Add feature placemarks for raster image output, based on a new parameter (KMPLACEMARK). The purpose of this change is to make the data returned more useful. Normally when a raster image is returned, there is no extra data returned such as the name of the feature. This is normally only returned when KML is returned. But in the example of your viewport zoomed in to a small area of a very large feature, it would normally have to return that whole large feature - which can be slow. With KMSCORE and KMZ output, you can speed this up by returning a raster image instead. But without additional information such as the name of the feature, this is not always useful.
The proposed feature applies when the output format is KMZ with raster images, as determined by the KMSCORE. In addition to returning the raster images, a folder of placemarks will be returned; one folder for each image and one placemark for each feature polygon. Each placemark will display the name of the polygon, placed in the same location as the polygon's label would be. The placemark's description field may contain non-spatial attribute values as per the feature type's description template.
This new optional feature will be turned on by setting the new parameter, KMPLACEMARK, to "true". A value of "false", or absence of the parameter, will turn off the inclusion of placemarks.
- Where are suitable locations for the proposed changes? org.vfny.geoserver.wms.responses.map.kml.KMLRasterTransformer as a start for placemark output? Use the GetMapRequest kvp system to add the new parameter?
- Are there any issues with using something similar to what's done in KMLVectorTransformer? i.e., getting the centroids of the geometries and encoding placemark names/snippets/descriptions/lookats/etc.
hooking into label module/system? any issues?