Community Questions¶

Developer Questions¶

Q: Where can I find uDig API javadocs

• The Reference section has a Javadocs
• http://udig.refractions.net/docs/api-udig

Q: Where can I find uDig extension Point Documentation

Extension point documentation is included as part of the SDK (or source code checkout)

for direct access during development.

To access click on the show extension point links in the Eclipse MANIFEST.MF and plugin.xml editors.

Development Questions¶

Q: How can I avoid startup dialogs when developing

1. Before launching uDig from your eclipse workspace, modify the run configuration
2. On the Arguments tab, add “-DUDIG_DEVELOPING” to your VM arguments.
3. Now when you clear your workspace and launch uDig, you won’t have to close the tips dialog nor navigate from the intro screen to the workbench.

Q: How to copy an existing Plugin

1. File > Import
2. Select Plugin Development and Plugins and Fragments wizard, and press Next
3. On the right you can see EVERYTHING in the uDig SDK
4. Select a plugin that is similar to what you want to do: net.refractions.udig.catalog.world.image
5. Press Add
6. Press Finish to the plugin(s) into your workspace

From here you can hack away:

• If you keep the plugin with the same name it will be used INSTEAD of the uDIG SDK entry
• Or you can rename the plugin and use it as a good starting point for your own work

Eclipse RCP Questions¶

Q: How can I display my views by default

The views initially displayed are provided by a perspective:

• Extend the current Map Perspective (that is an extension point); or
• Create your own perspective

An example of extending a perspective is in the net.refractions.udig.feature.editor/plugin.xml file.

If you define a new perspective and you want it to be loaded by default you have to create a fragment for net.refractions.udig.ui and in the fragment override the UDIGWorkbenchAdvisor class to return your perspective ID in the getInitialWindowPerspectiveId method.

Examples:

• Creating new Perspective Example
• Extending An Existing Perspective Example

Q: How do I configure keyboard shortcuts

You need to create a schema, which assigns keyboard bindings to commands. This allows a key press to invoke the same command that would otherwise be issued in response to an Action (such as a menu item or toolbar button being pressed).

The actual functionally invoked by a Command can be determined dynamically using a command handler. This is how the same keyboard short cuts (such as Control-C or Delete) can be performed differently depending on the view the user is working in.

Examples: Keyboard Shortcut Example

Q: How do I get an SWT layout to display properly

The size of a control sometimes defaults to zero width and height – this drives us all mad.

The solution lies in the parent composite (rather than the control).

If the composite is using a layout manager to control the size of the controls you can call the layout() method:

Composite parent = new Composite(grandParent, SWT.NONE);
...
(create controls)
...
parent.layout();

You should also review the layout manager configuration associated with each control.

If the composite is not using a layout manager you will need to do things by hand:

• Resize the composite programatically with parent.setSize(...)

Q: How should I use a Progress Monitor

A few tips on using Progress Monitors:

Always start the progress monitor and do at least 1 bit of work. For example:

monitor.beginTask("Working", 4);
monitor.work(1);

Always finish started job.

try{
  monitor.beginTask("Working", 4);
  monitor.work(1);
  // some work
}finally{
  monitor.done();
}

Make use of SubProgressMonitor if sending the monitor to another method:

try{
  monitor.beginTask("Working", 8);
  monitor.work(1);

  SubProgressMonitor sub=new SubProgressMonitor(monitor, 3);
  doSomeWork(sub);
  sub.done();  // don't forget to make sure the sub monitor is done

  sub=new SubProgressMonitor(monitor, 3);
  doSomeMoreWork(sub);
  sub.done();  // don't forget to make sure the sub monitor is done (callee might not use it)

}finally{
  monitor.done();
}

Q: How to wait in the display thread

All interactions with the user (or widgets) is funnelled through a single display thread. As such you should not take a trip out to disk or do any serious computation from the display thread or the application will appear to be frozen and unresponsive.

So how can you perform these tasks - and then update the user interface in response?

• Recommended: Start a background Job; and when it finishes start a Runnable to update the user interface

• Alternative: Use the display’s read and dispatch method to run other jobs that are waiting for the display thread. Only when there are no more jobs waiting(readAndDispatch returns false when no more jobs are waiting) then let the thread sleep.

  while( condition ) {
      //run a display event continue if there is more work todo.
      if ( display.readAndDispatch() ){
          continue;
      }
  
      //no more work to do in display thread, wait on request if request has not
      //finished
      if (condition)
          break ;
  
      Thread.sleep(300);
  }
  

Q: My plugins export but don’t work in uDig

My plugins export but don’t work in uDig. What is going on?

The common suspects are:

• The plugin depends on another plugin that is not part of the uDig your plugin is installed in. If a plugin depends on a missing plugin then that plugin will be deactivated
• Make sure that all the required resources are checked off in th build.properties editor. The build is the important one for running and exporting
• If you made a plugin that contains code as well as other jars then on the runtime tab of the Manifest.MF editor you must make sure that the all the extra jars and a . are in the Classpath list. (The period is intensional as it indicates the code of this plugin).
  • This last point is important only if the build.properties Runtime Information maps . to your source directory. If the mapping is to a jar then make sure that jar is in the classpath of the manifest.
• The exported plugin and the current udig build were compiled with different jdk. -debug -consoleLog as program arguments help to keep track of such errors.

GIS Application Questions¶

Q: How can I find what maps and projects are currently available

The ApplicationGIS class provides access to all the projects currently registered with the uDig instance.

ApplicationGIS.getProjects() will return all the projects currently registered with the system.

Given a project all the contained elements can be searched via the getElements() method. If a single type of elements is desired such as maps the getElements( Class<T> ) method can be used.

To access the first Map in the first project:

ApplicationGIS.openMap(ApplicationGIS.getProjects().get(0).getElements(IMap.class).get(0));

Q: How can I programmatically commit changes

There is a Commit Command that can be sent to the map to commit:

MapCommand commitCommand=EditCommandFactory.getInstance().createCommitCommand();
map.sendCommandSync(commitCommand);

Q: How can I set up a custom splash screen

Follow the Custom Application Tutorial for branding and splash screen information.

Alternative: Make a fragment which overrides the provided intro/root.xml or the introContent.xml in the net.refractions.udig.ui plugin.

Q: How do I add a layer programmatically

The ApplicationGIS class has a utility method for you:

ApplicationGIS.addLayersToMap(map, layers, startposition, project);

You can also send the command directly yourself:

IMap map;
map.sendCommandASync(new AddLayerCommand(layer) );

Q: How do I add default key bindings to my IWorkbenchPart

Simply add this line to the createPartControl() method of your workbench part.

ApplicationGIS.getToolManager().registerActionsWithPart(this);

This will allow your view to forward common keybindings to the MapEditor so the user can

change controls and pan around the Map using the keyboard.

Q: How do I create a map programmatically

There is a command to create a map, which you can send to the project:

CreateMapCommand command=new CreateMapCommand("MapName", listofGeoResources, project);
project.sendSync(command);
IMap map=command.getCreatedMap();

Q: How do I make a new type of layer

A layer in uDig is (very generally) a set of IGeoResources which are handles for the same dataset.

Example: A WFS and a WMS may be backed onto the same dataset, so the Layer should have access to a WFSGeoResource and a WMSGeoResource

To define a new type of layer one must create a new IGeoResource (and IService since IServices

contain the IGeoResources).

Example: A Web Terrain layer would require a Web Terrain IService and IGeoResource.

Once a new layer type is created it does not automatically get rendered unless one of its

IGeoResources resolves to a DataStore, GridCoverage or a WebMapServer. If one of the IGeoResources does resolve to one of those objects then you are done and the layer will render. If not then a new Renderer must be developed.

For more information:

• Catalog
• Renderers

Q: How is udig rendering different from geotools rendering

Our rendering challenge is a little bit different then the one that the geotools Streaming Renderer

focuses on.

Often in a OGC Open Web Services workflow there is more then one way to do it. Both a Web Feature

Server and a Web Map Server may provide the same information, or indeed be able to provide a picture for a layer in a map. To make matters interesting a Web Map Server may be able to draw more then one layer at the same time. Our api is trying to walk the line between capturing this complexity, and hiding it.

When we have figured out which workflow is going to be used for a layer we end up with

the concept of a Renderer.

Each Layer, or a set of Layers, get a Renderer. Each Renderer produced a raster. Renderers are

aranged into a stack, and enough event notification is used to so that everyone can play in their own thread.

Q: How the do Features and DataStores fit into the picture

At the uDig level there are IServices and IGeoResources. These are generic handles to something

“real.”

Example: An IService can be a handle to a WMS or a Datastore. An IGeoResource can be a handle to a “FeatureSource” or “GridCoverage”.

A Layer references an IGeoResource... Actually, since two IGeoResources can refer to the same data (for example a WMS Layer and a WFS FeatureCollection backed onto the same data) a layer can reference one or more IGeoResources BUT only 1 data. A layer has functionality allowing inspection of the georesources and also map specific information that is unrelated to the IGeoResource, for example a map name and a style.

Renderer is responsible for drawing a layer. There are many different types of renderers. Some renderers can render Vector data, others can render GridCoverages or make WMS requests. The BasicFeatureRenderer, for example, only works for Layers that has an IGeoResource that is a handle for a FeatureSource.

That is all at the uDig level. uDig uses the Geotools library extensively so most of the current IServices and IGeoResources are handles for Geotools objects. Geotools has DataStores and FeatureSources for Vector data. A DataStore is a peer of IService. An example of a Datastore is a PostGIS database. FeatureSources can be obtained from DataStores and are a peer of IGeoResource. FeatureSources can be used to obtain features of a particular feature type from a DataStore. FeatureSources are read-only. If the Datastore is read-write you can do an instance check on the FeatureSource to see if it is a FeatureStore. FeatureStores provide methods for adding/removing/modifying features.

The next obvious question is: If I am in uDig how do I get a FeatureSource? Here are some common scenarios that occur in uDig.

FeatureSource access from a layer:

layer.getResource( FeatureSource.class, monitor );

This is a blocking call you can do a non-blocking check to see if the layer has a FeatureSource by:

layer.hasResource( FeatureSource.class );

FeatureSource access from an IGeoResource:

if( resource.canResolve( FeatureSource.class ) )
    return resource.resolve( FeatureSource.class );

If it is possible, always obtain a FeatureSource from a layer. This is because uDig’s EditManager manages transactions for the user (and developer). FeatureSources by default use auto commit transactions where the FeatureSource obtained from a layer uses a transaction.

Q: How to get geometries from layer

To get geometries from a layer you need to ask for a FeatureSource object from the layer:

FeatureSource source=layer.getResource( FeatureSource.class, progressMonitor);

Once you have a feature source you can get all the features from the source by:

FeatureCollection collection=source.getFeatures();

Q: How to tell when something changes

The uDig “application model” is maintained with EMF (the eclipse modeling framework). As such it is very easy to morph uDig to meet your needs.

• Adaptor: Generic EMF notifications are provided through use of an “Adaptor”

  layer.eAdapters().add( new AdapterImpl(){
     public void notifyChanged( Notificaiton msg ) {
         if( msg.getNotifier() instnaceof Layer ){
             Layer layer = (Layer) msg.getNotifier();
             switch( msg.getFeatureID(Layer.class) ) {
             case ProjectPackage.LAYER__NAME:
                  System.out.println( layer.getName() +" renamed");
                  break;
             case ProjectPackage.LAYER__GEO_RESOURCES:
                  System.out.println( "We have new data!");
                  FeatureType schema = layer.getSchema();
                  if( schema != null ){
                       System.out.println( "changed to "+schema.getTypeName() );
                  }
                  break;
             }
         }
     }
  });
  

• Listener: the project sets add/remove listener methods as they are requested on the udig-devel email list:

  layer.addListener( new ILayerListener(){
     public refresh( LayerEvent event ){
         if( event.getType() == LayerEvent.EventType.STYLE ){
             ILayer layer = event.getSource();
             IStyleBlackboard style = (IStyleBlackboard ) event.getNewValue();
  
             // layer has changed style
         }
     }
  });
  

An adapter is a traditional pater when you want to use one data model and “morph” it to fit another interface. One of the side effects of this is you need to pay attention to the origional data, and pass any changes along.

You can see lots of examples of this idea in Java code. People setting up custom JTreeModels to visualize an internal data structure etc...

Since this need happens all the time the EMF crew decided design for it in mind. It is a much more difficult, and interesting, problem them simply listening for changes (Indeed it is a superset of change notification - basically change notification with interface change). The benifit is that you can “force” EMF models (and thus uDig) into about anything.

So the above example is “an adapter”, and we are only paying attention to the changes.

Example: Here is how you can watch the “Viewport Model” (ie. Zoom, Pan, CRS):

map.getViewportModel().addViewportModelListener(new IViewportModelListener()){
  public void changed(ViewportModel event){
    if( event.getType()==EventType.CRS }
      // crs has changed do something
    else if( event.getType()==EventType.BOUNDS ){
      // bounds have changed do something else
    }
  }
}

Example: Layer(s) added/removed

• IMapCompositionListener* tells you when layers were added to the map, were deleted from the map or when the layer list order has changed. For more information take a look at MapCompositionEvent.EventType:

  ApplicationGIS.getActiveMap().addMapCompositionListener(new IMapCompositionListener(){
      public void changed( MapCompositionEvent event ) {
          if (event.getType() == EventType.ADDED) {
              System.out.println("Layer added");
          } else if (event.getType() == EventType.REMOVED) {
              System.out.println("Layer removed");
          }
          //..
      }
  });
  

Example: Map opens/created/closes If you have to do something when a map opens, was created or is about to close, use the extension point net.refractions.udig.project.mapInterceptor in your plugin.xml:

[..]
  <extension
        point="net.refractions.udig.project.mapInterceptor">
        <mapOpening
              class="net.refractions.udig.yourplugin.OpenMapListener"
              id="net.refractions.udig.yourplugin.mapOpens">
        </mapOpening>
  </extension>
[..]

And create a new class:

import net.refractions.udig.project.interceptor.MapInterceptor;
import net.refractions.udig.project.internal.Map;

public class OpenMapListener implements MapInterceptor{

    public void run(Map map){
        System.out.println("map opens: " + map.getName());
    }
}

Example: MapEditor and other ViewParts get activated/opened/closed. If you just want to track changes of the MapEditor, take a look at LayersView.MapEditorListener.

getSite().getWorkbenchWindow().getPartService().addPartListener(new  IPartListener() {

    public void partActivated(IWorkbenchPart part) {
        System.out.println("partActivated: " + part.getTitle());
    }

    public void partBroughtToTop(IWorkbenchPart part) {
        System.out.println("partBroughtToTop: " + part.getTitle());
    }

    public void partClosed(IWorkbenchPart part) {
        System.out.println("partClosed: " + part.getTitle());
    }

    public void partDeactivated(IWorkbenchPart part) {
        System.out.println("partDeactivated: " + part.getTitle());
    }

    public void partOpened(IWorkbenchPart part) {
        System.out.println("partOpened: " + part.getTitle());
    }
});

Q: Providing Visual Feedback from a View

You will need to know which Map is being worked on; either by:

• looking up the current active part and asking it to “adapt” to a Map.
• having a workbench listener and watching the editors change; and remembering the Map from the last MapEditor used.

IMap map;

IMap getMap(){
    return map;
    // you will need to figure this out by listening to the workbench IPartListener
    // see LayerView for an example
}

• Creating and Executing a DrawShapeCommand: This technique is good for providing quick visual feedback on the screen; the draw command stays on the screen until you set it to be invalid. An example of this being used is to “flash” a feature when it is selected in the view. Most of the examples you see for using a DrawShapeCommand will be for using a ToolContext when a tool wants to provide visual feedback. A Context object is just a helper class - when you are working on a View you will need to do the work yourself.

  DrawShapeCommand command = new DrawShapeCommand(shape, paint, lineStyle, lineWidth);
  

  Here is the kind of work a tool context normally does for you:

  command.setMap( map );
  ViewportPane viewportPane = map.getRenderManager().getMapDisplay();
  viewportPane.addDrawCommand(command);
  Rectangle validArea;
  try{
      validArea = command.getValidArea();
  }catch (Exception e) {
      validArea=null;
  }
  if( validArea!=null )
      viewportPane.repaint(validArea.x, validArea.y, validArea.width, validArea.height);
  else
      viewportPane.repaint();
  

  A review of how draw commands work - they stay on the map until they are invalid. So at some point “later” you can take the command off the screen by:

  command.setValid(false);
  

If you want to issue a single command to schedule a series of shapes to be drawn look at the animation code example...it takes care using a timer to handle the the “later”.

• Implement a MapGraphic displaying data from a Blackboard

My last idea was to create public class TripMapGraphic implements MapGraphic {} , retrieving a List<Coordinate> from backboard, and drawing them.

This is a great way to provide visual feedback on the map. An example would be a TripMapGraphic that shows a route generated from the geotools graph module.

You can add a MapGraphic to the Map at any point by sending an “Add Layer” command to the Map. An example is the edit tool snapping functionality adding the Graph map graphic when “snapping to grid”.

This second way is good if you want to “mark up” the map for a specific purpose; say highlighting some roads on the screen in response to finding a route; using a MapGraphic will let your user pan and zoom around and still see your highlight.

The training course example coordinate map graphic and coordinate tool cover this kind of case. This example makes use of screen coordiantes; you will need to use the viewport world 2 screen and screen 2 world transformations to record the coordinates “on the map”.

Here is an example of storing points on the layer blackboard in lat/lon:

IBlackboard blackboard = map.getBlackboard();
List<Coordinate> points =  (List<Coordinate>) blackboard.get("route);

if (points == null) {
    points = new ArrayList<Coordinate>();
    blackboard.put(BLACKBOARD_KEY,points);
}
points.add( new Coordinate(lat,lon) ); // ie DefaultGeographicCRS.WGS84

You can draw these onto the screen in your mapgraphic render method

public void draw( MapGraphicContext context ) {

    //initialize the graphics handle
     ViewportGraphics g = context.getGraphics();
     g.setColor(Color.BLACK);
     g.setStroke(ViewportGraphics.LINE_SOLID, 1);

     //get the map blackboard
     IMap map = context.getLayer().getMap();
     IBlackboard blackboard = map.getBlackboard();

     List<Coordinate> coordinates = (List<Coordinate>) blackboard.get("route");

     if (coordinates == null) {
         return; //no coordinates to draw
     }

    MathTransform data2world = CRS.findMathTransform( DefaultGeographicCRS.WGS84, context.getCRS() );

     for (Coordinate coordinate : coordinates) {
        Ellipse2D e = new Ellipse2D.Double(
                coordinate.x-4,
                coordinate.y-4,
                10,10);
        g.draw(e);
        try {
            Coordinate worldCoord = JTS.transform(coordinate, null, data2world );
            Point point = context.worldToPixel(worldCoord);
            g.drawOval( point.x-1, point.y-2, 3, 3);
        } catch (TransformException e1) {
        }
     }
 }