Step 2: Creating a Navigation Connector

The navigation connector implements methods that define the following:

The initial (top-level) nodes of the connector. Once the navigation connector returns the initial nodes, the navigation service can query the nodes themselves for each one's children and related nodes.
The mapping between a navigation URL and the nodes defined by the connector.
For example, if the connector prefix is myPrefix, a navigation URL for the connector could be myPrefix://myParentNode2/myChildNode5. The connector must determine what INavigationConnectorNode object to return for this URL.

Procedure

Create a new class that extends AbstractNavigationConnector.
Implement the method getInitialNodes(), which returns a javax.naming.NamingEnumeration of the initial nodes. The method could return the nodes only if a condition is true, for example, the user is part of specific role.

The following returns the initial nodes if the user is part of the Java developer role:
Syntax
1. ```
public NamingEnumeration getInitialNodes(Hashtable environment) {
    if (isUserInJavaDeveloperRole(environment))
        return new NavigationEnum(initialNodes);
    else
        return NavigationEnum.EMPTY_ENUM;
}
```
End of the code.
Implement the method getNode(), which returns a INavigationConnectorNode based on a navigation URL passed into the method.
You can implement any logic for returning the INavigationConnectorNode object. One method is to store all the nodes in a Map with the navigation URL as the key. To implement this, you could do the following:
1. Store the nodes in a Map. You can add code for this in an initialization method.
  One implementation is to create a method that recursively traverses the navigation nodes that you created and store them in a map with the fully qualified name (navigation URL) as the key for each one, as follows:
  Syntax
  1. private void recursivelyFillNamesToNodesMap(NamingEnumeration enum) throws NamingException { while (enum.hasMoreElements()) { Binding b = (Binding) enum.nextElement(); myConnectorNode node = (myConnectorNode) b.getObject(); String nodeName = node.getName(); namesToNodes.put(nodeName, node); atomicNamesToNames.put(node.getAtomicName(),nodeName); recursivelyFillNamesToNodesMap(node.listBindings( "", INavigationConnectorNode.NAVIGATION_GET_CHILDREN)); } }
  End of the code.
  You could then call recursivelyFillNamesToNodesMap from the initialization method, passing in the initial nodes.
2. Look up the node in the Map based on the navigation URL. The following returns a node in the namesToNodes Map if the user is a member of the Java developer role:
  Syntax
  1. public INavigationConnectorNode getNode( Hashtable environment, String connectorNodeURL) { boolean memberOfRole = isUserInJavaDeveloperRole(environment); if (memberOfRole) return (myConnectorNode) namesToNodes.get(connectorNodeURL); else return null; }
  End of the code.

Implement the method getNodes(), which returns a javax.naming.NamingEnumeration of INavigationConnectorNode objects based on a set of navigation URLs passed into the method, as follows:

Syntax Syntax

public NamingEnumeration getNodes(
            Hashtable environment, Vector connectorNodeURLs) {
    myConnectorNode node;

    int size = connectorNodeURLs.size();
    List nodeBindings = new ArrayList(size);

    for (int i = 0; i < size; i++) {
        node = (myConnectorNode) getNode(
                        environment, (String) connectorNodeURLs.get(i));
        if (node != null) nodeBindings.add(
                        new Binding(node.getName(), node));
    }

    return new NavigationEnum(nodeBindings);
}

End of the code.

Implement the method getNodeByQuickLink() in order to map quick links strings to specific navigation nodes.
For more information, see Quick Links.
Implement getConnectorCacheDiscriminators(), which returns a set of keys, with each key representing a different entry point. The method enables caching of navigation nodes, improving performance.
For more information, see Navigation Cache.

Note

If your navigation connector does not support caching, implement getConnectorProfile(), and return a NavigationConnectorProfile object on which you called setCachable() and turned caching off.

End of the note.