Naming strategies

User supplied strategies for naming SEP nodes and filters

Introduction

The goal is to use a custom supplied strategy to customise node names and filters in the generated SEP. A custom naming strategy can add semantic value to member names that makes supporting the generated SEP easier. The default auto-generated names can be cryptic and confusing due to their similiarity..

Node naming

A developer can customise the node variable names declared within the final SEP by supplying Fluxtion with an instance of the NodeNameProducer interface. This can greatly help debugging and understanding the SEP code as contextually sensitive identifiers can be used within the SEP in place of the default cryptic auto-generated names. The interface is simple, Fluxtion compiler supplies the instance at generation time and the user returns a String name for the node. A null return delegates to the default auto-naming strategy.

public interface NodeNameProducer {
    String mappedNodeName(Object nodeToMap);
}

The Fluxtion maven plugin accepts the fully qualified class name of a NodeNameProducer in its configuration section:

<nodeNamingClass>[ NodeNameProducerClass ]</nodeNamingClass>

Filter naming

A developer can customise the filter names and comments declared within the final SEP by supplying Fluxtion with an instance of the FilterDescriptionProducer interface. This can greatly help debugging and understanding the SEP code as contextually sensitive identifiers can be used within the SEP in place of the default cryptic auto-generated names. The interface is simple as shown below. The developer returns a FilterDescription which also allows comments to be present at filter sites.

public interface FilterDescriptionProducer {
     
    default FilterDescription getFilterDescription(Class<? extends Event> event, int filterId){
        FilterDescription filter = new FilterDescription(event, filterId);
        filter.comment = null;
        filter.variableName = null;
        return filter;
    }
     
    default FilterDescription getFilterDescription(Class<? extends Event> event, String filterString){
        FilterDescription filter = new FilterDescription(event, filterString);
        filter.comment = null;
        filter.variableName = null;
        return filter;
    }
     
}

The Fluxtion maven plugin accepts the fully qualified class name of a FilterDescriptionProducer in its configuration section:

<filterNamingClass>[ NodeNameProducerClass ]</filterNamingClass>

Example

The following example implements node and filter naming strategies. The example code is located here.

Several data handlers are added to the graph in a SEPConfig. Each handler has a filter string, this value will be used to generate the node names in the SEP and the comments on the filters will be customised as well with this strategy:

public class NamingStrategy implements NodeNameProducer, FilterDescriptionProducer{

    @Override
    public String mappedNodeName(Object nodeToMap) {
        if(nodeToMap instanceof DataHandler){
            return "handler_" + ((DataHandler)nodeToMap).filter;
        }
        return null;
    }

    @Override
    public FilterDescription getFilterDescription(Class<? extends Event> event, String filterString) {
        FilterDescription filterDescription = FilterDescriptionProducer.super.getFilterDescription(event, filterString); 
        if(event == DataEvent.class){
            filterDescription.comment = "filtering for " + filterString;
        }
        return filterDescription;
    }
    
}

DataHandler node

public class DataHandler{

    final String filter;

    public DataHandler(String filter) {
        this.filter = filter;
    }
    
    @EventHandler(filterVariable = "filter")
    public void processUpdate(DataEvent event){
        
    }
    
}

SEPConfig

public class Builder extends SEPConfig{

    @Override
    public void buildConfig() {
        addNode(new DataHandler("FX"));
        addNode(new DataHandler("EQUITIES"));
        addNode(new DataHandler("BONDS"));
        //prevent mapped dispatch
        maxFiltersInline = 4;
    }
    
}

Generated SEP

The generated SEP now has nodes whose names are more human readable due to the naming strategy, see lines 4-6. Filter comments are customised in line with the filter strategy, lines 27, 32 and 37.

The SEP is much easier to map mentally making support a simpler task.

public class SampleProcessor implements EventHandler, BatchHandler, Lifecycle {

  //Node declarations
  private final DataHandler handler_FX = new DataHandler("FX");
  private final DataHandler handler_EQUITIES = new DataHandler("EQUITIES");
  private final DataHandler handler_BONDS = new DataHandler("BONDS");
  //Dirty flags

  //Filter constants

  public SampleProcessor() {}

  @Override
  public void onEvent(com.fluxtion.runtime.event.Event event) {
    switch (event.getClass().getName()) {
      case ("com.fluxtion.example.shared.DataEvent"):
        {
          DataEvent typedEvent = (DataEvent) event;
          handleEvent(typedEvent);
          break;
        }
    }
  }

  public void handleEvent(DataEvent typedEvent) {
    switch (typedEvent.filterString()) {
        //filtering for BONDS
      case ("BONDS"):
        handler_BONDS.processUpdate(typedEvent);
        afterEvent();
        return;
        //filtering for EQUITIES
      case ("EQUITIES"):
        handler_EQUITIES.processUpdate(typedEvent);
        afterEvent();
        return;
        //filtering for FX
      case ("FX"):
        handler_FX.processUpdate(typedEvent);
        afterEvent();
        return;
    }
    afterEvent();
  }
//code omitted for clarity
}