Bug #2499
closedPort spatial harvesting script from C++ to java
0%
Description
The current way to represent the metacat EML documents in a spatial dataset relies on saving the information to a temporary text file then calling a compiled C++/Shapelib application to generate a shapefile.
Since we are using Geoserver and the associated Geotools data library, this functionality can be handled entirely in java without the need for compiling binaries or external dependencies.
Updated by Matthew Perry over 17 years ago
The basic functionality is in place but will likely need to be overhauled completely in order to integrate with the indexing process ( bug 2183 ).
Namely there are three missing pieces:
- Supporting multiple geographicCoverages within the same document
- Determine when to handle the bounds as a point vs. a polygon
- update the shapefile based on docid rather than regenerate the entire thing (necessary since being tied to metacat events will mean it may get triggered often so it must be as lean as possible)
Updated by Matthew Perry over 17 years ago
The spatial harvesting classes have been entirely refactored in order to provide the necessary functionality:
- multiple geographic coverages supported
- represents datasets as points(centroids) and polygons(bounding boxes) where appropriate
- Handles inserts, updates and deletes as part of the indexing process so the spatial cache is always synced with metacat db.