ __ \____ _________ | | / (_)__ _ __
/ /_/ / __ `/ ___/ __ `/ | / / / _ \ | /| / /
/ ____/ /_/ / / / /_/ /| |/ / / __/ |/ |/ /
/_/ \__,_/_/ \__,_/ |___/_/\___/|__/|__/

Unleash the Beast

-----Original Message-----
Behalf Of Renato N. Elias
Sent: Friday, January 13, 2006 10:58 AM
Subject: ENC: [Paraview] Re:Point Cloud in Ensight Case Format
After loading your Ensight data, apply the Glyph filter with
the sphere glyph to create the particle effect in your points
and choose your scalar field to scale and/or colorize your
data. To colorize, go to "display" tab in the field "color
by" choose your scalar field. I guess it'll work...
Regards
Renato N. Elias
===============================================
PhD student - http://www.nacad.ufrj.br/~rnelias High
Performance Computing Center Federal University of Rio de
Janeiro Rio de Janeiro, Brazil
[phone]
-----Mensagem original-----
Em nome de Holger Bauer Enviada em: sexta-feira, 13 de
janeiro de 2006 13:40
Assunto: [Paraview] Re:Point Cloud in Ensight Case Format
Thanks for your answer. However, I am not really sure whether
I have understood what you are saying.
Yes, Paraview does show my points. But it does not colorize
it using the scalar
(data) values.
Applying the glyph filter adds arrows/cones/etc. while still
not showing any color. However, the arrows get propperly
scaled by the scalar values ...
So the question remains why I cannot apply the color as a
scalar scale to the points?
-Holger-

Post by Berk Geveci
It looks like you need to glyph the dataset. ParaView does

not display

Post by Berk Geveci
points that are not associated by geometry. The glyph

filters creates

Post by Berk Geveci
geometry (in your cae, you probably want a vertex per point)
associated with points. Apply the glyph filter to your dataset.
-Berk

_______________________________________________
ParaView mailing list
http://www.paraview.org/mailman/listinfo/paraview
_______________________________________________
ParaView mailing list
http://www.paraview.org/mailman/listinfo/paraview

[Renato N. Elias]

It's strange but try to change in the "display style" the representation to
"point" instead of surface. In fact, I guess you don't have any surface to
display...

Cheers

Renato N. Elias
===============================================
PhD student - http://www.nacad.ufrj.br/~rnelias
High Performance Computing Center
Federal University of Rio de Janeiro
Rio de Janeiro, Brazil
[phone]

-----Mensagem original-----
De: paraview-bounces+rnelias=***@paraview.org
[mailto:paraview-bounces+rnelias=***@paraview.org] Em nome de
Holger Bauer
Enviada em: sexta-feira, 13 de janeiro de 2006 15:24
Para: ***@paraview.org
Assunto: Re: RES: [Paraview] Re:Point Cloud in Ensight Case Format

o.k., I do not know whether sending images to the list are allowed, so I
try: This is what I get ...
It even shows me the "Point TEMPERATURE" value but the points still are
white ...
The scalar bar is o.k.

[Renato N. Elias]

I see in your image that you're trying to colorize your points without the
glyph filter. I think It should work, but the appearance will not be so
beautiful that with the glyphs. Furthermore, if your model doesn't have any
surface to display, only points and their coordinates, I'd suggest to change
the display style representation to "points" instead of "surface". At the
same tab, enlarge the point size in order to make the points more visible.

The glyphs would give you a pretty representation of your particles, so, I
suggest you also try to apply the glyph filter in your data.

NOTE: "The glyphs are also more hardware expensive to visualize"

Regards

Renato N. Elias
===============================================
PhD student - http://www.nacad.ufrj.br/~rnelias High Performance Computing
Center Federal University of Rio de Janeiro Rio de Janeiro, Brazil
[phone]

-----Mensagem original-----
De: paraview-bounces+rnelias=***@paraview.org
[mailto:paraview-bounces+rnelias=***@paraview.org] Em nome de
Holger Bauer Enviada em: sexta-feira, 13 de janeiro de 2006 15:24
Para: ***@paraview.org
Assunto: Re: RES: [Paraview] Re:Point Cloud in Ensight Case Format

o.k., I do not know whether sending images to the list are allowed, so I
try: This is what I get ...
It even shows me the "Point TEMPERATURE" value but the points still are
white ...
The scalar bar is o.k.

[Holger Bauer]

update on this:

a) I have sent a small example file set to Kitware for debugging the import.
b) I have extended a small xyz2vtk Konverter tool written in Python which I have
found somehwere on this list to convert the .CASE ensight file(s) to a single
.vtk file not taking any cell data into account. This means only the point
clouds and associated scalars are converted to a UNSTRUCTURED GRID VTK data
file.

The converted .vtk file just reads and displays the values fine (means taking
colors for the scalar values).

All other suggestions (including point to cell data converter within Paraview)
have failed.

Thanks for your help though.

- Holger -

[Dominik Szczerba]

but did you try cellData2PointData?
Dominik

Post by Holger Bauer
a) I have sent a small example file set to Kitware for debugging the import.
b) I have extended a small xyz2vtk Konverter tool written in Python which I have
found somehwere on this list to convert the .CASE ensight file(s) to a single
.vtk file not taking any cell data into account. This means only the point
clouds and associated scalars are converted to a UNSTRUCTURED GRID VTK data
file.
The converted .vtk file just reads and displays the values fine (means taking
colors for the scalar values).
All other suggestions (including point to cell data converter within Paraview)
have failed.
Thanks for your help though.
- Holger -
_______________________________________________
ParaView mailing list
http://www.paraview.org/mailman/listinfo/paraview

[Holger Bauer]

Post by Dominik Szczerba
but did you try cellData2PointData?
Dominik

as this is POINT data there is no way to select cellData2PointData ...

[Dominik Szczerba]

if I am not mistaken you showed a file once with points as cells (ID=1)
In some of my problems this confusion was to blame.
--dsz

Post by Holger Bauer

Post by Dominik Szczerba
but did you try cellData2PointData?
Dominik

as this is POINT data there is no way to select cellData2PointData ...

[Berk Geveci]

OK. Mystery solved. The problem is the fact that there are two parts
in this dataset. One seems to be some sort of structured dataset
whereas the other one contains the particles. The data arrays that you
want to color the particles by does not exist on the structured part.
Because of this, paraview refuses to color by those arrays. I am not
sure if this is a feature or a bug, I have to dig in more when I have
time. For now, you can extract the particles with the "Extract
datasets" filter and color the result.

-Berk

Post by Holger Bauer
a) I have sent a small example file set to Kitware for debugging the import.
b) I have extended a small xyz2vtk Konverter tool written in Python which I have
found somehwere on this list to convert the .CASE ensight file(s) to a single
.vtk file not taking any cell data into account. This means only the point
clouds and associated scalars are converted to a UNSTRUCTURED GRID VTK data
file.
The converted .vtk file just reads and displays the values fine (means taking
colors for the scalar values).
All other suggestions (including point to cell data converter within Paraview)
have failed.
Thanks for your help though.
- Holger -
_______________________________________________
ParaView mailing list
http://www.paraview.org/mailman/listinfo/paraview