Re: [udig-devel] [SoC] uDig GSoC / weekly report

[Craig Taverner <craig@xxxxxxxxxx>]

Hi Davide,

Some suggestions for the Neo4j catalog. Since Neo4j only allows a single process to access each database, but multiple threads can access the database through a common instance of the GraphDatabaseService, you could consider doing the following:

• Make an instance of the catalog entry (GeoResources) for each actual database location on disk. This will appear as an instance in the catalog tree, and could be named with a URL describing the location. For example I simply use file:/home/craig/.amanzi/neo for my applications URL.
• Then underneath that each layer has its own name, and the complete URL could be, for example: file:/home/craig/.amanzi/neo#Layer1
• This allows multiple databases containing multiple layers.
• You can drag the layers onto the map as with normal udig layers.
  

If you want to see sample code, let me know and I can give you access to my SVN server.

Cheers, Craig

On Tue, Jun 22, 2010 at 2:19 AM, Davide <davide.savazzi@xxxxxxxxx> wrote:

1) What did I do last week

- modified neo4j-spatial to support pluggable storage formats for
geometries (implemented WKT storage as an example)
- continued developing the routing Plugin:
added support for using writeable Neo4j instances (registering opened
instances and closing them when the Plugin stops)
created an uDig Operation to create a network / graph from a (Multi)
LineString Neo4j Layer.
- added a Delete Neo4j Layer Operation.

2) What do I plan to do this week
- fix Refresh Neo4j Service Operation (it doesn't update UI)
- fix Delete Operation OutOfMemory problems and add a progress bar
- add to neo4j-spatial an unindex method (needed if user doesn't want
to delete data but just remove them from layer / spatial index)
- continue developing the routing plugin

Regards,
--
Davide Savazzi
_______________________________________________
SoC mailing list
SoC@xxxxxxxxxxxxxxx
http://lists.osgeo.org/mailman/listinfo/soc