How to build an agent


With the release of OMERO 5.3.0, the OMERO.insight desktop client has entered maintenance mode, meaning it will only be updated if a major bug is discovered. Instead, the OME team will be focusing on developing the web clients. As a result, coding against this client is no longer recommended.

An agent is created and managed by the container. Building an agent is done in two steps:

  • write the agent code
  • declare the agent in the container.xml file located in the config directory

The agent intercepts the events posted on the event bus.


When a new version of the software is delivered, make sure you keep the container.xml shipped with the application and add the new agent entry to it.

Writing code

The following example creates a concrete agent MyBrowserAgent:

  • Create a myBrowser package in the agents package.
  • Create a class MyBrowserAgent, this class MUST implement the Agent interface to be initialized and the AgentListener to interact with other agents.
public class MyBrowserAgent
    implements Agent, AgentEventListener

    /** Reference to the registry. */
    private static Registry         registry;

    //no-arguments constructor required for initialization
    public MyBrowserAgent() {}

    //Follow methods required by the Agent Interface

    //No-op implementation in general
    public void activate()
        //this method will be invoked during the activation by the container

    //invoked before shutting down the application
    public boolean canTerminate() { return true; }

    //not yet implemented: invoked when shutting down the application
    public Map<String, Set> hasDataToSave() { return null; }

    //invoked while shutting down the application
    public void terminate() {}

    public void setContext(Registry ctx)
        //Must be a reference to the Agent Registry to access services.
        registry = ctx;

        //register the events the agent listens to e.g. BrowseImage
        EventBus bus = registry.getEventBus();
        bus.register(this, BrowseImage.class);


    //Follow methods required by the AgentEventListener Interface
    public void eventFired(AgentEvent e)
        if (e instanceof BrowseImage) {
            //Do something
            browseImage((BrowseImage) e);


Where to create the BrowseImage event

  • Create a myBrowser package in the package.
  • Create a BrowseImage event in the myBrowser package.
public class BrowseImage
    extends RequestEvent

    /** The id of the image to browse. */
    private long imageID;

     * Creates a new instance.
     * @param imageID The id of image to view.
    public BrowseImage(long imageID)
        if (imageID < 0)
            throw new IllegalArgumentException("ImageID not valid.");
        this.imageID = imageID;

     * Returns the ID of the image to browse.
     * @return See above.
    public long getImageID() { return imageID; }


Listening to the BrowseImage event

To listen to events posted on the event bus, the agent MUST implement the AgentListener Interface and register the events to listen to.

  • Register BrowseImage in the setContext(Registry) method of the Agent interface.
  • Listen to BrowseImage in the eventFired(AgentEvent) method of the AgentListener interface.

For example, when clicking on an image in the Data Manager, the following event is posted:

EventBus bus = registry.getEventBus(); BrowseImage(imageID));

The MyBrowserAgent handles the event

public void eventFired(AgentEvent e)
    if (e instanceof BrowseImage) {
        //Do something
        browseImage((BrowseImage) e);

Creating an agent’s view

See How to build an agent’s view

Declaring the agent

The MyBrowserAgent needs to be declared in the container.xml.

  • Open the container.xml located in the config folder (see Directory contents).

  • Add the following:

        <structuredEntry name="/agents" type="agents">
          <!-- NOTE FOR DEVELOPERS
               Add an agent tag for each of your Agents.
               The name tag specifies the human-readable name of the Agent.
               The active tag specifies if the agent is turned on or off.
               Set to true to turn the agent on, false otherwise.
               The class tag specifies the FQN of the Agent class.
               The config tag specifies the name of the Agent's
               configuration file within the config directory.
            <name>My Browser</name>
  • Create a mybrowser.xml and add it to the config directory:

    <?xml version="1.0" encoding="utf-8"?>
    <agent name="My Browser">
        <!-- This entry is turned into an instance of:
           This object can then be used to retrieve any image file within
           the directory pointed by the location tag. -->
           <structuredEntry name="/resources/icons/Factory" type = "icons">
             <!-- The location tag specifies the FQN of the package that contains
             the icon files. -->
           <!-- This entry is turned into an instance of java.awt.Font. -->
           <structuredEntry name="/resources/fonts/Titles" type="font">

    The file mybrowser.xml allows the agent to define specific parameters.