GeoTools

Interpolate function as per Symbology Encoding 1.1

Details

  • Type: New Feature New Feature
  • Status: Closed Closed
  • Priority: Major Major
  • Resolution: Fixed
  • Affects Version/s: None
  • Fix Version/s: 2.6.4, 2.7-M0
  • Component/s: None
  • Labels:
    None
  • Testcase included:
    yes

Description

The attached patch is a first go at implementing the Interpolate function for numeric and colour interpolation. Please review it carefully as I expect many changes and improvements will be required. The patch also has a new version of the unit tests for the Categorize function but more work is required for them (and probably that function).

Some design issues:
The SE 1.1 spec defines two optional parameters for this function: method ("numeric" or "color") and mode ("linear", "cosine" or "cubic"). Surely these terms are the wrong way round

The cubic algorithm used was chosen for its simplicity and, hopefully, speed relative to more complex cubic spline algorithms, but I've only assessed its performance briefly and qualitatively.

The cubic algorithm requires at least 4 interpolation points. If less than 4 are provided the function falls back to cosine interpolation - is this reasonable ?

  1. Text File
    se_functions_2.patch
    06/May/10 5:58 AM
    52 kB
    Michael Bedward
  2. Text File
    se_functions_3.patch
    12/May/10 3:28 AM
    53 kB
    Michael Bedward
  3. Text File
    se_functions.patch
    21/Mar/10 8:20 AM
    47 kB
    Michael Bedward

Activity

Ascending order - Click to sort in descending order
Hide
Permalink
Jody Garnett added a comment -
Not too woried about the naming; I can see "interpolation method" being numeric; while the mode defines how the interpolation function is acting?

It is reasonable to downgrade to cosine when are three points are available; or to linear when two points are provided.
Show
Jody Garnett added a comment - Not too woried about the naming; I can see "interpolation method" being numeric; while the mode defines how the interpolation function is acting? It is reasonable to downgrade to cosine when are three points are available; or to linear when two points are provided.
Hide
Permalink
Michael Bedward added a comment -
I still think they've stuffed up the names :-)

Cosine interpolation is just an elaboration of linear interpolation really and only requires two points.

I need to update the patch so that the function:
- checks for tied x values and throws an Exception if found
- adds an outlier point to make cubic interp possible with three points if the lookup value is within range
Show
Michael Bedward added a comment - I still think they've stuffed up the names :-) Cosine interpolation is just an elaboration of linear interpolation really and only requires two points. I need to update the patch so that the function: - checks for tied x values and throws an Exception if found - adds an outlier point to make cubic interp possible with three points if the lookup value is within range
Hide
Permalink
Andrea Aime added a comment -
Michael, I completely agree they swapped the names, but unfortunately that's the published schema, I think we should follow it for the sake of compatibility.
Show
Andrea Aime added a comment - Michael, I completely agree they swapped the names, but unfortunately that's the published schema, I think we should follow it for the sake of compatibility.
Hide
Permalink
Michael Bedward added a comment -
This updated patch has the Interpolate function ready for testing. It also adds InterpolateFunctionExample.java to the demo/example project. Still to do: Categoriize and Recode functions.
Show
Michael Bedward added a comment - This updated patch has the Interpolate function ready for testing. It also adds InterpolateFunctionExample.java to the demo/example project. Still to do: Categoriize and Recode functions.
Hide
Permalink
Andrea Aime added a comment -
Played with it, looks fine, managed to break it in a couple of ways:
- by removing the interpolation parameter in the function definition you get an NPE
- if I use bicubic interpolation I get this error:

{code}
GRAVE: Missing data or value in interpolation points ?
java.lang.IllegalArgumentException: Missing data or value in interpolation points ?
at org.geotools.filter.InterpolateFunction.initialize(InterpolateFunction.java:403)
at org.geotools.filter.InterpolateFunction.evaluate(InterpolateFunction.java:227)
at org.geotools.renderer.style.SLDStyleFactory.evalToColor(SLDStyleFactory.java:1391)
at org.geotools.renderer.style.SLDStyleFactory.getPaint(SLDStyleFactory.java:897)
at org.geotools.renderer.style.SLDStyleFactory.setPolygonStyleFill(SLDStyleFactory.java:393)
at org.geotools.renderer.style.SLDStyleFactory.createPolygonStyle(SLDStyleFactory.java:366)
at org.geotools.renderer.style.SLDStyleFactory.createStyleInternal(SLDStyleFactory.java:320)
at org.geotools.renderer.style.SLDStyleFactory.createStyle(SLDStyleFactory.java:285)
at org.geotools.renderer.lite.StreamingRenderer.processSymbolizers(StreamingRenderer.java:2062)
at org.geotools.renderer.lite.StreamingRenderer.process(StreamingRenderer.java:1988)
at org.geotools.renderer.lite.StreamingRenderer.drawOptimized(StreamingRenderer.java:1882)
at org.geotools.renderer.lite.StreamingRenderer.processStylers(StreamingRenderer.java:1795)
at org.geotools.renderer.lite.StreamingRenderer.paint(StreamingRenderer.java:700)
at org.geotools.swing.RenderingExecutor$Task.call(RenderingExecutor.java:149)
at org.geotools.swing.RenderingExecutor$Task.call(RenderingExecutor.java:1)
at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303)
at java.util.concurrent.FutureTask.run(FutureTask.java:138)
at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
at java.lang.Thread.run(Thread.java:619)
{code}
Show
Andrea Aime added a comment - Played with it, looks fine, managed to break it in a couple of ways: - by removing the interpolation parameter in the function definition you get an NPE - if I use bicubic interpolation I get this error: {code} GRAVE: Missing data or value in interpolation points ? java.lang.IllegalArgumentException: Missing data or value in interpolation points ? at org.geotools.filter.InterpolateFunction.initialize(InterpolateFunction.java:403) at org.geotools.filter.InterpolateFunction.evaluate(InterpolateFunction.java:227) at org.geotools.renderer.style.SLDStyleFactory.evalToColor(SLDStyleFactory.java:1391) at org.geotools.renderer.style.SLDStyleFactory.getPaint(SLDStyleFactory.java:897) at org.geotools.renderer.style.SLDStyleFactory.setPolygonStyleFill(SLDStyleFactory.java:393) at org.geotools.renderer.style.SLDStyleFactory.createPolygonStyle(SLDStyleFactory.java:366) at org.geotools.renderer.style.SLDStyleFactory.createStyleInternal(SLDStyleFactory.java:320) at org.geotools.renderer.style.SLDStyleFactory.createStyle(SLDStyleFactory.java:285) at org.geotools.renderer.lite.StreamingRenderer.processSymbolizers(StreamingRenderer.java:2062) at org.geotools.renderer.lite.StreamingRenderer.process(StreamingRenderer.java:1988) at org.geotools.renderer.lite.StreamingRenderer.drawOptimized(StreamingRenderer.java:1882) at org.geotools.renderer.lite.StreamingRenderer.processStylers(StreamingRenderer.java:1795) at org.geotools.renderer.lite.StreamingRenderer.paint(StreamingRenderer.java:700) at org.geotools.swing.RenderingExecutor$Task.call(RenderingExecutor.java:149) at org.geotools.swing.RenderingExecutor$Task.call(RenderingExecutor.java:1) at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303) at java.util.concurrent.FutureTask.run(FutureTask.java:138) at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908) at java.lang.Thread.run(Thread.java:619) {code}
Hide
Permalink
Michael Bedward added a comment -
Thanks Andrea. I'll do a bit more testing and update the patch. Let me know if you find any other problems in the meantime.
Show
Michael Bedward added a comment - Thanks Andrea. I'll do a bit more testing and update the patch. Let me know if you find any other problems in the meantime.
Hide
Permalink
Michael Bedward added a comment - - edited
The patch se_functions3 fixes the problems that Andrea uncovered. For the first problem, added a check for a mismatch between the color / numeric function parameter and the context arg of the evaluate method. In the event of a mismatch an IllegalArgumentException is thrown with a message that hopefully makes sense to the user. For the second problem, the issue was me not coding the creation of on-the-fly interpolation points properly (to handle cubic interp within the first or last interpolation segment). The demo now runs with cubic / color interp.
Show
Michael Bedward added a comment - - edited The patch se_functions3 fixes the problems that Andrea uncovered. For the first problem, added a check for a mismatch between the color / numeric function parameter and the context arg of the evaluate method. In the event of a mismatch an IllegalArgumentException is thrown with a message that hopefully makes sense to the user. For the second problem, the issue was me not coding the creation of on-the-fly interpolation points properly (to handle cubic interp within the first or last interpolation segment). The demo now runs with cubic / color interp.
Hide
Permalink
Andrea Aime added a comment -
Just tried out the latest patch, works fine for me too. Btw, it does not contain the example code anymore? I grabbed it from the previous patch.
Show
Andrea Aime added a comment - Just tried out the latest patch, works fine for me too. Btw, it does not contain the example code anymore? I grabbed it from the previous patch.
Hide
Permalink
Michael Bedward added a comment -
Ah sorry, forgot that I'd moved the example code into the org.geotools.demo.filter package. Would you like a new patch ?
Show
Michael Bedward added a comment - Ah sorry, forgot that I'd moved the example code into the org.geotools.demo.filter package. Would you like a new patch ?
Hide
Permalink
Andrea Aime added a comment -
Nope, go ahead :-)
Show
Andrea Aime added a comment - Nope, go ahead :-)
Hide
Permalink
Michael Bedward added a comment -
committed to trunk r35471 and 2.6.x branch r35473
Show
Michael Bedward added a comment - committed to trunk r35471 and 2.6.x branch r35473
Hide
Permalink
Andrea Aime added a comment -
Mass closing all issues that have been in "resolved" state for 2 months or more without any feedback or update
Show
Andrea Aime added a comment - Mass closing all issues that have been in "resolved" state for 2 months or more without any feedback or update

People

  • Assignee:
    Andrea Aime
    Reporter:
    Michael Bedward
Vote (0)
Watch (1)

Dates

  • Created:
    Updated:
    Resolved: