Mercurial > dive4elements > gnv-client
diff gnv-artifacts/ChangeLog @ 446:f5a041000357
Connected vertical cross section with chart generation.
gnv-artifacts/trunk@494 c6561f87-3c4e-4783-a992-168aeb5c3f6f
author | Sascha L. Teichmann <sascha.teichmann@intevation.de> |
---|---|
date | Fri, 01 Jan 2010 12:08:05 +0000 |
parents | f42ed4f10b79 |
children | 92b7ccbf6163 |
line wrap: on
line diff
--- a/gnv-artifacts/ChangeLog Wed Dec 30 23:02:10 2009 +0000 +++ b/gnv-artifacts/ChangeLog Fri Jan 01 12:08:05 2010 +0000 @@ -1,3 +1,45 @@ +2010-01-01 Sascha L. Teichmann <sascha.teichmann@intevation.de> + + * doc/conf/conf.xml: Deactivated gauss filter because + it blurs no data areas too. Moreover gauss filtering does not + seem to be necessary at all. Possible TODO: Mask the no data + areas while filtering. + + * src/main/java/de/intevation/gnv/state/profile/verticalcrosssection/VerticalCrossSectionOutputState.java: + Added generation of polygon data suitable for PolygonPlot. + + * src/main/java/de/intevation/gnv/state/timeseries/TimeSeriesOutputState.java: + Some hacks to display the new Chart. XXX: This class is + a real mess ... but I've said that before. + + * src/main/java/de/intevation/gnv/chart/Chart.java: Made it + serializable. + + * src/main/java/de/intevation/gnv/chart/VerticalCrossSectionChart.java: + New. Chart for vertical cross sections. Mainly a copy of + Ingo's prototype implementation. TODO: parameter based i18n. + + * src/main/java/de/intevation/gnv/math/Interpolation3D.java: Code + simplification. Made it Serializable to be cachable persistently. + + * src/main/java/de/intevation/gnv/math/AttributedXYColumns.java: + Added fields for the interpolation and the generated PolygonDataset. + Its used as the data carrying object in cache now. + + * src/main/java/de/intevation/gnv/jfreechart/PolygonPlot.java: + Removed println debug code. + + * src/main/java/de/intevation/gnv/jfreechart/PolygonRenderer.java: + Added a little hack to draw the polygons of the interpolation. + The values are in y-direction all below zero (depth is given + by negative values) so the drawing is mirrored along y axis. + This special behavior is determined by the position of data + bounding box. At first I suspected a ccw issue here but some + initial tests falsified this hypothesis. We need to have a + deeper look at this but till than it works. + + * ChangeLog: Fixed some typos. + 2009-12-30 Sascha L. Teichmann <sascha.teichmann@intevation.de> * doc/conf/conf.xml: Fixed defect XML @@ -151,7 +193,7 @@ Iso line generation is a bit more sophisticated than pure region tracing. Along a border of a region there could - be more than one type neighborhood. This is due to quantification + be more than one type of neighborhood. This is due to quantification errors introduced by the fact that steep value gradients are sampled to less points. The only ways out would be an increase of the sample resolution or an other algorithm