VTK › VTK - Users

Parallel coordinates Extraction

Classic

List

Threaded

14 messages Options

Darshan Pai-2

Parallel coordinates Extraction

This example

http://www.vtk.org/Wiki/VTK/Examples/Python/Infovis/ParallelCoordinatesExtraction

I can run it , but can someone please explain this part in the code

rep.SetInputArrayToProcess(0,0,0,0,'RTDataGradient')
rep.SetInputArrayToProcess(1,0,0,0,'RTData')
rep.SetInputArrayToProcess(2,0,0,0,'Elevation')
rep.SetInputArrayToProcess(3,0,0,0,'BrownianVectors')

How is this data acquired ? I never used Infovis before so I am not clear what these lines are doing exactly

Thanks in advance

Darshan

_______________________________________________
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 VTK FAQ at: http://www.vtk.org/Wiki/VTK_FAQ

Follow this link to subscribe/unsubscribe:
http://www.vtk.org/mailman/listinfo/vtkusers

David Doria-2-3

Re: Parallel coordinates Extraction

On Fri, Mar 25, 2011 at 8:12 PM, Darshan Pai <[hidden email]> wrote:

>
> This example
> http://www.vtk.org/Wiki/VTK/Examples/Python/Infovis/ParallelCoordinatesExtraction
> I can run it , but can someone please explain this part in the code
> rep.SetInputArrayToProcess(0,0,0,0,'RTDataGradient')
> rep.SetInputArrayToProcess(1,0,0,0,'RTData')
> rep.SetInputArrayToProcess(2,0,0,0,'Elevation')
> rep.SetInputArrayToProcess(3,0,0,0,'BrownianVectors')
>
> How is this data acquired ? I never used Infovis before so I am not clear what these lines are doing exactly
> Thanks in advance
> Darshan

The SetInputArrayToProcess function is among the most confusing ones I
have encountered (it doesn't have anything to do with vtkInfovis)

http://www.vtk.org/doc/nightly/html/classvtkAlgorithm.html#a6bea16e1329609dbccce0ff8d2367484

I believe all that is happening is the arrays RTDataGradient, RTData,
Elevation, BrownianVectors are being set as vertical axes in the
parallel coordinate representation. The first parameter, 'idx', is
setting the 0th array, the 1st array, etc. I just always put 0's for
the rest of the arguments.

It has always seemed to me that for saneness there should be a

void SetInputArrayToProcess (int idx, const char *name)

so the explanation of "I don't know, just put 0's" doesn't ever have to be used.

David
_______________________________________________
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 VTK FAQ at: http://www.vtk.org/Wiki/VTK_FAQ

Follow this link to subscribe/unsubscribe:
http://www.vtk.org/mailman/listinfo/vtkusers

Darshan Pai-2

Re: Parallel coordinates Extraction

Oh ok , so i guess it somehow picks up the internal arrays from the RTanalyticsSOurce.
Is there anyway to have some of the lines as different colors ? I mean like categorical data.
There is also a Parallel Coordinates in the Charts too .

On Fri, Mar 25, 2011 at 10:38 PM, David Doria <[hidden email]> wrote:

On Fri, Mar 25, 2011 at 8:12 PM, Darshan Pai <[hidden email]> wrote:
>
> This example
> http://www.vtk.org/Wiki/VTK/Examples/Python/Infovis/ParallelCoordinatesExtraction
> I can run it , but can someone please explain this part in the code
> rep.SetInputArrayToProcess(0,0,0,0,'RTDataGradient')
> rep.SetInputArrayToProcess(1,0,0,0,'RTData')
> rep.SetInputArrayToProcess(2,0,0,0,'Elevation')
> rep.SetInputArrayToProcess(3,0,0,0,'BrownianVectors')
>
> How is this data acquired ? I never used Infovis before so I am not clear what these lines are doing exactly
> Thanks in advance
> Darshan

The SetInputArrayToProcess function is among the most confusing ones I
have encountered (it doesn't have anything to do with vtkInfovis)

http://www.vtk.org/doc/nightly/html/classvtkAlgorithm.html#a6bea16e1329609dbccce0ff8d2367484

I believe all that is happening is the arrays RTDataGradient, RTData,
Elevation, BrownianVectors are being set as vertical axes in the
parallel coordinate representation. The first parameter, 'idx', is
setting the 0th array, the 1st array, etc. I just always put 0's for
the rest of the arguments.

It has always seemed to me that for saneness there should be a

void SetInputArrayToProcess (int idx, const char *name)

so the explanation of "I don't know, just put 0's" doesn't ever have to be used.

David

Eric E. Monson

Re: Parallel coordinates Extraction

Hey Darshan,

I don't remember if there's a way to color the lines in the older parallel coordinates view (and don't have time to look right now), but in the new Charts classes you can color plots by another array. This will be in 5.8 and it's been in git master for a while, but it wasn't in 5.6.

(There is a little bug lingering in the parallel coordinates charts plot which causes it to use the plot's pen opacity instead of the opacity of the lookup table you've passed it, but I think Marcus is trying to get around to checking in a fix sometime soon.)

Here's an example that uses categorical colors with the parallel coordinates:

Talk to you later,

-Eric

· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·

Eric E Monson

Duke Visualization Technology Group

# Translated to Python from [VTK]/Charts/Testing/Cxx/TestPCPlot.cxx

import vtk

import math

# Set up a 2D scene, add an XY chart to it

view = vtk.vtkContextView()

view.GetRenderer().SetBackground(1.0, 1.0, 1.0)

view.GetRenderWindow().SetSize(600,300)

chart = vtk.vtkChartParallelCoordinates()

view.GetScene().AddItem(chart)

# Test charting with a few more points...

numPoints = 100

numCats = 8

inc = 7.5 / (numPoints-1)

# Create arrays that will end up as table columns

arrX = vtk.vtkFloatArray()

arrX.SetName("XAxis")

arrX.SetNumberOfComponents(1)

arrX.SetNumberOfTuples(numPoints)

arrC = vtk.vtkFloatArray()

arrC.SetName("Cosine")

arrC.SetNumberOfComponents(1)

arrC.SetNumberOfTuples(numPoints)

arrS = vtk.vtkFloatArray()

arrS.SetName("Sine")

arrS.SetNumberOfComponents(1)

arrS.SetNumberOfTuples(numPoints)

arrCat = vtk.vtkIntArray()

arrCat.SetName("Category_ids")

arrCat.SetNumberOfComponents(1)

arrCat.SetNumberOfTuples(numPoints)

# For some reason table.SetValue() is not wrapped

# so need to build arrays and then add as columns

# Would be more elegant to use numpy, but wanted to keep it out for now...

for i in range(numPoints):

arrX.SetTuple1(i, i * inc)

arrC.SetTuple1(i, math.cos(i * inc) + 0.0)

arrS.SetTuple1(i, math.sin(i * inc) + 0.0)

arrCat.SetTuple1(i, int(float(numCats)*float(i)/float(numPoints)))

# Create a table with some points in it...

table = vtk.vtkTable()

table.AddColumn(arrX)

table.AddColumn(arrC)

table.AddColumn(arrS)

table.AddColumn(arrCat)

cl = []

cl.append([float(cc)/255.0 for cc in [27, 158, 119]]) # Colorbrewer Dark2

cl.append([float(cc)/255.0 for cc in [217, 95, 2]])

cl.append([float(cc)/255.0 for cc in [117, 112, 179]])

cl.append([float(cc)/255.0 for cc in [231, 41, 138]])

cl.append([float(cc)/255.0 for cc in [102, 166, 30]])

cl.append([float(cc)/255.0 for cc in [230, 171, 2]])

cl.append([float(cc)/255.0 for cc in [166, 118, 29]])

cl.append([float(cc)/255.0 for cc in [102, 102, 102]])

lut = vtk.vtkLookupTable()

lutNum = len(cl)

lut.SetNumberOfTableValues(lutNum)

lut.Build()

for ii,cc in enumerate(cl):

lut.SetTableValue(ii,cc[0],cc[1],cc[2],1.0)

lut.SetRange(0,numCats-1)

lut.SetAlpha(1.0)

chart.GetPlot(0).SetInput(table)

chart.GetPlot(0).SetScalarVisibility(1)

chart.GetPlot(0).SetLookupTable(lut)

chart.GetPlot(0).SelectColorArray("Category_ids")

chart.GetPlot(0).GetPen().SetOpacityF(0.8)

view.GetRenderWindow().SetMultiSamples(0)

view.GetRenderWindow().Render()

view.GetInteractor().Start()

On Mar 26, 2011, at 1:15 AM, Darshan Pai wrote:

Oh ok , so i guess it somehow picks up the internal arrays from the RTanalyticsSOurce.
Is there anyway to have some of the lines as different colors ? I mean like categorical data.
There is also a Parallel Coordinates in the Charts too .

On Fri, Mar 25, 2011 at 10:38 PM, David Doria <[hidden email]> wrote:

On Fri, Mar 25, 2011 at 8:12 PM, Darshan Pai <[hidden email]> wrote:
>
> This example
> http://www.vtk.org/Wiki/VTK/Examples/Python/Infovis/ParallelCoordinatesExtraction
> I can run it , but can someone please explain this part in the code
> rep.SetInputArrayToProcess(0,0,0,0,'RTDataGradient')
> rep.SetInputArrayToProcess(1,0,0,0,'RTData')
> rep.SetInputArrayToProcess(2,0,0,0,'Elevation')
> rep.SetInputArrayToProcess(3,0,0,0,'BrownianVectors')
>
> How is this data acquired ? I never used Infovis before so I am not clear what these lines are doing exactly
> Thanks in advance
> Darshan

The SetInputArrayToProcess function is among the most confusing ones I
have encountered (it doesn't have anything to do with vtkInfovis)

http://www.vtk.org/doc/nightly/html/classvtkAlgorithm.html#a6bea16e1329609dbccce0ff8d2367484

I believe all that is happening is the arrays RTDataGradient, RTData,
Elevation, BrownianVectors are being set as vertical axes in the
parallel coordinate representation. The first parameter, 'idx', is
setting the 0th array, the 1st array, etc. I just always put 0's for
the rest of the arguments.

It has always seemed to me that for saneness there should be a

void SetInputArrayToProcess (int idx, const char *name)

so the explanation of "I don't know, just put 0's" doesn't ever have to be used.

David

_______________________________________________
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 VTK FAQ at: http://www.vtk.org/Wiki/VTK_FAQ

Follow this link to subscribe/unsubscribe:
http://www.vtk.org/mailman/listinfo/vtkusers

Eric E. Monson

Re: Parallel coordinates Extraction

Hey Darshan,

How recent is your version of VTK? Marcus checked in a bug fix a week or two ago that may fix this. If you're using a very recent version then maybe you can post your test code and we can take a look.

-Eric

On Mar 29, 2011, at 7:24 PM, Darshan Pai wrote:

Working on it some more I see some more problems .
I have 5 rows ans select the 5th row for color which is the category value 0 and 1 .
I make a lookup table with 2 values having green and blue . I select the range from 0 -1 .
So ideally the plot should show the category value 0 with green and 1 with blue .

Instead I see that some of the lines of category zero are blue while some lines with category 1 are transparent . Also some lines change color in between . I don't understand what is wrong here .

added snapshot

On Mon, Mar 28, 2011 at 2:50 PM, Darshan Pai <[hidden email]> wrote:
Yes that was the problem . Got it working . Thanks

Darshan

On Mon, Mar 28, 2011 at 2:41 PM, Eric E. Monson <[hidden email]> wrote:

Those methods aren't defined in vtkPlot, but in vtkPlotPoints and vtkPlotParallelCoordinates separately. You'll have to cast the GetPlot() to the correct type.

-E

On Mar 28, 2011, at 2:39 PM, Darshan Pai wrote:

.\PC.cpp(77) : error C2039: 'SetScalarVisibility' : is not a member of 'vtkPlot'
1>        C:\vtkcvs\VTK\Charts\vtkPlot.h(42) : see declaration of 'vtkPlot'
1>..\PC.cpp(78) : error C2039: 'SetLookupTable' : is not a member of 'vtkPlot'
1>        C:\vtkcvs\VTK\Charts\vtkPlot.h(42) : see declaration of 'vtkPlot'
1>..\PC.cpp(79) : error C2039: 'SelectColorArray' : is not a member of 'vtkPlot'
1>        C:\vtkcvs\VTK\Charts\vtkPlot.h(42) : see declaration of 'vtkPlot'

Not sure if I am missing an include

On Mon, Mar 28, 2011 at 2:22 PM, Eric E. Monson <[hidden email]> wrote:

That's strange since the Python is just a wrapper around the C++...

-E

On Mar 28, 2011, at 2:15 PM, Darshan Pai wrote:

Thanks I get it .

BTW I cannot convert it for C++ . Is there a way to port it to C++. Esp ScalarVisibility and SelectColorArray is not defined in C++

On Mon, Mar 28, 2011 at 9:44 AM, Eric E. Monson <[hidden email]> wrote:

Hey Darshan,

I don't remember if there's a way to color the lines in the older parallel coordinates view (and don't have time to look right now), but in the new Charts classes you can color plots by another array. This will be in 5.8 and it's been in git master for a while, but it wasn't in 5.6.

(There is a little bug lingering in the parallel coordinates charts plot which causes it to use the plot's pen opacity instead of the opacity of the lookup table you've passed it, but I think Marcus is trying to get around to checking in a fix sometime soon.)

Here's an example that uses categorical colors with the parallel coordinates:

Talk to you later,
-Eric

· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
Eric E Monson
Duke Visualization Technology Group

<ParallelCoordinatesCatsColors.png>

# Translated to Python from [VTK]/Charts/Testing/Cxx/TestPCPlot.cxx

import vtk
import math

# Set up a 2D scene, add an XY chart to it
view = vtk.vtkContextView()

view.GetRenderer().SetBackground(1.0, 1.0, 1.0)
view.GetRenderWindow().SetSize(600,300)

chart = vtk.vtkChartParallelCoordinates()
view.GetScene().AddItem(chart)

# Test charting with a few more points...
numPoints = 100
numCats = 8
inc = 7.5 / (numPoints-1)

# Create arrays that will end up as table columns

arrX = vtk.vtkFloatArray()
arrX.SetName("XAxis")
arrX.SetNumberOfComponents(1)
arrX.SetNumberOfTuples(numPoints)
arrC = vtk.vtkFloatArray()
arrC.SetName("Cosine")

arrC.SetNumberOfComponents(1)
arrC.SetNumberOfTuples(numPoints)
arrS = vtk.vtkFloatArray()
arrS.SetName("Sine")
arrS.SetNumberOfComponents(1)
arrS.SetNumberOfTuples(numPoints)

arrCat = vtk.vtkIntArray()
arrCat.SetName("Category_ids")
arrCat.SetNumberOfComponents(1)
arrCat.SetNumberOfTuples(numPoints)

# For some reason table.SetValue() is not wrapped

# so need to build arrays and then add as columns
# Would be more elegant to use numpy, but wanted to keep it out for now...
for i in range(numPoints):

arrX.SetTuple1(i, i * inc)
arrC.SetTuple1(i, math.cos(i * inc) + 0.0)
arrS.SetTuple1(i, math.sin(i * inc) + 0.0)

arrCat.SetTuple1(i, int(float(numCats)*float(i)/float(numPoints)))

# Create a table with some points in it...
table = vtk.vtkTable()

table.AddColumn(arrX)
table.AddColumn(arrC)
table.AddColumn(arrS)
table.AddColumn(arrCat)

cl = []
cl.append([float(cc)/255.0 for cc in [27, 158, 119]]) # Colorbrewer Dark2

cl.append([float(cc)/255.0 for cc in [217, 95, 2]])
cl.append([float(cc)/255.0 for cc in [117, 112, 179]])
cl.append([float(cc)/255.0 for cc in [231, 41, 138]])
cl.append([float(cc)/255.0 for cc in [102, 166, 30]])

cl.append([float(cc)/255.0 for cc in [230, 171, 2]])
cl.append([float(cc)/255.0 for cc in [166, 118, 29]])
cl.append([float(cc)/255.0 for cc in [102, 102, 102]])

lut = vtk.vtkLookupTable()

lutNum = len(cl)
lut.SetNumberOfTableValues(lutNum)
lut.Build()
for ii,cc in enumerate(cl):
lut.SetTableValue(ii,cc[0],cc[1],cc[2],1.0)

lut.SetRange(0,numCats-1)
lut.SetAlpha(1.0)

chart.GetPlot(0).SetInput(table)
chart.GetPlot(0).SetScalarVisibility(1)
chart.GetPlot(0).SetLookupTable(lut)
chart.GetPlot(0).SelectColorArray("Category_ids")

chart.GetPlot(0).GetPen().SetOpacityF(0.8)

view.GetRenderWindow().SetMultiSamples(0)
view.GetRenderWindow().Render()

view.GetInteractor().Start()

On Mar 26, 2011, at 1:15 AM, Darshan Pai wrote:

Oh ok , so i guess it somehow picks up the internal arrays from the RTanalyticsSOurce.
Is there anyway to have some of the lines as different colors ? I mean like categorical data.
There is also a Parallel Coordinates in the Charts too .

On Fri, Mar 25, 2011 at 10:38 PM, David Doria <[hidden email]> wrote:

On Fri, Mar 25, 2011 at 8:12 PM, Darshan Pai <[hidden email]> wrote:
>
> This example
> http://www.vtk.org/Wiki/VTK/Examples/Python/Infovis/ParallelCoordinatesExtraction
> I can run it , but can someone please explain this part in the code
> rep.SetInputArrayToProcess(0,0,0,0,'RTDataGradient')
> rep.SetInputArrayToProcess(1,0,0,0,'RTData')
> rep.SetInputArrayToProcess(2,0,0,0,'Elevation')
> rep.SetInputArrayToProcess(3,0,0,0,'BrownianVectors')
>
> How is this data acquired ? I never used Infovis before so I am not clear what these lines are doing exactly
> Thanks in advance
> Darshan

The SetInputArrayToProcess function is among the most confusing ones I
have encountered (it doesn't have anything to do with vtkInfovis)

http://www.vtk.org/doc/nightly/html/classvtkAlgorithm.html#a6bea16e1329609dbccce0ff8d2367484

I believe all that is happening is the arrays RTDataGradient, RTData,
Elevation, BrownianVectors are being set as vertical axes in the
parallel coordinate representation. The first parameter, 'idx', is
setting the 0th array, the 1st array, etc. I just always put 0's for
the rest of the arguments.

It has always seemed to me that for saneness there should be a

void SetInputArrayToProcess (int idx, const char *name)

so the explanation of "I don't know, just put 0's" doesn't ever have to be used.

David

_______________________________________________
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 VTK FAQ at: http://www.vtk.org/Wiki/VTK_FAQ

Follow this link to subscribe/unsubscribe:
http://www.vtk.org/mailman/listinfo/vtkusers

<plotoutput.jpg>

Darshan Pai-2

Re: Parallel coordinates Extraction

I think it is the VTK nightly from some date in February . I will update the VTK and see whether the error still persists .

Regards
Darshan

On Tue, Mar 29, 2011 at 7:40 PM, Eric E. Monson <[hidden email]> wrote:

Hey Darshan,

How recent is your version of VTK? Marcus checked in a bug fix a week or two ago that may fix this. If you're using a very recent version then maybe you can post your test code and we can take a look.

-Eric

On Mar 29, 2011, at 7:24 PM, Darshan Pai wrote:

Working on it some more I see some more problems .
I have 5 rows ans select the 5th row for color which is the category value 0 and 1 .
I make a lookup table with 2 values having green and blue . I select the range from 0 -1 .
So ideally the plot should show the category value 0 with green and 1 with blue .

Instead I see that some of the lines of category zero are blue while some lines with category 1 are transparent . Also some lines change color in between . I don't understand what is wrong here .

added snapshot

On Mon, Mar 28, 2011 at 2:50 PM, Darshan Pai <[hidden email]> wrote:

Yes that was the problem . Got it working . Thanks

Darshan

On Mon, Mar 28, 2011 at 2:41 PM, Eric E. Monson <[hidden email]> wrote:

Those methods aren't defined in vtkPlot, but in vtkPlotPoints and vtkPlotParallelCoordinates separately. You'll have to cast the GetPlot() to the correct type.

-E

On Mar 28, 2011, at 2:39 PM, Darshan Pai wrote:

.\PC.cpp(77) : error C2039: 'SetScalarVisibility' : is not a member of 'vtkPlot'
1>        C:\vtkcvs\VTK\Charts\vtkPlot.h(42) : see declaration of 'vtkPlot'
1>..\PC.cpp(78) : error C2039: 'SetLookupTable' : is not a member of 'vtkPlot'
1>        C:\vtkcvs\VTK\Charts\vtkPlot.h(42) : see declaration of 'vtkPlot'
1>..\PC.cpp(79) : error C2039: 'SelectColorArray' : is not a member of 'vtkPlot'
1>        C:\vtkcvs\VTK\Charts\vtkPlot.h(42) : see declaration of 'vtkPlot'

Not sure if I am missing an include

On Mon, Mar 28, 2011 at 2:22 PM, Eric E. Monson <[hidden email]> wrote:

That's strange since the Python is just a wrapper around the C++...

-E

On Mar 28, 2011, at 2:15 PM, Darshan Pai wrote:

Thanks I get it .

BTW I cannot convert it for C++ . Is there a way to port it to C++. Esp ScalarVisibility and SelectColorArray is not defined in C++

On Mon, Mar 28, 2011 at 9:44 AM, Eric E. Monson <[hidden email]> wrote:

Hey Darshan,

I don't remember if there's a way to color the lines in the older parallel coordinates view (and don't have time to look right now), but in the new Charts classes you can color plots by another array. This will be in 5.8 and it's been in git master for a while, but it wasn't in 5.6.

(There is a little bug lingering in the parallel coordinates charts plot which causes it to use the plot's pen opacity instead of the opacity of the lookup table you've passed it, but I think Marcus is trying to get around to checking in a fix sometime soon.)

Here's an example that uses categorical colors with the parallel coordinates:

Talk to you later,
-Eric

· · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
Eric E Monson
Duke Visualization Technology Group

<ParallelCoordinatesCatsColors.png>

# Translated to Python from [VTK]/Charts/Testing/Cxx/TestPCPlot.cxx

import vtk
import math

# Set up a 2D scene, add an XY chart to it
view = vtk.vtkContextView()

view.GetRenderer().SetBackground(1.0, 1.0, 1.0)
view.GetRenderWindow().SetSize(600,300)

chart = vtk.vtkChartParallelCoordinates()
view.GetScene().AddItem(chart)

# Test charting with a few more points...
numPoints = 100
numCats = 8
inc = 7.5 / (numPoints-1)

# Create arrays that will end up as table columns

arrX = vtk.vtkFloatArray()
arrX.SetName("XAxis")
arrX.SetNumberOfComponents(1)
arrX.SetNumberOfTuples(numPoints)
arrC = vtk.vtkFloatArray()
arrC.SetName("Cosine")

arrC.SetNumberOfComponents(1)
arrC.SetNumberOfTuples(numPoints)
arrS = vtk.vtkFloatArray()
arrS.SetName("Sine")
arrS.SetNumberOfComponents(1)
arrS.SetNumberOfTuples(numPoints)

arrCat = vtk.vtkIntArray()
arrCat.SetName("Category_ids")
arrCat.SetNumberOfComponents(1)
arrCat.SetNumberOfTuples(numPoints)

# For some reason table.SetValue() is not wrapped

# so need to build arrays and then add as columns
# Would be more elegant to use numpy, but wanted to keep it out for now...
for i in range(numPoints):

arrX.SetTuple1(i, i * inc)
arrC.SetTuple1(i, math.cos(i * inc) + 0.0)
arrS.SetTuple1(i, math.sin(i * inc) + 0.0)

arrCat.SetTuple1(i, int(float(numCats)*float(i)/float(numPoints)))

# Create a table with some points in it...
table = vtk.vtkTable()

table.AddColumn(arrX)
table.AddColumn(arrC)
table.AddColumn(arrS)
table.AddColumn(arrCat)

cl = []
cl.append([float(cc)/255.0 for cc in [27, 158, 119]]) # Colorbrewer Dark2

cl.append([float(cc)/255.0 for cc in [217, 95, 2]])
cl.append([float(cc)/255.0 for cc in [117, 112, 179]])
cl.append([float(cc)/255.0 for cc in [231, 41, 138]])
cl.append([float(cc)/255.0 for cc in [102, 166, 30]])

cl.append([float(cc)/255.0 for cc in [230, 171, 2]])
cl.append([float(cc)/255.0 for cc in [166, 118, 29]])
cl.append([float(cc)/255.0 for cc in [102, 102, 102]])

lut = vtk.vtkLookupTable()

lutNum = len(cl)
lut.SetNumberOfTableValues(lutNum)
lut.Build()
for ii,cc in enumerate(cl):
lut.SetTableValue(ii,cc[0],cc[1],cc[2],1.0)

lut.SetRange(0,numCats-1)
lut.SetAlpha(1.0)

chart.GetPlot(0).SetInput(table)
chart.GetPlot(0).SetScalarVisibility(1)
chart.GetPlot(0).SetLookupTable(lut)
chart.GetPlot(0).SelectColorArray("Category_ids")

chart.GetPlot(0).GetPen().SetOpacityF(0.8)

view.GetRenderWindow().SetMultiSamples(0)
view.GetRenderWindow().Render()

view.GetInteractor().Start()

On Mar 26, 2011, at 1:15 AM, Darshan Pai wrote:

Oh ok , so i guess it somehow picks up the internal arrays from the RTanalyticsSOurce.
Is there anyway to have some of the lines as different colors ? I mean like categorical data.
There is also a Parallel Coordinates in the Charts too .

On Fri, Mar 25, 2011 at 10:38 PM, David Doria <[hidden email]> wrote:

On Fri, Mar 25, 2011 at 8:12 PM, Darshan Pai <[hidden email]> wrote:
>
> This example
> http://www.vtk.org/Wiki/VTK/Examples/Python/Infovis/ParallelCoordinatesExtraction
> I can run it , but can someone please explain this part in the code
> rep.SetInputArrayToProcess(0,0,0,0,'RTDataGradient')
> rep.SetInputArrayToProcess(1,0,0,0,'RTData')
> rep.SetInputArrayToProcess(2,0,0,0,'Elevation')
> rep.SetInputArrayToProcess(3,0,0,0,'BrownianVectors')
>
> How is this data acquired ? I never used Infovis before so I am not clear what these lines are doing exactly
> Thanks in advance
> Darshan

The SetInputArrayToProcess function is among the most confusing ones I
have encountered (it doesn't have anything to do with vtkInfovis)

http://www.vtk.org/doc/nightly/html/classvtkAlgorithm.html#a6bea16e1329609dbccce0ff8d2367484

I believe all that is happening is the arrays RTDataGradient, RTData,
Elevation, BrownianVectors are being set as vertical axes in the
parallel coordinate representation. The first parameter, 'idx', is
setting the 0th array, the 1st array, etc. I just always put 0's for
the rest of the arguments.

It has always seemed to me that for saneness there should be a

void SetInputArrayToProcess (int idx, const char *name)

so the explanation of "I don't know, just put 0's" doesn't ever have to be used.

David

_______________________________________________
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 VTK FAQ at: http://www.vtk.org/Wiki/VTK_FAQ

Follow this link to subscribe/unsubscribe:
http://www.vtk.org/mailman/listinfo/vtkusers

<plotoutput.jpg>

Darshan Pai-2

Re: Parallel coordinates Extraction

How do I set up the scalar range for each axis ?

On Wed, Mar 30, 2011 at 4:07 PM, Darshan Pai <[hidden email]> wrote:

So i updated to the current nightly build . Now I see that the category colors are correct , but the output is transparent . I changed the opacity of the plot and set it to 1 but still the output lines are very light . Is there any way to change this behaviour . Also can I change the line Size ?

Regards
Darshan

Eric E. Monson

Re: Parallel coordinates Extraction

In reply to this post by Darshan Pai-2

Hey Darshan,

Because there's still a slight bug in the plot class you have to set the line opacity through the plot's Pen rather than through the lookup table. You also set the line width through the Pen (this is Python, so you may have to do some casting in C++):

chart.GetPlot(0).GetPen().SetOpacityF(0.8)
chart.GetPlot(0).GetPen().SetWidth(6)

You set the scalar range for each axis (index i) by first calling chart.GetAxis(i).SetBehavior(1) to set it to a "fixed" range rather than auto-rescaling, and then you call chart.GetAxis(i).SetMaximum() and SetMinimum(). If you want/need to then specify where the ticks are and what their labels are, then you need to call chart.GetAxis(i).SetTickPositions() with a vtkDoubleArray of position values and SetTickLabels() with a vtkStringArray of labels.

-Eric

On Mar 30, 2011, at 4:07 PM, Darshan Pai wrote:

> So i updated to the current nightly build . Now I see that the category colors are correct , but the output is transparent . I changed the opacity of the plot and set it to 1 but still the output lines are very light . Is there any way to change this behaviour . Also can I change the line Size ?
>
> Regards
> Darshan
> <Updated_PC.jpg>

_______________________________________________
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 VTK FAQ at: http://www.vtk.org/Wiki/VTK_FAQ

Follow this link to subscribe/unsubscribe:
http://www.vtk.org/mailman/listinfo/vtkusers

Darshan Pai-2

Re: Parallel coordinates Extraction

thanks eric , works great!!

On Thu, Mar 31, 2011 at 1:38 PM, Eric E. Monson <[hidden email]> wrote:

Hey Darshan,

Because there's still a slight bug in the plot class you have to set the line opacity through the plot's Pen rather than through the lookup table. You also set the line width through the Pen (this is Python, so you may have to do some casting in C++):

chart.GetPlot(0).GetPen().SetOpacityF(0.8)

chart.GetPlot(0).GetPen().SetWidth(6)

You set the scalar range for each axis (index i) by first calling chart.GetAxis(i).SetBehavior(1) to set it to a "fixed" range rather than auto-rescaling, and then you call chart.GetAxis(i).SetMaximum() and SetMinimum(). If you want/need to then specify where the ticks are and what their labels are, then you need to call chart.GetAxis(i).SetTickPositions() with a vtkDoubleArray of position values and SetTickLabels() with a vtkStringArray of labels.

-Eric

On Mar 30, 2011, at 4:07 PM, Darshan Pai wrote:

> So i updated to the current nightly build . Now I see that the category colors are correct , but the output is transparent . I changed the opacity of the plot and set it to 1 but still the output lines are very light . Is there any way to change this behaviour . Also can I change the line Size ?
>
> Regards
> Darshan

> <Updated_PC.jpg>

Darshan Pai-2

Re: Parallel coordinates Extraction

Eric,

Is there an upper limit to the number of axis the Parallel Coordinates defaults too ?
I was just trying for 17 , but i does not allow me anything greater than 10 .

Regards
Darshan

On Thu, Mar 31, 2011 at 2:00 PM, Darshan Pai <[hidden email]> wrote:

thanks eric , works great!!

On Thu, Mar 31, 2011 at 1:38 PM, Eric E. Monson <[hidden email]> wrote:

Hey Darshan,

Because there's still a slight bug in the plot class you have to set the line opacity through the plot's Pen rather than through the lookup table. You also set the line width through the Pen (this is Python, so you may have to do some casting in C++):

chart.GetPlot(0).GetPen().SetOpacityF(0.8)

chart.GetPlot(0).GetPen().SetWidth(6)

You set the scalar range for each axis (index i) by first calling chart.GetAxis(i).SetBehavior(1) to set it to a "fixed" range rather than auto-rescaling, and then you call chart.GetAxis(i).SetMaximum() and SetMinimum(). If you want/need to then specify where the ticks are and what their labels are, then you need to call chart.GetAxis(i).SetTickPositions() with a vtkDoubleArray of position values and SetTickLabels() with a vtkStringArray of labels.

-Eric

On Mar 30, 2011, at 4:07 PM, Darshan Pai wrote:

> So i updated to the current nightly build . Now I see that the category colors are correct , but the output is transparent . I changed the opacity of the plot and set it to 1 but still the output lines are very light . Is there any way to change this behaviour . Also can I change the line Size ?
>
> Regards
> Darshan

> <Updated_PC.jpg>

Eric E. Monson

Re: Parallel coordinates Extraction

Hey Darshan,

I think there is still a default limit of 10 columns, which was probably just so the chart would still end up looking reasonable even if you fed it a table with a thousand columns. All you need to do is loop through your table column names and the call SetColumnVisibility(vtkStdString name, bool visible) for the ones you want to appear. Note that there is also a convenience method SetColumnVisibilityAll(bool visible) in case you want to turn them all on, or want all off so you can just turn on a select few.

-Eric

On Mar 31, 2011, at 6:38 PM, Darshan Pai wrote:

Eric,

Is there an upper limit to the number of axis the Parallel Coordinates defaults too ?
I was just trying for 17 , but i does not allow me anything greater than 10 .

Regards
Darshan

On Thu, Mar 31, 2011 at 2:00 PM, Darshan Pai <[hidden email]> wrote:
thanks eric , works great!!

On Thu, Mar 31, 2011 at 1:38 PM, Eric E. Monson <[hidden email]> wrote:

Hey Darshan,

Because there's still a slight bug in the plot class you have to set the line opacity through the plot's Pen rather than through the lookup table. You also set the line width through the Pen (this is Python, so you may have to do some casting in C++):

chart.GetPlot(0).GetPen().SetOpacityF(0.8)

chart.GetPlot(0).GetPen().SetWidth(6)

You set the scalar range for each axis (index i) by first calling chart.GetAxis(i).SetBehavior(1) to set it to a "fixed" range rather than auto-rescaling, and then you call chart.GetAxis(i).SetMaximum() and SetMinimum(). If you want/need to then specify where the ticks are and what their labels are, then you need to call chart.GetAxis(i).SetTickPositions() with a vtkDoubleArray of position values and SetTickLabels() with a vtkStringArray of labels.

-Eric

On Mar 30, 2011, at 4:07 PM, Darshan Pai wrote:

> So i updated to the current nightly build . Now I see that the category colors are correct , but the output is transparent . I changed the opacity of the plot and set it to 1 but still the output lines are very light . Is there any way to change this behaviour . Also can I change the line Size ?
>
> Regards
> Darshan

> <Updated_PC.jpg>

Darshan Pai-2

Re: Parallel coordinates Extraction

Eric,

I see what you are saying but then I guess there is a bug in the implementation or else I am doing something wrong .
If I hide one of the columns less than 10 and unhide one column greater than 10 it crashes .
Also if I try to set the max and min of columns greater than 10 it crashes too .

It looks as if it only defaults to the first 10 columns.

Regards
Darshan

On Fri, Apr 1, 2011 at 9:43 AM, Eric E. Monson <[hidden email]> wrote:

Hey Darshan,

I think there is still a default limit of 10 columns, which was probably just so the chart would still end up looking reasonable even if you fed it a table with a thousand columns. All you need to do is loop through your table column names and the call SetColumnVisibility(vtkStdString name, bool visible) for the ones you want to appear. Note that there is also a convenience method SetColumnVisibilityAll(bool visible) in case you want to turn them all on, or want all off so you can just turn on a select few.

-Eric

On Mar 31, 2011, at 6:38 PM, Darshan Pai wrote:

Eric,

Is there an upper limit to the number of axis the Parallel Coordinates defaults too ?
I was just trying for 17 , but i does not allow me anything greater than 10 .

Regards
Darshan

On Thu, Mar 31, 2011 at 2:00 PM, Darshan Pai <[hidden email]> wrote:
thanks eric , works great!!

On Thu, Mar 31, 2011 at 1:38 PM, Eric E. Monson <[hidden email]> wrote:

Hey Darshan,

Because there's still a slight bug in the plot class you have to set the line opacity through the plot's Pen rather than through the lookup table. You also set the line width through the Pen (this is Python, so you may have to do some casting in C++):

chart.GetPlot(0).GetPen().SetOpacityF(0.8)

chart.GetPlot(0).GetPen().SetWidth(6)

You set the scalar range for each axis (index i) by first calling chart.GetAxis(i).SetBehavior(1) to set it to a "fixed" range rather than auto-rescaling, and then you call chart.GetAxis(i).SetMaximum() and SetMinimum(). If you want/need to then specify where the ticks are and what their labels are, then you need to call chart.GetAxis(i).SetTickPositions() with a vtkDoubleArray of position values and SetTickLabels() with a vtkStringArray of labels.

-Eric

On Mar 30, 2011, at 4:07 PM, Darshan Pai wrote:

> So i updated to the current nightly build . Now I see that the category colors are correct , but the output is transparent . I changed the opacity of the plot and set it to 1 but still the output lines are very light . Is there any way to change this behaviour . Also can I change the line Size ?
>
> Regards
> Darshan

> <Updated_PC.jpg>

Eric E. Monson

Re: Parallel coordinates Extraction

Try setting all columns invisible first, and then setting the ones you want to visible. Also, as you've seen, don't try setting attributes of axes/columns that aren't visible since it only keeps a list of and only creates vtkAxis objects for visible columns.

The PC chart is a little brittle with regard to these types of operations, but it's been a while since I looked at the code, so I can't remember exactly why.

-Eric

On Apr 1, 2011, at 2:05 PM, Darshan Pai wrote:

Eric,

I see what you are saying but then I guess there is a bug in the implementation or else I am doing something wrong .
If I hide one of the columns less than 10 and unhide one column greater than 10 it crashes .
Also if I try to set the max and min of columns greater than 10 it crashes too .

It looks as if it only defaults to the first 10 columns.

Regards
Darshan

On Fri, Apr 1, 2011 at 9:43 AM, Eric E. Monson <[hidden email]> wrote:

Hey Darshan,

I think there is still a default limit of 10 columns, which was probably just so the chart would still end up looking reasonable even if you fed it a table with a thousand columns. All you need to do is loop through your table column names and the call SetColumnVisibility(vtkStdString name, bool visible) for the ones you want to appear. Note that there is also a convenience method SetColumnVisibilityAll(bool visible) in case you want to turn them all on, or want all off so you can just turn on a select few.

-Eric

On Mar 31, 2011, at 6:38 PM, Darshan Pai wrote:

Eric,

Is there an upper limit to the number of axis the Parallel Coordinates defaults too ?
I was just trying for 17 , but i does not allow me anything greater than 10 .

Regards
Darshan

On Thu, Mar 31, 2011 at 2:00 PM, Darshan Pai <[hidden email]> wrote:
thanks eric , works great!!

On Thu, Mar 31, 2011 at 1:38 PM, Eric E. Monson <[hidden email]> wrote:

Hey Darshan,

Because there's still a slight bug in the plot class you have to set the line opacity through the plot's Pen rather than through the lookup table. You also set the line width through the Pen (this is Python, so you may have to do some casting in C++):

chart.GetPlot(0).GetPen().SetOpacityF(0.8)

chart.GetPlot(0).GetPen().SetWidth(6)

You set the scalar range for each axis (index i) by first calling chart.GetAxis(i).SetBehavior(1) to set it to a "fixed" range rather than auto-rescaling, and then you call chart.GetAxis(i).SetMaximum() and SetMinimum(). If you want/need to then specify where the ticks are and what their labels are, then you need to call chart.GetAxis(i).SetTickPositions() with a vtkDoubleArray of position values and SetTickLabels() with a vtkStringArray of labels.

-Eric

On Mar 30, 2011, at 4:07 PM, Darshan Pai wrote:

> So i updated to the current nightly build . Now I see that the category colors are correct , but the output is transparent . I changed the opacity of the plot and set it to 1 but still the output lines are very light . Is there any way to change this behaviour . Also can I change the line Size ?
>
> Regards
> Darshan

> <Updated_PC.jpg>

Marcus D. Hanwell-2

Re: Parallel coordinates Extraction

Hi,

Commit 733af6b should address the transparency issue, and I added
extra API so that all four color components can be set in one call for
the vtkPen and vtkBrush classes in commit 802784b. If there is
something missing there let me know, but that was also merged to
release for VTK 5.8.

If there is some delicacy around setting the visibility of columns in
the parallel coordinates please file bugs describing them to me. I am
really busy right now, but will certainly try to get to them as soon
as I can.

Marcus
--
Marcus D. Hanwell, Ph.D.
R&D Engineer, Kitware Inc.
(518) 881-4937

On Fri, Apr 1, 2011 at 4:51 PM, Eric E. Monson <[hidden email]> wrote:

> Try setting all columns invisible first, and then setting the ones you want
> to visible. Also, as you've seen, don't try setting attributes of
> axes/columns that aren't visible since it only keeps a list of and only
> creates vtkAxis objects for visible columns.
> The PC chart is a little brittle with regard to these types of operations,
> but it's been a while since I looked at the code, so I can't remember
> exactly why.
> -Eric
>
> On Apr 1, 2011, at 2:05 PM, Darshan Pai wrote:
>
> Eric,
>
> I see what you are saying but then I guess there is a bug in the
> implementation or else I am doing something wrong .
> If I hide one of the columns less than 10 and unhide one column greater
> than 10 it crashes .
> Also if I try to set the max and min of columns greater than 10 it crashes
> too .
>
> It looks as if it only defaults to the first 10 columns.
>
> Regards
> Darshan
>
> On Fri, Apr 1, 2011 at 9:43 AM, Eric E. Monson <[hidden email]> wrote:
>>
>> Hey Darshan,
>> I think there is still a default limit of 10 columns, which was probably
>> just so the chart would still end up looking reasonable even if you fed it a
>> table with a thousand columns. All you need to do is loop through your table
>> column names and the call SetColumnVisibility(vtkStdString name, bool
>> visible) for the ones you want to appear. Note that there is also a
>> convenience method SetColumnVisibilityAll(bool visible) in case you want to
>> turn them all on, or want all off so you can just turn on a select few.
>> -Eric
>>
>> On Mar 31, 2011, at 6:38 PM, Darshan Pai wrote:
>>
>> Eric,
>>
>> Is there an upper limit to the number of axis the Parallel Coordinates
>> defaults too ?
>> I was just trying for 17 , but i does not allow me anything greater than
>> 10 .
>>
>> Regards
>> Darshan
>>
>>
>> On Thu, Mar 31, 2011 at 2:00 PM, Darshan Pai <[hidden email]> wrote:
>>>
>>> thanks eric , works great!!
>>>
>>> On Thu, Mar 31, 2011 at 1:38 PM, Eric E. Monson <[hidden email]>
>>> wrote:
>>>>
>>>> Hey Darshan,
>>>>
>>>> Because there's still a slight bug in the plot class you have to set the
>>>> line opacity through the plot's Pen rather than through the lookup table.
>>>> You also set the line width through the Pen (this is Python, so you may have
>>>> to do some casting in C++):
>>>>
>>>> chart.GetPlot(0).GetPen().SetOpacityF(0.8)
>>>> chart.GetPlot(0).GetPen().SetWidth(6)
>>>>
>>>> You set the scalar range for each axis (index i) by first calling
>>>> chart.GetAxis(i).SetBehavior(1) to set it to a "fixed" range rather than
>>>> auto-rescaling, and then you call chart.GetAxis(i).SetMaximum() and
>>>> SetMinimum(). If you want/need to then specify where the ticks are and what
>>>> their labels are, then you need to call chart.GetAxis(i).SetTickPositions()
>>>> with a vtkDoubleArray of position values and SetTickLabels() with a
>>>> vtkStringArray of labels.
>>>>
>>>> -Eric
>>>>
>>>>
>>>> On Mar 30, 2011, at 4:07 PM, Darshan Pai wrote:
>>>>
>>>> > So i updated to the current nightly build . Now I see that the
>>>> > category colors are correct , but the output is transparent . I changed the
>>>> > opacity of the plot and set it to 1 but still the output lines are very
>>>> > light . Is there any way to change this behaviour . Also can I change the
>>>> > line Size ?
>>>> >
>>>> > Regards
>>>> > Darshan
>>>> > <Updated_PC.jpg>
>>>>
>>>
>>
>>
>
>
>
> _______________________________________________
> 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 VTK FAQ at:
> http://www.vtk.org/Wiki/VTK_FAQ
>
> Follow this link to subscribe/unsubscribe:
> http://www.vtk.org/mailman/listinfo/vtkusers
>
>