[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [Newsgroup Home]
[news.eclipse.technology.ohf] Re: STEM : Can we visualise at street level?

Amit,
    yes, there is a match between the definition of new graph components in 
stem.internal.data and lat/long data in stem.geography.  Each of the 
properties files in internal.data supports full specification of Dublin Core 
attributes for the data.  One of the DC attributes is "SPATIAL".  In STEM, 
SPATIAL is used to specify a URI for a GML file that contains the lat/long 
data that goes with the data.  The polygons in the GML file are identified 
by the ISO-3166 keys used in the properties file.

Mostly for testing purposes, there is an alternative URI format that is 
recognized for the SPATIAL attribute that allows the lat/long data to be 
specified in the URI itself. eg. "inline:///lat,long,lat,long,lat,long"  See 
org.eclipse.ohf.stem.tests.definitions/src/org.eclipse.ohf.stem.definitions.adapters.spatial.geo.InlineLatLongDataProviderTest.java

All property files can specify the SPATIAL attribute.  This is most 
typically used for the definitions of nodes that represent geographic 
regions, but it can also apply to edges that represent things like roads.

There are several types of predefined data sets that come with STEM.  Some 
such as the basic graph components (nodes, edges,labels) are generated 
directly from the properties files.  The data set called "Models" is 
programmatically generated from the graph components by systematically 
combining them for each country.  Some models for specific countries are 
configured by hand, you'll find these in 
org.eclipse.ohf.stem.internal.data/resources/data/models.  The data set 
called "Scenarios" is configured by hand you'll find the files in 
org.eclipse.ohf.stem.internal.data/resources/data/scenarios.

You are correct there is no model for "India(1,2) Full Infrastructure", 
which is the country of India (IND) with administration levels 1 and 2 (and 
not 0).  The manner in which the built-in full infrastructure models are 
generated always includes the full sequence of levels (i.e, 0, 0,1, 0,1,2, 
0,1,2,3....).  It is quite easy to build and model you wish by using the 
editors and wizards.  Create a project (project wizard) and then a new Model 
(model wizard) and then drag which ever built-in components (models, graphs, 
etc) you like to the new model and drop them.  You can also drag the 
built-in components to your project and drop them and create a copy of the 
built-in component which can be edited with the editors (double click to 
open).  This is good way to explore the structure of any of the built-in 
components and use them as examples of how to create your own.  Note, that 
Drag and drop in the project explorer is not complete, because of a quirk of 
the yet to be completed implementation, you can only drop on the "category" 
in the project and not the file and you can only drag from the file.  This 
will be fixed in some upcoming enhancements to the project explorer.

Connecting nodes requires the definition of edges.  You'll find these files 
under org.eclipse.ohf.stem.internal.data/resources/data/relationship

We're working on additional documentation so you're not missing anything. 
We're stuck between completing functionality and writing about it.

-- 
Daniel Ford
IBM Almaden Research Center
San Jose, CA