[Geotools-devel] DataSources

[<ra...@ei...>]

Hi All. Been cleaning up DataSources for Geotools2.

Currently my version of the code is in an in-between state as I
incorporate 
JTS into the Framework - attached is the javadoc of the code as it 
stands.
Some of the code does not compile, but the javadoc should give 
you an idea
of the general structure.

The only confusing thing is whether to wrap FeatureIndexes around 
a 
FeatureTable, or have Indexes stored as part of that FeatureTable. 
There
are 
ups and downs to both approaches. It should become more clear as
development 
proceeds.

Currently, the pipeline works like this :

A Viewer requests Features within a given Extent - this is sent 
through to
the FeatureTable, which copies it (new method called cloneExtent() 
added
today) and, depending on the load mode, performs a check to see 
if part or
all of the given Extent has already been loaded, if so, the Extent is 
only
partially loaded, using Extent.difference() to find out the extra area.
The 
Extents to be loaded are sent, one at a time, to the DataSource 
referenced
by the Featuretable, which return arrays of Features. The Features 
are 
filtered against the already-loaded Extents to make sure they 
haven't 
already been loaded. This is needed for Features which may be 
along the 
edges or two Extents, or large Features, such as roads or rivers, 
which 
would cut across two or more Extents.
When the load is complete, a TableChangedEvent is fired, which 
notifies
any 
listeners of the update (Viewers, FeatureIndexes, etc.).

That's it. A bit concise, but getting there.

Ray.