Geo: Do not normalize the longitude with value -180 for Lucene shapes #37299
Conversation
iverase
added
>bug
:Analytics/Geo
|
Pinging @elastic/es-analytics-geo |
iverase
changed the title
| @@ -357,7 +357,7 @@ protected static Polygon polygonS4J(GeometryFactory factory, Coordinate[][] poly | |||
| double[] x = new double[shell.length]; | |||
| double[] y = new double[shell.length]; | |||
| for (int i = 0; i < shell.length; ++i) { | |||
| x[i] = normalizeLon(shell[i].x); | |||
| x[i] = Math.abs(shell[i].x) > 180 ? normalizeLon(shell[i].x) : shell[i].x; | |||
There was a problem hiding this comment.
Could you clarify why this is done only here and not in toPolygonLucene() and in LineStringBuilder?
There was a problem hiding this comment.
I am not an expert in JTS, but I guess the implementation understand that -180/+180 are the same point and it creates the right polygon. Lucene implementation, and in particular the tesselator implementation does not understand that and it won't create the right polygon. It will join the points without crossing the dateline.
I don't think it is a bug in Lucene's implementation but a feature and we need to adapt accordingly.
There was a problem hiding this comment.
Discussed it with @iverase and he clarified this. Thanks! I think it would be nice to add a javadoc comment here to explain why this is not needed for holes and in toPolygonLucene() to explain why it is not needed there.
| @@ -357,7 +357,7 @@ protected static Polygon polygonS4J(GeometryFactory factory, Coordinate[][] poly | |||
| double[] x = new double[shell.length]; | |||
| double[] y = new double[shell.length]; | |||
| for (int i = 0; i < shell.length; ++i) { | |||
| x[i] = normalizeLon(shell[i].x); | |||
| x[i] = Math.abs(shell[i].x) > 180 ? normalizeLon(shell[i].x) : shell[i].x; | |||
There was a problem hiding this comment.
Discussed it with @iverase and he clarified this. Thanks! I think it would be nice to add a javadoc comment here to explain why this is not needed for holes and in toPolygonLucene() to explain why it is not needed there.
iverase
changed the title
…#37299) Lucene based shapes should not normalize the longitude value -180 to 180.
…#37299) Lucene based shapes should not normalize the longitude value -180 to 180.
Lucene based shapes should not normalize the longitude value -180 to 180.
Closes #37297