JBoss.orgCommunity Documentation


1. The Rule Engine
1.1. What is a Rule Engine?
1.1.1. Introduction and Background
1.2. Why use a Rule Engine?
1.2.1. Advantages of a Rule Engine
1.2.2. When should you use a Rule Engine?
1.2.3. When not to use a Rule Engine
1.2.4. Scripting or Process Engines
1.2.5. Strong and Loose Coupling
2. Quick Start
2.1. The Basics
2.1.1. Stateless Knowledge Session
2.1.2. Stateful Knowledge Session
2.2. A Little Theory
2.2.1. Methods versus Rules
2.2.2. Cross Products
2.2.3. Activations, Agenda and Conflict Sets.
2.3. More on building and deploying
2.3.1. Knowledge Base by Configuration Using Changesets
2.3.2. Knowledge Agent
3. User Guide
3.1. Building
3.1.1. Building using Code
3.1.2. Building using Configuration and the ChangeSet XML
3.2. Deploying
3.2.1. KnowledgePackage and Knowledge Definitions
3.2.2. KnowledgeBase
3.2.3. In-Process Building and Deployment
3.2.4. Building and Deployment in Separate Processes
3.2.5. StatefulknowledgeSessions and KnowledgeBase Modifications
3.2.6. KnowledgeAgent
3.3. Running
3.3.1. KnowledgeBase
3.3.2. StatefulKnowledgeSession
3.3.3. KnowledgeRuntime
3.3.4. Agenda
3.3.5. Event Model
3.3.6. KnowledgeRuntimeLogger
3.3.7. StatelessKnowledgeSession
3.3.8. Pipeline
3.3.9. Commands and the CommandExecutor
3.3.10. Marshalling
3.3.11. Persistence and Transactions
4. The Rule Language
4.1. Overview
4.1.1. A rule file
4.1.2. What makes a rule
4.2. Keywords
4.3. Comments
4.3.1. Single line comment
4.3.2. Multi-line comment
4.4. Error Messages
4.4.1. Message format
4.4.2. Error Messages Description
4.4.3. Other Messages
4.5. Package
4.5.1. import
4.5.2. global
4.6. Function
4.7. Type Declaration
4.7.1. Declaring New Types
4.7.2. Declaring Metadata
4.7.3. Declaring Metadata for Existing Types
4.7.4. Accessing Declared Types from the Application Code
4.8. Rule
4.8.1. Rule Attributes
4.8.2. Left Hand Side (when) Conditional Elements
4.8.3. The Right Hand Side (then)
4.8.4. A Note on Auto-boxing and Primitive Types
4.9. Query
4.10. Domain Specific Languages
4.10.1. When to use a DSL
4.10.2. Editing and managing a DSL
4.10.3. Using a DSL in your rules
4.10.4. Adding constraints to facts
4.10.5. How it works
4.10.6. Creating a DSL from scratch
4.10.7. Scope and keywords
4.10.8. DSLs in the BRMS and IDE
4.11. XML Rule Language
4.11.1. When to use XML
4.11.2. The XML format
4.11.3. Legacy Drools 2.x XML rule format
4.11.4. Automatic transforming between formats (XML and DRL)
5. Authoring
5.1. Decision Tables in Spreadsheets
5.1.1. When to use Decision tables
5.1.2. Overview
5.1.3. How decision tables work
5.1.4. Keywords and Syntax
5.1.5. Creating and integrating Spreadsheet based Decision Tables
5.1.6. Managing business rules in decision tables.
5.1.7. Rule Templates
5.2. Templates
5.2.1. The Rule Template File
5.2.2. Expanding a Template
5.2.3. Example
6. The Java Rule Engine API
6.1. Introduction
6.2. How To Use
6.2.1. Building and Registering RuleExecutionSets
6.2.2. Using Stateful and Stateless RuleSessions
6.3. References
7. The Rule IDE (Eclipse)
7.1. Features Outline
7.2. Creating a Rule Project
7.3. Creating a New Rule and Wizards
7.4. Textual Rule Editor
7.5. The Guided Editor (Rule GUI)
7.6. Drools Views
7.6.1. The Working Memory View
7.6.2. The Agenda View
7.6.3. The Global Data View
7.6.4. The Audit View
7.7. Domain Specific Languages
7.7.1. Editing languages
7.8. The Rete View
7.9. Large DRL Files
7.10. Debugging Rules
7.10.1. Creating Breakpoints
7.10.2. Debugging Rules
8. Examples
8.1. Getting the Examples
8.2. Hello World
8.3. State Example
8.3.1. Understanding the State Example
8.4. Fibonacci Example
8.5. Banking Tutorial
8.6. Pricing Rule Decision Table Example
8.6.1. Executing the example
8.6.2. The decision table
8.7. Pet Store Example
8.8. Honest Politician Example
8.9. Sudoku Example
8.9.1. Sudoku Overview
8.9.2. Running the Example
8.9.3. Java Source and Rules Overview
8.9.4. Sudoku Validator Rules (validatorSudoku.drl)
8.9.5. Sudoku Solving Rules (solverSudoku.drl)
8.9.6. Suggestions for Future Developments
8.10. Number Guess
8.11. Miss Manners and Benchmarking
8.11.1. Introduction
8.11.2. Indepth Discussion
8.11.3. Output Summary
8.12. Conway's Game Of Life

Artificial Intelligence (A.I.) is a very broad research area that focuses on "Making computers think like people" and includes disciplines such as Neural Networks, Genetic Algorithms, Decision Trees, Frame Systems and Expert Systems. Knowledge representation is the area of A.I. concerned with how knowledge is represented and manipulated. Expert Systems use Knowledge representation to facilitate the codification of knowledge into a knowledge base which can be used for reasoning, i.e., we can process data with this knowledge base to infer conclusions. Expert Systems are also known as Knowledge-based Systems and Knowledge-based Expert Systems and are considered to be "applied artificial intelligence". The process of developing with an Expert System is Knowledge Engineering. EMYCIN was one of the first "shells" for an Expert System, which was created from the MYCIN medical diagnosis Expert System. Whereas early Expert Systems had their logic hard-coded, "shells" separated the logic from the system, providing an easy to use environment for user input. Drools is a Rule Engine that uses the rule-based approach to implement an Expert System and is more correctly classified as a Production Rule System.

The term "Production Rule" originates from formal grammars where it is described as "an abstract structure that describes a formal language precisely, i.e., a set of rules that mathematically delineates a (usually infinite) set of finite-length strings over a (usually finite) alphabet" ( wikipedia ).

Business Rule Management Systems build additional value on top of a general purpose Rule Engine by providing business user focused systems for rule creation, management, deployment, collaboration, analysis and end user tools. Further adding to this value is the fast evolving and popular methodology "Business Rules Approach", which is a helping to formalize the role of Rule Engines in the enterprise.

The term Rule Engine is quite ambiguous in that it can be any system that uses rules, in any form, that can be applied to data to produce outcomes. This includes simple systems like form validation and dynamic expression engines. The book "How to Build a Business Rules Engine (2004)" by Malcolm Chisholm exemplifies this ambiguity. The book is actually about how to build and alter a database schema to hold validation rules. The book then shows how to generate VB code from those validation rules to validate data entry. This, while a very valid and useful topic for some, caused quite a surprise to this author, unaware at the time in the subtleties of Rules Engines' differences, who was hoping to find some hidden secrets to help improve the Drools engine. JBoss jBPM uses expressions and delegates in its Decision nodes which control the transitions in a Workflow. At each node it evaluates ther is a rule set that dictates the transition to undertake, and so this is also a Rule Engine. While a Production Rule System is a kind of Rule Engine and also an Expert System, the validation and expression evaluation Rule Engines mentioned previously are not Expert Systems.

A Production Rule System is Turing complete, with a focus on knowledge representation to express propositional and first order logic in a concise, non-ambiguous and declarative manner. The brain of a Production Rules System is an Inference Engine that is able to scale to a large number of rules and facts. The Inference Engine matches facts and data against Production Rules - also called Productions or just Rules - to infer conclusions which result in actions. A Production Rule is a two-part structure using First Order Logic for reasoning over knowledge representation.

when
    <conditions>
then
    <actions>;

The process of matching the new or existing facts against Production Rules is called Pattern Matching, which is performed by the Inference Engine. There are a number of algorithms used for Pattern Matching by Inference Engines including:

Drools implements and extends the Rete algorithm; Leaps used to be provided but was retired as it became unmaintained. The Drools Rete implementation is called ReteOO, signifying that Drools has an enhanced and optimized implementation of the Rete algorithm for object oriented systems. Other Rete based engines also have marketing terms for their proprietary enhancements to Rete, like RetePlus and Rete III. The most common enhancements are covered in "Production Matching for Large Learning Systems (Rete/UL)" (1995) by Robert B. Doorenbos.

The Rules are stored in the Production Memory and the facts that the Inference Engine matches against are kept in the Working Memory. Facts are asserted into the Working Memory where they may then be modified or retracted. A system with a large number of rules and facts may result in many rules being true for the same fact assertion; these rules are said to be in conflict. The Agenda manages the execution order of these conflicting rules using a Conflict Resolution strategy.


There are two methods of execution for a rule system: Forward Chaining and Backward Chaining; systems that implement both are called Hybrid Rule Systems. Understanding these two modes of operation is the key to understanding why a Production Rule System is different and how to get the best from it. Forward chaining is "data-driven" and thus reactionary, with facts being asserted into working memory, which results in one or more rules being concurrently true and scheduled for execution by the Agenda. In short, we start with a fact, it propagates and we end in a conclusion. Drools is a forward chaining engine.


Backward chaining is "goal-driven", meaning that we start with a conclusion which the engine tries to satisfy. If it can't it then searches for conclusions that it can satisfy; these are known as subgoals, that will help satisfy some unknown part of the current goal. It continues this process until either the initial conclusion is proven or there are no more subgoals. Prolog is an example of a Backward Chaining engine; Drools plans to provide support for Backward Chaining in a future release.


Some frequently asked questions:

We will attempt to address these questions below.

The shortest answer to this is "when there is no satisfactory traditional programming approach to solve the problem.". Given that short answer, some more explanation is required. The reason why there is no "traditional" approach is possibly one of the following:

If rules are a new technology for your project teams, the overhead in getting going must be factored in. It is not a trivial technology, but this document tries to make it easier to understand.

Typically in a modern OO application you would use a rule engine to contain key parts of your business logic, especially the really messy parts. This is an inversion of the OO concept of encapsulating all the logic inside your objects. This is not to say that you throw out OO practices, on the contrary in any real world application, business logic is just one part of the application. If you ever notice lots of conditional statements such as "if" and "switch", an overabundance of strategy patterns and other messy logic in your code that just doesn't feel right: that would be a place for rules. If there is some such logic and you keep coming back to fix it, either because you got it wrong, or the logic or your understanding changes: think about using rules. If you are faced with tough problems for which there are no algorithms or patterns: consider using rules.

Rules could be used embedded in your application or perhaps as a service. Often a rule engine works best as "stateful" component, being an integral part of an application. However, there have been successful cases of creating reusable rule services which are stateless.

For your organization it is important to decide about the process you will use for updating rules in systems that are in production. The options are many, but different organizations have different requirements. Frequently, rules maintenance is out of the control of the application vendors or project developers.

To quote a Drools mailing list regular:

 

It seems to me that in the excitement of working with rules engines, that people forget that a rules engine is only one piece of a complex application or solution. Rules engines are not really intended to handle workflow or process executions nor are workflow engines or process management tools designed to do rules. Use the right tool for the job. Sure, a pair of pliers can be used as a hammering tool in a pinch, but that's not what it's designed for.

 
 --Dave Hamu

As rule engines are dynamic (dynamic in the sense that the rules can be stored and managed and updated as data), they are often looked at as a solution to the problem of deploying software. (Most IT departments seem to exist for the purpose of preventing software being rolled out.) If this is the reason you wish to use a rule engine, be aware that rule engines work best when you are able to write declarative rules. As an alternative, you can consider data-driven designs (lookup tables), or script processing engines where the scripts are managed in a database and are able to be updated on the fly.

Hopefully the preceding sections have explained when you may want to use a rule engine.

Alternatives are script-based engines that provide the drive for "changes on the fly", and there are many such solutions.

Alternatively Process Engines (also capable of workflow) such as jBPM allow you to graphically (or programmatically) describe steps in a process. Those steps can also involve decision points which are in themselves a simple rule. Process engines and rules often can work nicely together, so they are not mutually exclusive.

One key point to note with rule engines is that some rule engines are really scripting engines. The downside of scripting engines is that you are tightly coupling your application to the scripts. If they are rules, you are effectively calling rules directly and this may cause more difficulty in future maintenance, as they tend to grow in complexity over time. The upside of scripting engines is that they can be easier to implement initially, producing results quickly, and are conceptually simpler for imperative programmers.

Many people have also implemented data-driven systems successfully in the past (where there are control tables that store meta-data that changes your applications behavior) - these can work well when the control can remain very limited. However, they can quickly grow out of control if extended too much (such that only the original creators can change the applications behavior) or they cause the application to stagnate as they are too inflexible.

So where do we get started, there are so many use cases and so much functionality in a rule engine such as Drools that it becomes beguiling. Have no fear my intrepid adventurer, the complexity is layered and you can ease yourself into with simple use cases.

Stateless session, not utilising inference, forms the simplest use case. A stateless session can be called like a function passing it some data and then receiving some results back. Some common use cases for stateless sessions are, but not limited to:

So let's start with a very simple example using a driving license application.

public class Applicant {
    private String name;
    private int age;
    private boolean valid;
    // getter and setter methods here
}

Now that we have our data model we can write our first rule. We assume that the application uses rules to refute invalid applications. As this is a simple validation use case we will add a single rule to disqualify any applicant younger than 18.

package com.company.license

rule "Is of valid age"
when
    $a : Applicant( age < 18 )
then
    $a.setValid( false );
end

To make the engine aware of data, so it can be processed against the rules, we have to insert the data, much like with a database. When the Applicant instance is inserted into the engine it is evaluated against the constraints of the rules, in this case just two constraints for one rule. We say two because the type Applicant is the first object type constraint, and age < 18 is the second field constraint. An object type constraint plus its zero or more field constraints is referred to as a pattern. When an inserted instance satisfies both the object type constraint and all the field constraints, it is said to be matched. The $a is a binding variable which permits us to reference the matched object in the consequence. There its properties can be updated. The dollar character ('$') is optional, but it helps to differentiate variable names from field names. The process of matching patterns against the inserted data is, not surprisingly, often referred to as pattern matching.

Let's assume that the rules are in the same folder as the classes, so we can use the classpath resource loader to build our first KnowledgeBase. A Knowledge Base is what we call our collection of compiled rules, which are compiled using the KnowledgeBuilder.

KnowledgeBuilder kbuilder = KnowledgeBuilderFactory.newKnowledgeBuilder();
kbuilder.add( ResourceFactory.newClassPathResource( "licenseApplication.drl", getClass() ),
              ResourceType.DRL );
if ( kbuilder.hasErrors() ) {
    System.err.println( builder.getErrors().toString() );
}
kbase.addKnowledgePackages( kbuilder.getKnowledgePackages() );

The above code snippet looks on the classpath for the licenseApplication.drl file, using the method newClassPathResource(). The resource type is DRL, short for "Drools Rule Language". Once the DRL file has been added we can check the Knowledge Builder object for any errors. If there are no errors, we can add the resulting packages to our Knowledge Base. Now we are ready to build our session and execute against some data:

StatelessKnowledgeSession ksession = kbase.newStatelessKnowledgeSession();
Applicant applicant = new Applicant( "Mr John Smith", 16 );
assertTrue( applicant.isValid() );
ksession.execute( applicant );
assertFalse( applicant.isValid() );

The preceding code executes the data against the rules. Since the applicant is under the age of 18, the application is marked as invalid.

So far we've only used a single instance, but what if we want to use more than one? We can execute against any object implementing Iterable, such as a collection. Let's add another class called Application, which has the date of the application, and we'll also move the boolean valid field to the Application class.

public class Applicant {
    private String name;
    private int age;
    // getter and setter methods here
}

public class Application {
    private Date dateApplied;
    private boolean valid;
    // getter and setter methods here
}

We can also add another rule to validate that the application was made within a period of time.

package com.company.license

rule "Is of valid age"
when
    Applicant( age < 18 )
    $a : Application()     
then
    $a.setValid( false );
end

rule "Application was made this year"
when
    $a : Application( dateApplied > "01-jan-2009" )     
then
    $a.setValid( false );
end

Unfortunately a Java array does not implement the Iterable interface, so we have to use the JDK converter method Arrays.asList(...). The code shown below executes against an iterable list, where all collection elements are inserted before any matched rules are fired.

StatelessKnowledgeSession ksession = kbase.newStatelessKnowledgeSession();
Applicant applicant = new Applicant( "Mr John Smith", 16 );
Application application = new Application();
assertTrue( application() );
ksession.execute( Arrays.asList( new Object[] { application, applicant } ) );
assertFalse( application() );

The two execute methods execute(Object object) and execute(Iterable objects) are actually convenience methods for the interface BatchExecutor's method execute(Command command).

A CommandFactory is used to create commands, so that the following is equivalent to execute(Iterable it):

ksession.execute( CommandFactory.newInsertIterable( new Object[] { application, applicant } ) );

Batch Executor and Command Factory are particularly useful when working with multiple Commands and with output identifiers for obtaining results.

List<Command> cmds = new ArrayList<Command>();
cmds.add( CommandFactory.newInsert( new Person( "Mr John Smith" ), "mrSmith" );
cmds.add( CommandFactory.newInsert( new Person( "Mr John Doe" ), "mrDoe" );
BatchExecutionResults results = ksession.execute( CommandFactory.newBatchExecution( cmds ) );
assertEquals( new Person( "Mr John Smith" ), results.getValue( "mrSmith" ) );

CommandFactory supports many other Commands that can be used in the BatchExecutor like StartProcess, Query, and SetGlobal.

Stateful Sessions are longer lived and allow iterative changes over time. Some common use cases for Stateful Sessions are, but not limited to:

In contrast to a Stateless Session, the dispose() method must be called afterwards to ensure there are no memory leaks, as the Knowledge Base contains references to Stateful Knowledge Sessions when they are created. StatefulKnowledgeSession also supports the BatchExecutor interface, like StatelessKnowledgeSession, the only difference being that the FireAllRules command is not automatically called at the end for a Stateful Session.

We illustrate the monitoring use case with an example for raising a fire alarm. Using just four classes, we represent rooms in a house, each of which has one sprinkler. If a fire starts in a room, we represent that with a single Fire instance.

public class Room {
    private String name
    // getter and setter methods here
}
public classs Sprinkler {
    private Room room;
    private boolean on;
    // getter and setter methods here
}
public class Fire {
    private Room room;
    // getter and setter methods here
}
public class Alarm {
}

In the previous section on Stateless Sessions the concepts of inserting and matching against data was introduced. That example assumed that only a single instance of each object type was ever inserted and thus only used literal constraints. However, a house has many rooms, so rules must express relationships between objects, such as a sprinkler being in a certain room. This is best done by using a binding variable as a constraint in a pattern. This "join" process results in what is called cross products, which are covered in the next section.

When a fire occurs an instance of the Fire class is created, for that room, and inserted into the session. The rule uses a binding on the room field of the Fire object to constrain matching to the sprinkler for that room, which is currently off. When this rule fires and the consequence is executed the sprinkler is turned on.

rule "When there is a fire turn on the sprinkler"
when
    Fire($room : room)
    $sprinkler : Sprinkler( room == $room, on == false )
then
    modify( $sprinkler ) { setOn( true ) };
    System.out.println( "Turn on the sprinkler for room " + $room.getName() );
end

Whereas the Stateless Session uses standard Java syntax to modify a field, in the above rule we use the modify statement, which acts as a sort of "with" statement. It may contain a series of comma separated Java expressions, i.e., calls to setters of the object selected by the modify statement's control expression. This modifies the data, and makes the engine aware of those changes so it can reason over them once more. This process is called inference, and it's essential for the working of a Stateful Session. Stateless Sessions typically do not use inference, so the engine does not need to be aware of changes to data. Inference can also be turned off explicitly by using the sequential mode.

So far we have rules that tell us when matching data exists, but what about when it does not exist? How do we determine that a fire has been extinguished, i.e., that there isn't a Fire object any more? Previously the constraints have been sentences according to Propositional Logic, where the engine is constraining against individual intances. Drools also has support for First Order Logic that allows you to look at sets of data. A pattern under the keyword not matches when something does not exist. The rule given below turns the sprinkler off as soon as the fire in that room has disappeared.

rule "When the fire is gone turn off the sprinkler"
when
    $room : Room( )
    $sprinkler : Sprinkler( room == $room, on == true )
    not Fire( room == $room )
then
    modify( $sprinkler ) { setOn( false ) };
    System.out.println( "Turn off the sprinkler for room " + $room.getName() );
end

While there is one sprinkler per room, there is just a single alarm for the building. An Alarm object is created when a fire occurs, but only one Alarm is needed for the entire building, no matter how many fires occur. Previously not was introduced to match the absence of a fact; now we use its complement exists which matches for one or more instances of some category.

rule "Raise the alarm when we have one or more fires"
when
    exists Fire()
then
    insert( new Alarm() );
    System.out.println( "Raise the alarm" );
end

Likewise, when there are no fires we want to remove the alarm, so the not keyword can be used again.

rule "Cancel the alarm when all the fires have gone"
when
    not Fire()
    $alarm : Alarm()
then
    retract( $alarm );
    System.out.println( "Cancel the alarm" );
end

Finally there is a general health status message that is printed when the application first starts and after the alarm is removed and all sprinklers have been turned off.

rule "Status output when things are ok"
when
    not Alarm()
    not Sprinkler( on === true ) 
then
    System.out.println( "Everything is ok" );
end

The above rules should be placed in a single DRL file and saved to some directory on the classpath and using the file name fireAlarm.drl, as in the Stateless Session example. We can then build a Knowledge Base, as before, just using the new name fireAlarm.drl. The difference is that this time we create a Stateful Session from the Knowledge Base, whereas before we created a Stateless Session.

KnowledgeBuilder kbuilder = KnowledgeBuilderFactory.newKnowledgeBuilder();
kbuilder.add( ResourceFactory.newClassPathResource( "fireAlarm.drl", getClass() ),
              ResourceType.DRL );
if ( kbuilder.hasErrors() ) {
    System.err.println( builder.getErrors().toString() );
}
kbase.addKnowledgePackages( kbuilder.getKnowledgePackages() );
StatefulKnowledgeSession ksession = kbase.newStatefulKnowledgeSession();

With the session created it is now possible to iteratvely work with it over time. Four Room objects are created and inserted, as well as one Sprinkler object for each room. At this point the engine has done all of its matching, but no rules have fired yet. Calling ksession.fireAllRules() allows the matched rules to fire, but without a fire that will just produce the health message.

String[] names = new String[]{"kitchen", "bedroom", "office", "livingroom"};
Map<String,Room> name2room = new HashMap<String,Room>();
for( String name: names ){
    Room room = new Room( name );
    name2room.put( name, room );
    ksession.insert( room );
    Sprinkler sprinkler = new Sprinkler( room );
    ksession.insert( sprinkler );
}

ksession.fireAllRules()
> Everything is ok

We now create two fires and insert them; this time a reference is kept for the returned FactHandle. A Fact Handle is an internal engine reference to the inserted instance and allows instances to be retracted or modified at a later point in time. With the fires now in the engine, once fireAllRules() is called, the alarm is raised and the respective sprinklers are turned on.

Fire kitchenFire = new Fire( name2room.get( "kitchen" ) );
Fire officeFire = new Fire( name2room.get( "office" ) );

FactHandle kitchenFireHandle = ksession.insert( kitchenFire );
FactHandle officeFireHandle = ksession.insert( officeFire );

ksession.fireAllRules();
> Raise the alarm
> Turn on the sprinkler for room kitchen
> Turn on the sprinkler for room office

After a while the fires will be put out and the Fire instances are retracted. This results in the sprinklers being turned off, the alarm being cancelled, and eventually the health message is printed again.

ksession.retract( kitchenFireHandle );
ksession.retract( officeFireHandle );

ksession.fireAllRules();
> Turn on the sprinkler for room office
> Turn on the sprinkler for room kitchen
> Cancel the alarm
> Everything is ok

Everyone still with me? That wasn't so hard and already I'm hoping you can start to see the value and power of a declarative rule system.

Earlier the term "cross product" was mentioned, which is the result of a join. Imagine for a moment that the data from the fire alarm example were used in combination with the following rule where there ar no field constraints:

rule
when
    $room : Room()
    $sprinkler : Sprinkler()
then
    System.out.println( "room:" + $room.getName() +
                        " sprinkler:" + $sprinkler.getRoom().getName() );
end

In SQL terms this would be like doing select * from Room, Sprinkler and every row in the Room table would be joined with every row in the Sprinkler table resulting in the following output:

room:office sprinker:office
room:office sprinkler:kitchen
room:office sprinkler:livingroom
room:office sprinkler:bedroom
room:kitchen sprinkler:office
room:kitchen sprinkler:kitchen
room:kitchen sprinkler:livingroom
room:kitchen sprinkler:bedroom
room:livingroom sprinkler:office
room:livingroom sprinkler:kitchen
room:livingroom sprinkler:livingroom
room:livingroom sprinkler:bedroom
room:bedroom sprinkler:office
room:bedroom sprinkler:kitchen
room:bedroom sprinkler:livingroom
room:bedroom sprinkler:bedroom

These cross products can obviously become huge, and they may very well contain spurious data. The size of cross products is often the source of performance problems for new rule authors. From this it can be seen that it's always desirable to constrain the cross products, which is done with the variable constraint.

rule
when
    $room : Room()
    $sprinkler : Sprinkler( room == $room )
then
    System.out.println( "room:" + $room.getName() +
                        " sprinkler:" + $sprinkler.getRoom().getName() );
end

This results in just four rows of data, with the correct Sprinkler for each Room. In SQL (actually HQL) the corresponding query would be select * from Room, Sprinkler where Room == Sprinkler.room.

room:office sprinkler:office
room:kitchen sprinkler:kitchen
room:livingroom sprinkler:livingroom
room:bedroom sprinkler:bedroom

So far the data and the matching process has been simple and small. To mix things up a bit a new example will be explored that handles cashflow calculations over date periods. The state of the engine will be illustratively shown at key stages to help get a better understanding of what is actually going on under the hood. Three classes will be used, as shown below.

public class CashFlow {
    private Date   date;
    private double amount;
    private int    type;
    long           accountNo;
    // getter and setter methods here
}

public class Account {
    private long   accountNo;
    private double balance;
    // getter and setter methods here
}

public AccountPeriod {
    private Date start;
    private Date end;
    // getter and setter methods here
}

By now you already know how to create Knowledge Bases and how to instantiate facts to populate the StatefulKnowledgeSession, so tables will be used to show the state of the inserted data, as it makes things clearer for illustration purposes. The tables below show that a single fact was inserted for the Account. Also inserted are a series of debits and credits as CashFlow objects for that account, extending over two quarters.


Two rules can be used to determine the debit and credit for that quarter and update the Account balance. The two rules below constrain the cashflows for an account for a given time period. Notice the "&&" which use short cut syntax to avoid repeating the field name twice.

rule "increase balance for credits"
when
  ap : AccountPeriod()
  acc : Account( $accountNo : accountNo )
  CashFlow( type == CREDIT,
    accountNo == $accountNo,
    date >= ap.start && <= ap.end,
    $amount : amount )
then
  acc.balance  += $amount;
end
rule "decrease balance for debits" 
when 
  ap : AccountPeriod() 
  acc : Account( $accountNo : accountNo ) 
  CashFlow( type == DEBIT, 
    accountNo == $accountNo,
    date >= ap.start && <= ap.end, 
    $amount : amount ) 
then 
  acc.balance -= $amount; 
end

If the AccountPeriod is set to the first quarter we constrain the rule "increase balance for credits" to fire on two rows of data and "decrease balance for debits" to act on one row of data.


The two cashflow tables above represent the matched data for the two rules. The data is matched during the insertion stage and, as you discovered in the previous chapter, does not fire straight away, but only after fireAllRules() is called. Meanwhile, the rule plus its matched data is placed on the Agenda and referred to as an Activation. The Agenda is a table of Activations that are able to fire and have their consequences executed, as soon as fireAllRules() is called. Activations on the Agenda are executed in turn. Notice that the order of execution so far is considered arbitrary.


After all of the above activations are fired, the account has a balance of -25.


If the AccountPeriod is updated to the second quarter, we have just a single matched row of data, and thus just a single Activation on the Agenda.


The firing of that Activation results in a balance of 25.


What if you don't want the order of Activation execution to be arbitrary? When there is one or more Activations on the Agenda they are said to be in conflict, and a conflict resolver strategy is used to determine the order of execution. At the simplest level the default strategy uses salience to determine rule priority. Each rule has a default value of 0, the higher the value the higher the priority. To illustrate this we add a rule to print the account balance, where we want this rule to be executed after all the debits and credits have been applied for all accounts. We achieve this by assigning a negative salience to this rule so that it fires after all rules with the default salience 0.

rule "Print balance for AccountPeriod"
        salience -50
    when
        ap : AccountPeriod()
        acc : Account()        
    then
        System.out.println( acc.accountNo + " : " + acc.balance );    
end

The table below depicts the resulting Agenda. The three debit and credit rules are shown to be in arbitrary order, while the print rule is ranked last, to execute afterwards.


Earlier we showed how rules would equate to SQL, which can often help people with an SQL background to understand rules. The two rules above can be represented with two views and a trigger for each view, as below:

select * from Account acc,
              Cashflow cf,
              AccountPeriod ap      
where acc.accountNo == cf.accountNo and 
      cf.type == CREDIT and
      cf.date >= ap.start and 
      cf.date <= ap.end
select * from Account acc, 
              Cashflow cf,
              AccountPeriod ap 
where acc.accountNo == cf.accountNo and 
      cf.type == DEBIT and
      cf.date >= ap.start and 
      cf.date <= ap.end
trigger : acc.balance += cf.amount
trigger : acc.balance -= cf.amount

Drools also features ruleflow-group attributes which allows workflow diagrams to declaratively specify when rules are allowed to fire. The screenshot below is taken from Eclipse using the Drools plugin. It has two ruleflow-group nodes which ensures that the calculation rules are executed before the reporting rules.

The use of the ruleflow-group attribute in a rule is shown below.

rule "increase balance for credits"
  ruleflow-group "calculation"
when
  ap : AccountPeriod()
  acc : Account( $accountNo : accountNo )
  CashFlow( type == CREDIT,
            accountNo == $accountNo,
            date >= ap.start && <= ap.end,
            $amount : amount )
then
  acc.balance  += $amount;
end
rule "Print balance for AccountPeriod"
  ruleflow-group "report"
when
  ap : AccountPeriod()
  acc : Account()
then
  System.out.println( acc.accountNo +
                      " : " + acc.balance );    
end

So far, the programmatic API has been used to build a Knowledge Base. Quite often it's more desirable to do this via configuration. To facilitate this, Drools supports the "Changeset" feature. The file changeset.xml contains a list of resources, and it may also point recursively to another changeset XML file. Currently the changeset has only a single "add" element, but support for remove and modify will be added in the future, for more powerful incremental changes over time. Currently there is no XML schema for the changeset XML, but we hope to add one soon. A few examples will be shown to give you the gist of things. A resource approach is employed that uses a prefix to indicate the protocol. All the protocols provided by java.net.URL, such as "file" and "http", are supported, as well as an additional "classpath". Currently the type attribute must always be specified for a resource, as it is not inferred from the file name extension. Here is a simple example that points to a http location for some rules.

 <change-set xmlns='http://drools.org/drools-5.0/change-set'
             xmlns:xs='http://www.w3.org/2001/XMLSchema-instance'
             xs:schemaLocation='http://drools.org/drools-5.0/change-set drools-change-set-5.0.xsd' >
   <add>
       <resource source='http:org/domain/myrules.drl' type='DRL' />
   </add>
 </change-set>

To use the above XML, the code is almost identical as before, except we change the resource type to CHANGE_SET.

KnowledgeBuilder kbuilder = KnowledgeBuilderFactory.newKnowledgeBuilder();
kbuilder.add( ResourceFactory.newClasspathResource( "myChangeSet.xml", getClass() ),
              ResourceType.CHANGE_SET );
if ( kbuilder.hasErrors() ) {
    System.err.println( builder.getErrors().toString() );
}

Changesets can include any number of resources, and they even support additional configuration information, which currently is only needed for decision tables. The example below is expanded to load the rules from a http URL location, and an Excel decision table from the classpath.

 <change-set xmlns='http://drools.org/drools-5.0/change-set'
             xmlns:xs='http://www.w3.org/2001/XMLSchema-instance'
             xs:schemaLocation='http://drools.org/drools-5.0/change-set.xsd' >
   <add>
       <resource source='http:org/domain/myrules.drl' type='DRL' />
       <resource source='classpath:data/IntegrationExampleTest.xls' type="DTABLE">
           <decisiontable-conf input-type="XLS" worksheet-name="Tables_2" />
       </resource>
   </add>
 </change-set>

It is also possible to specify a directory, to add the contents of that directory. It is expected that all the files are of the specified type, since type is not yet inferred from the file name extensions.

 <change-set xmlns='http://drools.org/drools-5.0/change-set'
             xmlns:xs='http://www.w3.org/2001/XMLSchema-instance'
             xs:schemaLocation='http://drools.org/drools-5.0/change-set.xsd' >
   <add>
       <resource source='file://myfolder/' type='DRL' />
   </add>
 </change-set>

The KnowledgeBuilder is responsible for taking source files, such as a DRL file or an Excel file, and turning them into a Knowledge Package of rule and process definitions which a Knowledge Base can consume. An object of the class ResourceType indicates the type of resource it is being asked to build.

The ResourceFactory provides capabilities to load resources from a number of sources, such as Reader, ClassPath, URL, File, or ByteArray. Binaries, such as decision tables (Excel .xls files), should not use a Reader based resource handler, which is only suitable for text based resources.


The KnowlegeBuilder is created using the KnowledgeBuilderFactory.


A KnowledgeBuilder can be created using the default configuration.


A configuration can be created using the KnowledgeBuilderFactory. This allows the behavior of the Knowledge Builder to be modified. The most common usage is to provide a custom class loader so that the KnowledgeBuilder object can resolve classes that are not in the default classpath. The first parameter is for properties and is optional, i.e., it may be left null, in which case the default options will be used. The options parameter can be used for things like changing the dialect or registering new accumulator functions.


Resources of any type can be added iteratively. Below, a DRL file is added. Unlike Drools 4.0 Package Builder, the Knowledge Builder can now handle multiple namespaces, so you can just keep adding resources regardless of namespace.


It is best practice to always check the hasErrors() method after an addition. You should not add more resources or retrieve the Knowledge Packages if there are errors. getKnowledgePackages() returns an empty list if there are errors.


When all the resources have been added and there are no errors the collection of Knowledge Packages can be retrieved. It is a Collection because there is one Knowledge Package per package namespace. These Knowledge Packages are serializable and often used as a unit of deployment.


The final example puts it all together.


Instead of adding the resources to create definitions programmatically it is also possible to do it by configuration, via the ChangeSet XML. The simple XML file supports three elements: add, remove, and modify, each of which has a sequence of <resource> subelements defining a configuration entity. The following XML schema is not normative and intended for illustration only.

Example 3.7. XML Schema for ChangeSet XML (not normative)

<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"
           xmlns="http://drools.org/drools-5.0/change-set"
           targetNamespace="http://drools.org/drools-5.0/change-set">

  <xs:element name="change-set" type="ChangeSet"/>

  <xs:complexType name="ChangeSet">
    <xs:choice maxOccurs="unbounded">
      <xs:element name="add"    type="Operation"/>
      <xs:element name="remove" type="Operation"/>
      <xs:element name="modify" type="Operation"/>
    </xs:choice>
  </xs:complexType>

  <xs:complexType name="Operation">
    <xs:sequence>
      <xs:element name="resource" type="Resource"
                  maxOccurs="unbounded"/>
    </xs:sequence>
  </xs:complexType>

  <xs:complexType name="Resource">
    <xs:sequence>
      
      <xs:element name="decisiontable-conf" type="DecTabConf"
                  minOccurs="0"/>
    </xs:sequence>
    
    <xs:attribute name="source" type="xs:string"/>
    <xs:attribute name="type"   type="ResourceType"/>
  </xs:complexType>

  <xs:complexType name="DecTabConf">
    <xs:attribute name="input-type"     type="DecTabInpType"/>
    <xs:attribute name="worksheet-name" type="xs:string"
                  use="optional"/>
  </xs:complexType>

  
  <xs:simpleType name="ResourceType">
    <xs:restriction base="xs:string">
      <xs:enumeration value="DRL"/>
      <xs:enumeration value="XDRL"/>
      <xs:enumeration value="DSL"/>
      <xs:enumeration value="DSLR"/>
      <xs:enumeration value="DRF"/>
      <xs:enumeration value="DTABLE"/>
      <xs:enumeration value="PKG"/>
      <xs:enumeration value="BRL"/>
      <xs:enumeration value="CHANGE_SET"/>
    </xs:restriction>
  </xs:simpleType>

  
  <xs:simpleType name="DecTabInpType">
    <xs:restriction base="xs:string">
      <xs:enumeration value="XLS"/>
      <xs:enumeration value="CSV"/>
    </xs:restriction>
  </xs:simpleType>

</xs:schema>

Currently only the add element is supported, but the others will be implemented to support iterative changes. The following example loads a single DRL file.


Notice the file: prefix, which signifies the protocol for the resource. The Change Set supports all the protocols provided by java.net.URL, such as "file" and "http", as well as an additional "classpath". Currently the type attribute must always be specified for a resource, as it is not inferred from the file name extension. Using the ClassPath resource loader in Java allows you to specify the Class Loader to be used to locate the resource but this is not possible from XML. Instead, the Class Loader will default to the one used by the Knowledge Builder unless the ChangeSet XML is itself loaded by the ClassPath resource, in which case it will use the Class Loader specified for that resource.

Currently you still need to use the API to load that ChangeSet, but we will add support for containers such as Spring in the future, so that the process of creating a Knowledge Base can be done completely by XML configuration. Loading resources using an XML file couldn't be simpler, as it's just another resource type.


ChangeSets can include any number of resources, and they even support additional configuration information, which currently is only needed for decision tables. Below, the example is expanded to load rules from a http URL location, and an Excel decision table from the classpath.


The ChangeSet is especially useful when working with a Knowledge Agent, as it allows for change notification and automatic rebuilding of the Knowledge Base, which is covered in more detail in the section on the Knowledge Agent, under Deploying.

Directories can also be specified, to add all resources in that folder. Currently it is expected that all resources in that folder are of the same type. If you use the Knowledge Agent it will provide a continous scanning for added, modified or removed resources and rebuild the cached Knowledge Base. The KnowledgeAgent provides more information on this.


The Knowlege Base is a repository of all the application's knowledge definitions. It may contain rules, processes, functions, and type models. The Knowledge Base itself does not contain instance data, known as facts; instead, sessions are created from the Knowledge Base into which data can be inserted and where process instances may be started. Creating the Knowlege Base can be heavy, whereas session creation is very light, so it is recommended that Knowledge Bases be cached where possible to allow for repeated session creation.


A KnowledgeBase object is also serializable, and some people may prefer to build and then store a KnowledgeBase, treating it also as a unit of deployment, instead of the Knowledge Packages.

The KnowlegeBase is created using the KnowledgeBaseFactory.


A KnowledgeBase can be created using the default configuration.


If a custom class loader was used with the KnowledgeBuilder to resolve types not in the default class loader, then that must also be set on the KnowledgeBase. The technique for this is the same as with the KnowledgeBuilder.


Both the KnowledgeBase and the KnowledgePackage are units of deployment and serializable. This means you can have one machine do any necessary building, requiring drools-compiler.jar, and have another machine deploy and execute everything, needing only drools-core.jar.

Although serialization is standard Java, we present an example of how one machine might write out the deployment unit and how another machine might read in and use that deployment unit.



The KnowledgeBase is also serializable and some people may prefer to build and then store the KnowledgeBase itself, instead of the Knowledge Packages.

Drools Guvnor, our server side management system, uses this deployment approach. After Guvnor has compiled and published serialized Knowledge Packages on a URL, Drools can use the URL resource type to load them.

The KnowlegeAgent provides automatic loading, caching and re-loading of resources and is configured from a properties files. The Knowledge Agent can update or rebuild this Knowlege Base as the resources it uses are changed. The strategy for this is determined by the configuration given to the factory, but it is typically pull-based using regular polling. We hope to add push-based updates and rebuilds in future versions. The Knowledge Agent will continuously scan all the added resources, using a default polling interval of 60 seconds. If their date of the last modification is updated it will rebuild the cached Knowledge Base using the new resources.


The KnowlegeBuilder is created using a KnowledgeBuilderFactory object. The agent must specify a name, which is used in the log files to associate a log entry with the corresponding agent.



The following example constructs an agent that will build a new KnowledgeBase from the specified ChangeSet. (See section "Building" for more details on the ChangeSet format.) Note that the method can be called iteratively to add new resources over time. At the moment the remove and modify elements of the ChangeSet XML are not implemented, but future versions will provide this as well, giving you more control over those incremental changes. The Knowledge Agent polls the resources added from the ChangeSet every 60 seconds, the default interval, to see if they are updated. Whenever changes are found it will construct a new Knowledge Base. If the change set specifies a resource that is a directory its contents will be scanned for changes, too.


Resource scanning is not on by default, it's a service and must be started, and the same is true for notification. Both can be done via the ResourceFactory.


The default resource scanning period may be changed via the ResourceChangeScannerService. A suitably updated ResourceChangeScannerConfiguration object is passed to the service's configure() method, which allows for the service to be reconfigured on demand.


Knowledge Agents can take an empty Knowledge Base or a populated one. If a populated Knowledge Base is provided, the Knowledge Agent will run an iterator from Knowledge Base and subscribe to the resources that it finds. While it is possible for the Knowledge Builder to build all resources found in a directory, that information is lost by the Knowledge Builder so that those directories will not be continuously scanned. Only directories specified as part of the applyChangeSet(Resource) method are monitored.

One of the advantages of providing KnowledgeBase as the starting point is that you can provide it with a KnowledgeBaseConfiguration. When resource changes are detected and a new KnowledgeBase object is instantiated, it will use the KnowledgeBaseConfiguration of the previous KnowledgeBase object.


In the above example getKnowledgeBase() will return the same provided kbase instance until resource changes are detected and a new Knowledge Base is built. When the new Knowledge Base is built, it will be done with the KnowledgeBaseConfiguration that was provided to the previous KnowledgeBase.

As mentioned previously, a ChangeSet XML can specify a directory and all of its contents will be added. If this ChangeSet XML is used with the applyChangeSet() method it will also add any directories to the scanning process. When the directory scan detects an additional file, it will be added to the Knowledge Base; any removed file is removed from the Knowledge Base, and modified files will, as usual, force the build of a new Knowledge Base using the latest version.


Note that for the resource type PKG the drools-compiler dependency is not needed as the Knowledge Agent is able to handle those with just drools-core.

The KnowledgeAgentConfiguration can be used to modify a Knowledge Agent's default behavior. You could use this to load the resources from a directory, while inhibiting the continuous scan for changes of that directory.


Previously we mentioned Drools Guvnor and how it can build and publish serialized Knowledge Packages on a URL, and that the ChangeSet XML can handle URLs and Packages. Taken together, this forms an importanty deployment scenario for the Knowledge Agent.

The WorkingMemoryEntryPoint provides the methods around inserting, updating and retrieving facts. The term "entry point" is related to the fact that we have multiple partitions in a Working Memory and you can choose which one you are inserting into, although this use case is aimed at event processing and covered in more detail in the Fusion manual. Most rule based applications will work with the default entry point alone.

The KnowledgeRuntime interface provides the main interaction with the engine. It is available in rule consequences and process actions. In this manual the focus is on the methods and interfaces related to rules, and the methods pertaining to processes will be ignored for now. But you'll notice that the KnowledgeRuntime inherits methods from both the WorkingMemory and the ProcessRuntime, thereby providing a unified API to work with processes and rules. When working with rules, three interfaces form the KnowledgeRuntime: WorkingMemoryEntryPoint, WorkingMemory and the KnowledgeRuntime itself.


Insertion is the act of telling the WorkingMemory about a fact, which you do by ksession.insert(yourObject), for example. When you insert a fact, it is examined for matches against the rules. This means all of the work for deciding about firing or not firing a rule is done during insertion; no rule, however, is executed until you call fireAllRules(), which you call after you have finished inserting your facts. It is a common misunderstanding for people to think the condition evaluation happens when you call fireAllRules(). Expert systems typically use the term assert or assertion to refer to facts made available to the system. However, due to "assert" being a keyword in most languages, we have decided to use the insert keyword; so expect to hear the two used interchangeably.

When an Object is inserted it returns a FactHandle. This FactHandle is the token used to represent your inserted object within the WorkingMemory. It is also used for interactions with the WorkingMemory when you wish to retract or modify an object.

Cheese stilton = new Cheese("stilton");
FactHandle stiltonHandle = ksession.insert( stilton );      

As mentioned in the Knowledge Base section, a Working Memory may operate in two assertion modes, i.e., equality or identity, with identity being the default.

Identity means that the Working Memory uses an IdentityHashMap to store all asserted objects. New instance assertions always result in the return of a new FactHandle, but if an instance is asserted again then it returns the original fact handle, i.e., it ignores repeated insertions for the same fact.

Equality means that the Working Memory uses a HashMap to store all asserted Objects. New instance assertions will only return a new FactHandle if no equal objects have been asserted.

The Agenda is a Rete feature. During actions on the WorkingMemory, rules may become fully matched and eligible for execution; a single Working Memory Action can result in multiple eligible rules. When a rule is fully matched an Activation is created, referencing the rule and the matched facts, and placed onto the Agenda. The Agenda controls the execution order of these Activations using a Conflict Resolution strategy.

The engine cycles repeatedly through two phases:


The process repeats until the agenda is clear, in which case control returns to the calling application. When Working Memory Actions are taking place, no rules are being fired.


The event package provides means to be notified of rule engine events, including rules firing, objects being asserted, etc. This allows you, for instance, to separate logging and auditing activities from the main part of your application (and the rules).

The KnowlegeRuntimeEventManager interface is implemented by the KnowledgeRuntime which provides two interfaces, WorkingMemoryEventManager and ProcessEventManager. We will only cover the WorkingMemoryEventManager here.


The WorkingMemoryEventManager allows for listeners to be added and removed, so that events for the working memory and the agenda can be listened to.


The following code snippet shows how a simple agenda listener is declared and attached to a session. It will print activations after they have fired.


Drools also provides DebugWorkingMemoryEventListener and DebugAgendaEventListener which implement each method with a debug print statement. To print all Working Memory events, you add a listener like this:


All emitted events implement the KnowlegeRuntimeEvent interface which can be used to retrieve the actual KnowlegeRuntime the event originated from.


The events currently supported are:

  • ActivationCreatedEvent

  • ActivationCancelledEvent

  • BeforeActivationFiredEvent

  • AfterActivationFiredEvent

  • AgendaGroupPushedEvent

  • AgendaGroupPoppedEvent

  • ObjectInsertEvent

  • ObjectRetractedEvent

  • ObjectUpdatedEvent

  • ProcessCompletedEvent

  • ProcessNodeLeftEvent

  • ProcessNodeTriggeredEvent

  • ProcessStartEvent

The StatelessKnowledgeSession wraps the StatefulKnowledgeSession, instead of extending it. Its main focus is on decision service type scenarios. It avoids the need to call dispose(). Stateless sessions do not support iterative insertions and the method call fireAllRules() from Java code; the act of calling execute() is a single-shot method that will internally instantiate a StatefulKnowledgeSession, add all the user data and execute user commands, call fireAllRules(), and then call dispose(). While the main way to work with this class is via the BatchExecution (a subinterface of Command) as supported by the CommandExecutor interface, two convenience methods are provided for when simple object insertion is all that's required. The CommandExecutor and BatchExecution are talked about in detail in their own section.


Our simple example shows a stateless session executing a given collection of Java objects using the convenience API. It will iterate the collection, inserting each element in turn.


If this was done as a single Command it would be as follows:


If you wanted to insert the collection itself, and the collection's individual elements, then CommandFactory.newInsert(collection) would do the job.

Methods of the CommandFactory create the supported commands, all of which can be marshalled using XStream and the BatchExecutionHelper. BatchExecutionHelper provides details on the XML format as well as how to use Drools Pipeline to automate the marshalling of BatchExecution and ExecutionResults.

StatelessKnowledgeSession supports globals, scoped in a number of ways. I'll cover the non-command way first, as commands are scoped to a specific execution call. Globals can be resolved in three ways.

The CommandExecutor interface also offers the ability to export data via "out" parameters. Inserted facts, globals and query results can all be returned.


With Rete you have a stateful session where objects can be asserted and modified over time, and where rules can also be added and removed. Now what happens if we assume a stateless session, where after the initial data set no more data can be asserted or modified and rules cannot be added or removed? Certainly it won't be necessary to re-evaluate rules, and the engine will be able to operate in a simplified way.

The LeftInputAdapterNode no longer creates a Tuple, adding the Object, and then propagate the Tuple – instead a Command object is created and added to a list in the Working Memory. This Command object holds a reference to the LeftInputAdapterNode and the propagated object. This stops any left-input propagations at insertion time, so that we know that a right-input propagation will never need to attempt a join with the left-inputs (removing the need for left-input memory). All nodes have their memory turned off, including the left-input Tuple memory but excluding the right-input object memory, which means that the only node remembering an insertion propagation is the right-input object memory. Once all the assertions are finished and all right-input memories populated, we can then iterate the list of LeftInputAdatperNode Command objects calling each in turn. They will propagate down the network attempting to join with the right-input objects, but they won't be remembered in the left input as we know there will be no further object assertions and thus propagations into the right-input memory.

There is no longer an Agenda, with a priority queue to schedule the Tuples; instead, there is simply an array for the number of rules. The sequence number of the RuleTerminalNode indicates the element within the array where to place the Activation. Once all Command objects have finished we can iterate our array, checking each element in turn, and firing the Activations if they exist. To improve performance, we remember the first and the last populated cell in the array. The network is constructed, with each RuleTerminalNode being given a sequence number based on a salience number and its order of being added to the network.

Typically the right-input node memories are Hash Maps, for fast object retraction; here, as we know there will be no object retractions, we can use a list when the values of the object are not indexed. For larger numbers of objects indexed Hash Maps provide a performance increase; if we know an object type has only a few instances, indexing is probably not advantageous, and a list can be used.

Sequential mode can only be used with a Stateless Session and is off by default. To turn it on, either call RuleBaseConfiguration.setSequential(true), or set the rulebase configuration property drools.sequential to true. Sequential mode can fall back to a dynamic agenda by calling setSequentialAgenda with SequentialAgenda.DYNAMIC. You may also set the "drools.sequential.agenda" property to "sequential" or "dynamic".

The PipelineFactory and associated classes are there to help with the automation of getting information into and out of Drools, especially when using services such as Java Message Service (JMS), and other data sources that aren't Java objects. Transformers for Smooks, JAXB, XStream and jXLS are povided. Smooks is an ETL (extract, transform, load) tool and can work with a variety of data sources. JAXB is a Java standard for XML binding capable of working with XML schemas. XStream is a simple and fast XML serialisation framework. jXLS finally allows for loading of Java objects from an Excel spreadsheet. Minimal information on these technologies will be provided here; beyond this, you should consult the relevant user guide for each of these tools.


Pipeline is not meant as a replacement for products like the more powerful Apache Camel. It is a simple framework aimed at the specific Drools use cases.

In Drools, a pipeline is a series of stages that operate on and propagate a given payload. Typically this starts with a Pipeline instance which is responsible for taking the payload, creating a PipelineContext for it and propagating that to the first receiver stage. Two subtypes of Pipeline are provided, both requiring a different PipelineContext: StatefulKnowledgeSessionPipeline and StatelessKnowledgeSessionPipeline. PipelineFactory provides methods to create both of the two Pipeline subtypes. Notice that both factory methods take the relevant session as an argument. The construction of a StatefulKnowledgeSessionPipeline is shown below, where also its receiver is set.


A pipeline is then made up of a chain of Stages that implement both the Emitter and the Receiver interfaces. The Emitter interface enables the Stage to propagate a payload, and the Receiver interface lets it receive a payload. This is why the Pipeline interface only implements Emitter and Stage and not Receiver, as it is the first instance in the chain. The Stage interface allows a custom exception handler to be set on the Stage object.


The Transformer interface extends Stage, Emitter and Receiver, providing those interface methods as a single type. Its other purpose is that of a marker interface indicating this particulare role of the implementing class. (We have several other marker interfaces such as Expression and Action, both of which also extend Stage, Emitter and Receiver.) One of the stages should be responsible for setting a result value on the PipelineContext. It's the responsibility of the ResultHandler interface, to be implemented by the user, to process on these results. It may do so by inserting them into some suitable object, whence the user's code may retrieve them.


While the above example shows a simple handler that merely assigns the result to a field that the user can access, it could do more complex work like sending the object as a message.

Pipeline provides an adapter to insert the payload and to create the correct Pipeline Context internally.

In general it is easier to construct the pipelines in reverse. In the following example XML data is loaded from disk, transformed with XStream and finally inserted into the session.


While the above example is for loading a resource from disk, it is also possible to work from a running messaging service. Drools currently provides a single service for JMS, called JmsMessenger. Support for other services will be added later. The code below shows part of a unit test which illustrates part of the JmsMessenger in action:

Example 3.38. Using JMS with Pipeline

// As this is a service, it's more likely that
// the results will be logged or sent as a return message.
Action resultHandlerStage = PipelineFactory.newExecuteResultHandler();

// Insert the transformed object into the session associated with the PipelineContext
KnowledgeRuntimeCommand insertStage = PipelineFactory.newStatefulKnowledgeSessionInsert();
insertStage.setReceiver( resultHandlerStage );

// Create the transformer instance and create the Transformer stage where we are
// going from XML to Pojo. JAXB needs an array of the available classes.
JAXBContext jaxbCtx = KnowledgeBuilderHelper.newJAXBContext( classNames,
                                                              kbase );
Unmarshaller unmarshaller = jaxbCtx.createUnmarshaller();
Transformer transformer = PipelineFactory.newJaxbFromXmlTransformer( unmarshaller );
transformer.setReceiver( insertStage );

// Payloads for JMS arrive in a Message wrapper: we need to unwrap this object.
Action unwrapObjectStage = PipelineFactory.newJmsUnwrapMessageObject();
unwrapObjectStage.setReceiver( transformer );

// Create the start adapter Pipeline for StatefulKnowledgeSessions
Pipeline pipeline = PipelineFactory.newStatefulKnowledgeSessionPipeline( ksession );
pipeline.setReceiver( unwrapObjectStage );

// Services, like JmsMessenger take a ResultHandlerFactory implementation.
// This is because a result handler must be created for each incoming message.
ResultHandlerFactory factory = new ResultHandlerFactoryImpl();
Service messenger = PipelineFactory.newJmsMessenger( pipeline,
                                                     props,
                                                     destinationName,
                                                     factory );
messenger.start();

The Transformer objects are JaxbFromXmlTransformer and JaxbToXmlTransformer. The former uses an javax.xml.bind.Unmarshaller for converting an XML document into a content tree; the latter serializes a content tree to XML by passing it to a javax.xml.bind.Marshaller. Both of these objects can be obtained from a JAXBContext object.

A JAXBContext maintains the set of Java classes that are bound to XML elements. Such classes may be generated from an XML schema, by compiling it with JAXB's schema compiler xjc. Alternatively, handwritten classes can be augmented with annotations from jaxb.xml.bind.annotation.

Unmarshalling an XML document results in an object tree. Inserting objects from this tree as facts into a session can be done by walking the tree and inserting nodes as appropriate. This could be done in the context of a pipeline by a custom Transformer that emits the nodes one by one to its receiver.




This transformer creates a new JmsMessenger which runs as a service in its own thread. It expects an existing JNDI entry for "ConnectionFactory", used to create the MessageConsumer which will feed into the specified pipeline.

Example 3.47. JMS Messenger stage

// As this is a service, it's more likely the results will be logged
// or sent as a return message.
Action resultHandlerStage = PipelineFactory.newExecuteResultHandler();

// Insert the transformed object into the session associated with the PipelineContext
KnowledgeRuntimeCommand insertStage = PipelineFactory.newStatefulKnowledgeSessionInsert();
insertStage.setReceiver( resultHandlerStage );

// Create the transformer instance and create the Transformer stage,
// where we are going from XML to Java object.
// JAXB needs an array of the available classes
JAXBContext jaxbCtx = KnowledgeBuilderHelper.newJAXBContext( classNames, kbase );
Unmarshaller unmarshaller = jaxbCtx.createUnmarshaller();
Transformer transformer = PipelineFactory.newJaxbFromXmlTransformer( unmarshaller );
transformer.setReceiver( insertStage );

// Payloads for JMS arrive in a Message wrapper, we need to unwrap this object.
Action unwrapObjectStage = PipelineFactory.newJmsUnwrapMessageObject();
unwrapObjectStage.setReceiver( transformer );

// Create the start adapter Pipeline for StatefulKnowledgeSessions
Pipeline pipeline = PipelineFactory.newStatefulKnowledgeSessionPipeline( ksession );
pipeline.setReceiver( unwrapObjectStage );

// Services like JmsMessenger take a ResultHandlerFactory implementation.
// This is so because a result handler must be created for each incoming message.
ResultHandleFactoryImpl factory = new ResultHandleFactoryImpl();
Service messenger = PipelineFactory.newJmsMessenger( pipeline,
                                                     props,
                                                     destinationName,
                                                     factory );

Drools has the concept of stateful or stateless sessions. We've already covered stateful sessions, which use the standard working memory that can be worked with iteratively over time. Stateless is a one-off execution of a working memory with a provided data set. It may return some results, with the session being disposed at the end, prohibiting further iterative interactions. You can think of stateless as treating a rule engine like a function call with optional return results.

In Drools 4 we supported these two paradigms but the way the user interacted with them was different. StatelessSession used an execute(...) method which would insert a collection of objects as facts. StatefulSession didn't have this method, and insert used the more traditional insert(...) method. The other issue was that the StatelessSession did not return any results, so that users themselves had to map globals to get results, and it wasn't possible to do anything besides inserting objects; users could not start processes or execute queries.

Drools 5.0 addresses all of these issues and more. The foundation for this is the CommandExecutor interface, which both the stateful and stateless interfaces extend, creating consistency and ExecutionResults:



The CommandFactory allows for commands to be executed on those sessions, the only difference being that the Stateless Knowledge Session executes fireAllRules() at the end before disposing the session. The currently supported commands are:

  • FireAllRules

  • GetGlobal

  • SetGlobal

  • InsertObject

  • InsertElements

  • Query

  • StartProcess

  • BatchExecution

InsertObject will insert a single object, with an optional "out" identifier. InsertElements will iterate an Iterable, inserting each of the elements. What this means is that a Stateless Knowledge Session is no longer limited to just inserting objects, it can now start processes or execute queries, and do this in any order.


The execute method always returns an ExecutionResults instance, which allows access to any command results if they specify an out identifier such as the "stilton_id" above.


The execute method only allows for a single command. That's where BatchExecution comes in, which represents a composite command, created from a list of commands. Now, execute will iterate over the list and execute each command in turn. This means you can insert some objects, start a process, call fireAllRules and execute a query, all in a single execute(...) call, which is quite powerful.

As mentioned previosly, the Stateless Knowledge Session will execute fireAllRules() automatically at the end. However the keen-eyed reader probably has already noticed the FireAllRules command and wondered how that works with a StatelessKnowledgeSession. The FireAllRules command is allowed, and using it will disable the automatic execution at the end; think of using it as a sort of manual override function.

Commands support out identifiers. Any command that has an out identifier set on it will add its results to the returned ExecutionResults instance. Let's look at a simple example to see how this works.


In the above example multiple commands are executed, two of which populate the ExecutionResults. The query command defaults to use the same identifier as the query name, but it can also be mapped to a different identifier.

A custom XStream marshaller can be used with the Drools Pipeline to achieve XML scripting, which is perfect for services. Here are two simple XML samples, one for the BatchExecution and one for the ExecutionResults.



The previously mentioned pipeline allows for a series of Stage objects, combined to help with getting data into and out of sessions. There is a Stage implementing the CommandExecutor interface that allows the pipeline to script either a stateful or stateless session. The pipeline setup is trivial:


The key thing here to note is the use of the BatchExecutionHelper to provide a specially configured XStream with custom converters for our Command objects and the new BatchExecutor stage.

Using the pipeline is very simple. You must provide your own implementation of the ResultHandler which is called when the pipeline executes the ExecuteResultHandler stage.




Earlier a BatchExecution was created with Java to insert some objects and execute a query. The XML representation to be used with the pipeline for that example is shown below, with parameters added to the query.


The CommandExecutor returns an ExecutionResults, and this is handled by the pipeline code snippet as well. A similar output for the <batch-execution> XML sample above would be:


The BatchExecutionHelper provides a configured XStream instance to support the marshalling of Batch Executions, where the resulting XML can be used as a message format, as shown above. Configured converters only exist for the commands supported via the Command Factory. The user may add other converters for their user objects. This is very useful for scripting stateless or stateful knowledge sessions, especially when services are involved.

There is currently no XML schema to support schema validation. The basic format is outlined here, and the drools-transformer-xstream module has an illustrative unit test in the XStreamBatchExecutionTest unit test. The root element is <batch-execution> and it can contain zero or more commands elements.


This contains a list of elements that represent commands, the supported commands is limited to those Commands provided by the Command Factory. The most basic of these is the <insert> element, which inserts objects. The contents of the insert element is the user object, as dictated by XStream.


The insert element features an "out-identifier" attribute, demanding that the inserted object will also be returned as part of the result payload.


It's also possible to insert a collection of objects using the <insert-elements> element. This command does not support an out-identifier. The org.domain.UserClass is just an illustrative user object that XStream would serialize.


Next, there is the <set-global> element, which sets a global for the session.


<set-global> also supports two other optional attributes, out and out-identifier. A true value for the boolean out will add the global to the <batch-execution-results> payload, using the name from the identifier attribute. out-identifier works like out but additionally allows you to override the identifier used in the <batch-execution-results> payload.


There is also a <get-global> element, without contents, with just an out-identifier attribute. (There is no need for an out attribute because retrieving the value is the sole purpose of a <get-global> element.


While the out attribute is useful in returning specific instances as a result payload, we often wish to run actual queries. Both parameter and parameterless queries are supported. The name attribute is the name of the query to be called, and the out-identifier is the identifier to be used for the query results in the <execution-results> payload.


The <start-process> command accepts optional parameters. Other process related methods will be added later, like interacting with work items.





Support for more commands will be added over time.

The MarshallerFactory is used to marshal and unmarshal Stateful Knowledge Sessions.


At the simplest the MarshallerFactory can be used as follows:


However, with marshalling you need more flexibility when dealing with referenced user data. To achieve this we have the ObjectMarshallingStrategy interface. Two implementations are provided, but users can implement their own. The two supplied strategies are IdentityMarshallingStrategy and SerializeMarshallingStrategy. SerializeMarshallingStrategy is the default, as used in the example above, and it just calls the Serializable or Externalizable methods on a user instance. IdentityMarshallingStrategy instead creates an integer id for each user object and stores them in a Map, while the id is written to the stream. When unmarshalling it accesses the IdentityMarshallingStrategy map to retrieve the instance. This means that if you use the IdentityMarshallingStrategy, it is stateful for the life of the Marshaller instance and will create ids and keep references to all objects that it attempts to marshal. Below is he code to use an Identity Marshalling Strategy.


For added flexability we can't assume that a single strategy is suitable. Therefore we have added the ObjectMarshallingStrategyAcceptor interface that each Object Marshalling Strategy contains. The Marshaller has a chain of strategies, and when it attempts to read or write a user object it iterates the strategies asking if they accept responsability for marshalling the user object. One of the provided implementations is ClassFilterAcceptor. This allows strings and wild cards to be used to match class names. The default is "*.*", so in the above example the Identity Marshalling Strategy is used which has a default "*.*" acceptor.

Assuming that we want to serialize all classes except for one given package, where we will use identity lookup, we could do the following:


Note that the acceptance checking order is in the natural order of the supplied array.

Longterm out of the box persistence with Java Persistence API (JPA) is possible with Drools. You will need to have some implementation of the Java Transaction API (JTA) installed. For development purposes we recommend the Bitronix Transaction Manager, as it's simple to set up and works embedded, but for production use JBoss Transactions is recommended.


To use a JPA, the Environment must be set with both the EntityManagerFactory and the TransactionManager. If rollback occurs the ksession state is also rolled back, so you can continue to use it after a rollback. To load a previously persisted Stateful Knowledge Session you'll need the id, as shown below:


To enable persistence several classes must be added to your persistence.xml, as in the example below:


The jdbc JTA data source would have to be configured first. Bitronix provides a number of ways of doing this, and its documentation should be contsulted for details. For a quick start, here is the programmatic approach:


Bitronix also provides a simple embedded JNDI service, ideal for testing. To use it add a jndi.properties file to your META-INF and add the following line to it:


Drools has a "native" rule language. This format is very light in terms of punctuation, and supports natural and domain specific languages via "expanders" that allow the language to morph to your problem domain. This chapter is mostly concerted with this native rule format. The diagrams used to present the syntax are known as "railroad" diagrams, and they are basically flow charts for the language terms. The technically very keen may also refer to DRL.g which is the Antlr3 grammar for the rule language. If you use the Rule Workbench, a lot of the rule structure is done for you with content assistance, for example, type "ru" and press ctrl+space, and it will build the rule structure for you.

Drools 5 introduces the concept of hard and soft keywords.

Hard keywords are reserved, you cannot use any hard keyword when naming your domain objects, properties, methods, functions and other elements that are used in the rule text.

Here is the list of hard keywords that must be avoided as identifiers when writing rules:

Soft keywords are just recognized in their context, enabling you to use these words in any other place you wish. Here is a list of the soft keywords:

Of course, you can have these (hard and soft) words as part of a method name in camel case, like notSomething() or accumulateSomething() - there are no issues with that scenario.

Another improvement of the DRL language is the ability to escape hard keywords on rule text. This feature enables you to use your existing domain objects without worrying about keyword collision. To escape a word, simply enclose it in grave accents, like this:

Holiday( `when` == "july" )

The escape should be used everywehere in rule text, except within code expressions in the LHS or RHS code block. Here are examples of proper usage:

rule "validate holiday by eval" 
dialect "mvel"
when
    h1 : Holiday( )
    eval( h1.when == "july" )
then
    System.out.println(h1.name + ":" + h1.when);
end
rule "validate holiday" 
dialect "mvel"
when
    h1 : Holiday( `when` == "july" )
then
    System.out.println(h1.name + ":" + h1.when);
end

Drools 5 introduces standardized error messages. This standardization aims to help users to find and resolve problems in a easier and faster way. In this section you will learn how to identify and interpret those error messages, and you will also receive some tips on how to solve the problems associated with them.

A package is a collection of rules and other related constructs, such as imports and globals. The package members are typically related to each other - perhaps HR rules, for instance. A package represents a namespace, which ideally is kept unique for a given grouping of rules. The package name itself is the namespace, and is not related to files or folders in any way.

It is possible to assemble rules from multiple rule sources, and have one top level package configuration that all the rules are kept under (when the rules are assembled). Although, it is not possible to merge into the same package resources declared under different names. A single Rulebase may, however, contain multiple packages built on it. A common structure is to have all the rules for a package in the same file as the package declaration (so that is it entirely self-contained).

The following railroad diagram shows all the components that may make up a package. Note that a package must have a namespace and be declared using standard Java conventions for package names; i.e., no spaces, unlike rule names which allow spaces. In terms of the order of elements, they can appear in any order in the rule file, with the exception of the package statement, which must be at the top of the file. In all cases, the semicolons are optional.


Notice that any rule atttribute (as described the section Rule Attributes) may also be written at package level, superseding the attribute's default value. The modified default may still be replaced by an attribute setting within a rule.


With global you define global variables. They are used to make application objects available to the rules. Typically, they are used to provide data or services that the rules use, especially application services used in rule consequences, and to return data from the rules, like logs or values added in rule consequences, or for the rules to interact with the application, doing callbacks. Globals are not inserted into the Working Memory, and therefore a global should never be used to establish conditions in rules except when it has a constant immutable value. The engine cannot be notified about value changes of globals and does not track their changes. Incorrect use of globals in constraints may yield surprising results - surprising in a bad way.

If multiple packages declare globals with the same identifier they must be of the same type and all of them will reference the same global value.

In order to use globals you must:

  1. Declare your global variable in your rules file and use it in rules. Example:

    global java.util.List myGlobalList;
    
    rule "Using a global"
    when
        eval( true )
    then
        myGlobalList.add( "Hello World" );
    end
    
  2. Set the global value on your working memory. It is a best practice to set all global values before asserting any fact to the working memory. Example:

    List list = new ArrayList();
    WorkingMemory wm = rulebase.newStatefulSession();
    wm.setGlobal( "myGlobalList", list );
    

Note that these are just named instances of objects that you pass in from your application to the working memory. This means you can pass in any object you want: you could pass in a service locator, or perhaps a service itself. With the new from element it is now common to pass a Hibernate session as a global, to allow from to pull data from a named Hibernate query.

One example may be an instance of a Email service. In your integration code that is calling the rule engine, you obtain your emailService object, and then set it in the working memory. In the DRL, you declare that you have a global of type EmailService, and give it the name "email". Then in your rule consequences, you can use things like email.sendSMS(number, message).

Globals are not designed to share data between rules and they should never be used for that purpose. Rules always reason and react to the working memory state, so if you want to pass data from rule to rule, assert the data as facts into the working memory.

It is strongly discouraged to set or change a global value from inside your rules. We recommend to you always set the value from your application using the working memory interface.



Type declarations have two main goals in the rules engine: to allow the declaration of new types, and to allow the declaration of metadata for types.

  • Declaring new types: Drools works out of the box with plain Java objects as facts. Sometimes, however, users may want to define the model directly to the rules engine, without worrying about creating models in a lower level language like Java. At other times, there is a domain model already built, but eventually the user wants or needs to complement this model with additional entities that are used mainly during the reasoning process.

  • Declaring metadata: facts may have meta information associated to them. Examples of meta information include any kind of data that is not represented by the fact attributes and is consistent among all instances of that fact type. This meta information may be queried at runtime by the engine and used in the reasoning process.

To declare a new type, all you need to do is use the keyword declare, followed by the list of fields, and the keyword end.


The previous example declares a new fact type called Address. This fact type will have three attributes: number, streetName and city. Each attribute has a type that can be any valid Java type, including any other class created by the user or even other fact types previously declared.

For instance, we may want to declare another fact type Person:


As we can see on the previous example, dateOfBirth is of type java.util.Date, from the Java API, while address is of the previously defined fact type Address.

You may avoid having to write the fully qualified name of a class every time you write it by using the import clause, as previously discussed.


When you declare a new fact type, Drools will, at compile time, generate bytecode that implements a Java class representing the fact type. The generated Java class will be a one-to-one Java Bean mapping of the type definition. So, for the previous example, the generated Java class would be:


Since the generated class is a simple Java class, it can be used transparently in the rules, like any other fact.


Declared types are usually used inside rules files, while Java models are used when sharing the model between rules and applications. Although, sometimes, the application may need to access and handle facts from the declared types, especially when the application is wrapping the rules engine and providing higher level, domain specific user interfaces for rules management.

In such cases, the generated classes can be handled as usual with the Java Reflection API, but, as we know, that usually requires a lot of work for small results. Therefore, Drools provides a simplified API for the most common fact handling the application may want to do.

The first important thing to realize is that a declared fact will belong to the package where it was declared. So, for instance, in the example below, Person will belong to the org.drools.examples package, and so the fully qualified name of the generated class will be org.drools.examples.Person.


Declared types, as discussed previously, are generated at knowledge base compilation time, i.e., the application will only have access to them at application run time. Therefore, these classes are not available for direct reference from the application.

Drools then provides an interface through which users can handle declared types from the application code: org.drools.definition.type.FactType. Through this interface, the user can instantiate, read and write fields in the declared fact types.


The API also includes other helpful methods, like setting all the attributes at once, reading values from a Map, or reading all attributes at once, into a Map.

Although the API is similar to Java reflection (yet much simpler to use), it does not use reflection underneath, relying on much more performant accessors implemented in generated bytecode.


A rule specifies that when a particular set of conditions occur, specified in the Left Hand Side (LHS), then do what is specified as a list of actions in the Right Hand Side (RHS). A common question from users is "Why use when instead of if?" "When" was chosen over "if" because "if" is normally part of a procedural execution flow, where, at a specific point in time, a condition is to be checked. In contrast, "when" indicates that the condition evaluation is not tied to a specific evaluation sequence or point in time, but that it happens continually, at any time during the life time of the engine; whenever the condition is met, the actions are executed.

A rule must have a name, unique within its rule package. If you define a rule twice in the same DRL it produces an error while loading. If you add a DRL that includes a rule name already in the package, it replaces the previous rule. If a rule name is to have spaces, then it will need to be enclosd in double quotes (it is best to always use double quotes).

Attributes - described below - are optional. They are best written one per line.

The LHS of the rule follows the when keyword (ideally on a new line), similarly the RHS follows the then keyword (again, ideally on a newline). The rule is terminated by the keyword end. Rules cannot be nested.



Rule attributes provide a declarative way to influence the behavior of the rule. Some are quite simple, while others are part of complex subsystems such as ruleflow. To get the most from Drools you should make sure you have a proper understanding of each attribute.


no-loop

default value: false

type: Boolean

When the rule's consequence modifies a fact it may cause the Rule to activate again, causing recursion. Setting no-loop to true means the attempt to create the Activation for the current set of data will be ignored.

ruleflow-group

default value: N/A

type: String

Ruleflow is a Drools feature that lets you exercise control over the firing of rules. Rules that are assembled by the same ruleflow-group identifier fire only when their group is active.

lock-on-active

default value: false

type: Boolean

Whenever a ruleflow-group becomes active or an agenda-group receives the focus, any rule within that group that has lock-on-active set to true will not be activated any more; irrespective of the origin of the update, the activation of a matching rule is discarded. This is a stronger version of no-loop, because the change could now be caused not only by the rule itself. It's ideal for calculation rules where you have a number of rules that modify a fact and you don't want any rule re-matching and firing again. Only when the ruleflow-group is no longer active or the agenda-group loses the focus those rules with lock-on-active set to true become eligible again for their activations to be placed onto the agenda.

salience

default value : 0

type : integer

Each rule has a salience attribute that can be assigned an integer number, which defaults to zero and can be negative or positive. Salience is a form of priority where rules with higher salience values are given higher priority when ordered in the Activation queue.

agenda-group

default value: MAIN

type: String

Agenda groups allow the user to partition the Agenda providing more execution control. Only rules in the agenda group that has acquired the focus are allowed to fire.

auto-focus

default value: false

type: Boolean

When a rule is activated where the auto-focus value is true and the rule's agenda group does not have focus yet, then it is given focus, allowing the rule to potentially fire.

activation-group

default value: N/A

type: String

Rules that belong to the same activation-group, identified by this attribute's string value, will only fire exclusively. In other words, the first rule in an activation-group to fire will cancel the other rules' activations, i.e., stop them from firing.

Note: This used to be called Xor group, but technically it's not quite an Xor. You may still hear people mention Xor group; just swap that term in your mind with activation-group.

dialect

default value: as specified by the package

type: String

possible values: "java" or "mvel"

The dialect species the language to be used for any code expressions in the LHS or the RHS code block. Currently two dialects are available, Java and MVEL. While the dialect can be specified at the package level, this attribute allows the package definition to be overridden for a rule.

date-effective

default value: N/A

type: String, containing a date and time definition

A rule can only activate if the current date and time is after date-effective attribute.

date-expires

default value: N/A

type: String, containing a date and time definition

A rule cannot activate if the current date and time is after the date-expires attribute.

duration

default value: no default value

type: long

The duration dictates that the rule will fire after a specified duration, if it is still true.


The Left Hand Side (LHS) is a common name for the conditional part of the rule. It consists of zero or more Conditional Elements. If the LHS is left empty, it is re-written as eval(true), which means that the rule's condition is always true. It will be activated, once, when a new Working Memory session is created.



Conditional elements work on one or more patterns (which are described below). The most common one is and, which is implicit when you have multiple patterns in the LHS of a rule that are not connected in any way. Note that an and cannot have a leading declaration binding like or. This is obvious, since a declaration can only reference a single fact, and when the and is satisfied it matches more than one fact - so which fact would the declaration bind to?

The pattern element is the most important Conditional Element. The entity relationship diagram below provides an overview of the various parts that make up the pattern's constraints and how they work together; each is then covered in more detail with railroad diagrams and examples.


At the top of the ER diagram you can see that the pattern consists of zero or more constraints and has an optional pattern binding. The railroad diagram below shows the syntax for this.


In its simplest form, with no constraints, a pattern matches against a fact of the given type. In the following case the type is Cheese, which means that the pattern will match against all Cheese objects in the Working Memory.

Notice that the type need not be the actual class of some fact object. Patterns may refer to superclasses or even interfaces, thereby potentially matching facts from many different classes.


For referring to the matched object, use a pattern binding variable such as $c. The prefixed dollar symbol ('$') is optional; it can be useful in complex rules where it helps to more easily differentiate between variables and fields.


Inside of the pattern parenthesis is where all the action happens. A constraint can be either a Field Constraint, Inline Eval, or a Constraint Group. Constraints can be separated by the following symbols: ',', '&&' or '||'.




The comma character (',') is used to separate constraint groups. It has implicit and connective semantics.


The above example has three constraint groups, each with a single constraint:

The '&&' (and) and '||' (or) constraint connectives allow constraint groups to have multiple constraints. Example:


The above example has two constraint groups. The first has two constraints and the second has one constraint.

The connectives are evaluated in the following order, from first to last:

  1. &&

  2. ||

  3. ,

It is possible to change the evaluation priority by using parentheses, as in any logic or mathematical expression. Example:


In the above example, the use of parentheses evaluates the connective '||' before the connective '&&'.

Also, it is important to note that besides having the same semantics, the connectives '&&' and ',' are resolved with different priorities, and ',' cannot be embedded in a composite constraint expression.


A Field constraint specifies a restriction to be used on a named field; the field name can have an optional variable binding.


There are three types of restrictions: Single Value Restriction, Compound Value Restriction, and Multi Restriction.



A Single Value Restriction is a binary relation, applying a binary operator to the field value and another value, which may be a literal, a variable, a parenthesized expression ("return value"), or a qualified identifier, i.e., an enum constant.


The operators '==' and '!=' are valid for all types. Other relational operatory may be used whenever the type values are ordered; for date fields, '<' means "before". The pair matches and not matches is only applicable to string fields, contains and not contains require the field to be of some Collection type. Coercion to the correct value for the evaluator and the field will be attempted, as mentioned in the "Values" section.

The Conditional Element or is used to group other Conditional Elements into a logical disjunction. Drools supports both prefix or and infix or, but prefix is the preferred option as its implicit grouping avoids confusion. The behavior of the Conditional Element or is different from the connective '||' for constraints and restrictions in field constraints. The engine actually has no understanding of the Conditional Element or; instead, via a number of different logic transformations, a rule with or is rewritten as a number of subrules. This process ultimately results in a rule that has a single or as the root node and one subrule for each of its CEs. Each subrule can activate and fire like any normal rule; there is no special behavior or interaction between these subrules. - This can be most confusing to new rule authors.



Infix or is supported along with explicit grouping with parentheses, should it be needed. The symbol '||', as an alternative to or, is deprecated although it is still supported in the syntax for legacy support reasons.



The Conditional Element or also allows for optional pattern binding. This means that each resulting subrule will bind its pattern to the pattern binding. Each pattern must be bound separately, using eponymous variables:


Since the conditional element or results in multiple subrule generation, one for each possible logically outcome, the example above would result in the internal generation of two rules. These two rules work independently within the Working Memory, which means both can match, activate and fire - there is no shortcutting.

The best way to think of the conditional element or is as a shortcut for generating two or more similar rules. When you think of it that way, it's clear that for a single rule there could be multiple activations if two or more terms of the disjunction are true.


The Conditional Element forall completes the First Order Logic support in Drools. The Conditional Element forall evaluates to true when all facts that match the first pattern match all the remaining patterns. Example:

rule "All English buses are red"
when
    forall( $bus : Bus( type == 'english') 
                   Bus( this == $bus, color = 'red' ) )
then
    # all english buses are red
end

In the above rule, we "select" all Bus objects whose type is "english". Then, for each fact that matches this pattern we evaluate the following patterns and if they match, the forall CE will evaluate to true.

To state that all facts of a given type in the working memory must match a set of constraints, forall can be written with a single pattern for simplicity. Example:


Another example shows multiple patterns inside the forall:


Forall can be nested inside other CEs for complete expressiveness. For instance, forall can be used inside a not CE. Note that only single patterns have optional parentheses, so that with a nested forall parentheses must be used :


As a side note, not( forall( p1 p2 p3...)) is equivalent to writing:

not(p1 and not(and p2 p3...))

Also, it is important to note that forall is a scope delimiter. Therefore, it can use any previously bound variable, but no variable bound inside it will be available for use outside of it.


The Conditional Element from enables users to specify an arbitrary source for data to be matched by LHS patterns. This allows the engine to reason over data not in the Working Memory. The data source could be a sub-field on a bound variable or the results of a method call. It is a powerful construction that allows out of the box integration with other application components and frameworks. One common example is the integration with data retrieved on-demand from databases using hibernate named queries.

The expression used to define the object source is any expression that follows regular MVEL syntax. Therefore, it allows you to easily use object property navigation, execute method calls and access maps and collections elements.

Here is a simple example of reasoning and binding on another pattern sub-field:

rule "validate zipcode"
when
    Person( $personAddress : address ) 
    Address( zipcode == "23920W") from $personAddress 
then
    # zip code is ok
end

With all the flexibility from the new expressiveness in the Drools engine you can slice and dice this problem many ways. This is the same but shows how you can use a graph notation with the 'from':

rule "validate zipcode"
when
    $p : Person( ) 
    $a : Address( zipcode == "23920W") from $p.address 
then
    # zip code is ok
end

Previous examples were evaluations using a single pattern. The CE from also support object sources that return a collection of objects. In that case, from will iterate over all objects in the collection and try to match each of them individually. For instance, if we want a rule that applies 10% discount to each item in an order, we could do:

rule "apply 10% discount to all items over US$ 100,00 in an order"
when
    $order : Order()
    $item  : OrderItem( value > 100 ) from $order.items
then
    # apply discount to $item
end

The above example will cause the rule to fire once for each item whose value is greater than 100 for each given order.

You must take caution, however, when using from, especially in conjunction with the lock-on-active rule attribute as it may produce unexpected results. Consider the example provided earlier, but now slightly modified as follows:

rule "Assign people in North Carolina (NC) to sales region 1"
ruleflow-group "test"
lock-on-active true
when
    $p : Person( ) 
    $a : Address( state == "NC") from $p.address 
then
    modify ($p) {} #Assign person to sales region 1 in a modify block
end

rule "Apply a discount to people in the city of Raleigh"
ruleflow-group "test"
lock-on-active true
when
    $p : Person( ) 
    $a : Address( city == "Raleigh") from $p.address 
then
    modify ($p) {} #Apply discount to person in a modify block
end

In the above example, persons in Raleigh, NC should be assigned to sales region 1 and receive a discount; i.e., you would expect both rules to activate and fire. Instead you will find that only the second rule fires.

If you were to turn on the audit log, you would also see that when the second rule fires, it deactivates the first rule. Since the rule attribute lock-on-active prevents a rule from creating new activations when a set of facts change, the first rule fails to reactivate. Though the set of facts have not changed, the use of from returns a new fact for all intents and purposes each time it is evaluated.

First, it's important to review why you would use the above pattern. You may have many rules across different rule-flow groups. When rules modify working memory and other rules downstream of your RuleFlow (in different rule-flow groups) need to be reevaluated, the use of modify is critical. You don't, however, want other rules in the same rule-flow group to place activations on one another recursively. In this case, the no-loop attribute is ineffective, as it would only prevent a rule from activating itself recursively. Hence, you resort to lock-on-active.

There are several ways to address this issue:

  • Avoid the use of from when you can assert all facts into working memory or use nested object references in your constraint expressions (shown below).

  • Place the variable assigned used in the modify block as the last sentence in your condition (LHS).

  • Avoid the use of lock-on-active when you can explicitly manage how rules within the same rule-flow group place activations on one another (explained below).

The preferred solution is to minimize use of from when you can assert all your facts into working memory directly. In the example above, both the Person and Address instance can be asserted into working memory. In this case, because the graph is fairly simple, an even easier solution is to modify your rules as follows:

rule "Assign people in North Carolina (NC) to sales region 1"
ruleflow-group "test"
lock-on-active true
when
    $p : Person(address.state == "NC" )  
then
    modify ($p) {} #Assign person to sales region 1 in a modify block
end

rule "Apply a discount to people in the city of Raleigh"
ruleflow-group "test"
lock-on-active true
when
    $p : Person(address.city == "Raleigh" )  
then
    modify ($p) {} #Apply discount to person in a modify block
end

Now, you will find that both rules fire as expected. However, it is not always possible to access nested facts as above. Consider an example where a Person holds one or more Addresses and you wish to use an existential quantifier to match people with at least one address that meets certain conditions. In this case, you would have to resort to the use of from to reason over the collection.

There are several ways to use from to achieve this and not all of them exhibit an issue with the use of lock-on-active. For example, the following use of from causes both rules to fire as expected:

rule "Assign people in North Carolina (NC) to sales region 1"
ruleflow-group "test"
lock-on-active true
when
    $p : Person($addresses : addresses)
    exists (Address(state == "NC") from $addresses)  
then
    modify ($p) {} #Assign person to sales region 1 in a modify block
end

rule "Apply a discount to people in the city of Raleigh"
ruleflow-group "test"
lock-on-active true
when
    $p : Person($addresses : addresses)
    exists (Address(city == "Raleigh") from $addresses)  
then
    modify ($p) {} #Apply discount to person in a modify block
end

However, the following slightly different approach does exhibit the problem:

rule "Assign people in North Carolina (NC) to sales region 1"
ruleflow-group "test"
lock-on-active true
when
    $assessment : Assessment()
    $p : Person()
    $addresses : List() from $p.addresses
    exists (Address( state == "NC") from $addresses) 
then
    modify ($assessment) {} #Modify assessment in a modify block
end

rule "Apply a discount to people in the city of Raleigh"
ruleflow-group "test"
lock-on-active true
when
    $assessment : Assessment()
    $p : Person()
    $addresses : List() from $p.addresses 
    exists (Address( city == "Raleigh") from $addresses)
then
    modify ($assessment) {} #Modify assessment in a modify block
end

In the above example, the $addresses variable is returned from the use of from. The example also introduces a new object, assessment, to highlight one possible solution in this case. If the $assessment variable assigned in the condition (LHS) is moved to the last condition in each rule, both rules fire as expected.

Though the above examples demonstrate how to combine the use of from with lock-on-active where no loss of rule activations occurs, they carry the drawback of placing a dependency on the order of conditions on the LHS. In addition, the solutions present greater complexity for the rule author in terms of keeping track of which conditions may create issues.

A better alternative is to assert more facts into working memory. In this case, a person's addresses may be asserted into working memory and the use of from would not be necessary.

There are cases, however, where asserting all data into working memory is not practical and we need to find other solutions. Another option is to reevaluate the need for lock-on-active. An alternative to lock-on-active is to directly manage how rules within the same rule-flow group activate one another by including conditions in each rule that prevent rules from activating each other recursively when working memory is modified. For example, in the case above where a discount is applied to citizens of Raleigh, a condition may be added to the rule that checks whether the discount has already been applied. If so, the rule does not activate.


The Conditional Element collect allows rules to reason over a collection of objects obtained from the given source or from the working memory. In First Oder Logic terms this is the cardinality quantifier. A simple example:

import java.util.ArrayList

rule "Raise priority if system has more than 3 pending alarms"
when
    $system : System()
    $alarms : ArrayList( size >= 3 )
              from collect( Alarm( system == $system, status == 'pending' ) )
then
    # Raise priority, because system $system has
    # 3 or more alarms pending. The pending alarms
    # are $alarms.
end

In the above example, the rule will look for all pending alarms in the working memory for each given system and group them in ArrayLists. If 3 or more alarms are found for a given system, the rule will fire.

The result pattern of collect can be any concrete class that implements the java.util.Collection interface and provides a default no-arg public constructor. This means that you can use Java collections like ArrayList, LinkedList, HashSet, etc., or your own class, as long as it implements the java.util.Collection interface and provide a default no-arg public constructor.

Both source and result patterns can be constrained as any other pattern.

Variables bound before the collect CE are in the scope of both source and result patterns and therefore you can use them to constrain both your source and result patterns. But note that collect is a scope delimiter for bindings, so that any binding made inside of it is not available for use outside of it.

Collect accepts nested from CEs. The following example is a valid use of "collect":

import java.util.LinkedList;

rule "Send a message to all mothers"
when
    $town : Town( name == 'Paris' )
    $mothers : LinkedList() 
               from collect( Person( gender == 'F', children > 0 ) 
                             from $town.getPeople() 
                           )
then
    # send a message to all mothers
end

The Conditional Element accumulate is a more flexible and powerful form of collect, the sense that it can be used to do what collect does and also achieve things that the CE collect is not capable of doing. Basically, what it does is that it allows a rule to iterate over a collection of objects, executing custom actions for each of the elements, and at the end it returns a result object.

The general syntax of the accumulate CE is:

<result pattern> from accumulate( <source pattern>,
                                  init( <init code> ),
                                  action( <action code> ),
                                  reverse( <reverse code> ),
                                  result( <result expression> ) )

The meaning of each of the elements is the following:

  • <source pattern>: the source pattern is a regular pattern that the engine will try to match against each of the source objects.

  • <init code>: this is a semantic block of code in the selected dialect that will be executed once for each tuple, before iterating over the source objects.

  • <action code>: this is a semantic block of code in the selected dialect that will be executed for each of the source objects.

  • <reverse code>: this is an optional semantic block of code in the selected dialect that if present will be executed for each source object that no longer matches the source pattern. The objective of this code block is to undo any calculation done in the <action code> block, so that the engine can do decremental calculation when a source object is modified or retracted, hugely improving performance of these operations.

  • <result expression>: this is a semantic expression in the selected dialect that is executed after all source objects are iterated.

  • <result pattern>: this is a regular pattern that the engine tries to match against the object returned from the <result expression>. If it matches, the accumulate conditional element evaluates to true and the engine proceeds with the evaluation of the next CE in the rule. If it does not matches, the accumulate CE evaluates to false and the engine stops evaluating CEs for that rule.

It is easier to understand if we look at an example:

rule "Apply 10% discount to orders over US$ 100,00"
when
    $order : Order()
    $total : Number( doubleValue > 100 ) 
             from accumulate( OrderItem( order == $order, $value : value ),
                              init( double total = 0; ),
                              action( total += $value; ),
                              reverse( total -= $value; ),
                              result( total ) )
then
    # apply discount to $order
end

In the above example, for each Order in the Working Memory, the engine will execute the init code initializing the total variable to zero. Then it will iterate over all OrderItem objects for that order, executing the action for each one (in the example, it will sum the value of all items into the total variable). After iterating over all OrderItem objects, it will return the value corresponding to the result expression (in the above example, the value of variable total). Finally, the engine will try to match the result with the Number pattern, and if the double value is greater than 100, the rule will fire.

The example used Java as the semantic dialect, and as such, note that the usage of the semicolon as statement delimiter is mandatory in the init, action and reverse code blocks. The result is an expression and, as such, it does not admit ';'. If the user uses any other dialect, he must comply to that dialect's specific syntax.

As mentioned before, the reverse code is optional, but it is strongly recommended that the user writes it in order to benefit from the improved performance on update and retract.

The accumulate CE can be used to execute any action on source objects. The following example instantiates and populates a custom object:

rule "Accumulate using custom objects"
when
    $person   : Person( $likes : likes )
    $cheesery : Cheesery( totalAmount > 100 )
                from accumulate( $cheese : Cheese( type == $likes ),
                                 init( Cheesery cheesery = new Cheesery(); ),
                                 action( cheesery.addCheese( $cheese ); ),
                                 reverse( cheesery.removeCheese( $cheese ); ),
                                 result( cheesery ) );
then
    // do something
end

The accumulate CE is a very powerful CE, but it gets real declarative and easy to use when using predefined functions that are known as Accumulate Functions. They work exactly like accumulate, but instead of explicitly writing custom code in every accumulate CE, the user can use predefined code for common operations.

For instance, the rule to apply discount on orders written in the previous section, could be written in the following way, using Accumulate Functions:

rule "Apply 10% discount to orders over US$ 100,00"
when
    $order : Order()
    $total : Number( doubleValue > 100 ) 
             from accumulate( OrderItem( order == $order, $value : value ),
                              sum( $value ) )
then
    # apply discount to $order
end

In the above example, sum is an Accumulate Function and will sum the $value of all OrderItems and return the result.

Drools ships with the following built-in accumulate functions:

These common functions accept any expression as input. For instance, if someone wants to calculate the average profit on all items of an order, a rule could be written using the average function:

rule "Average profit"
when
    $order : Order()
    $profit : Number() 
              from accumulate( OrderItem( order == $order, $cost : cost, $price : price )
                               average( 1 - $cost / $price ) )
then
    # average profit for $order is $profit
end

Accumulate Functions are all pluggable. That means that if needed, custom, domain specific functions can easily be added to the engine and rules can start to use them without any restrictions. To implement a new Accumulate Functions all one needs to do is to create a Java class that implements the org.drools.base.acumulators.AccumulateFunction interface and add a line to the configuration file or set a system property to let the engine know about the new function. As an example of an Accumulate Function implementation, the following is the implementation of the average function:

/*
 * Copyright 2007 JBoss Inc
 * 
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 * 
 *      http://www.apache.org/licenses/LICENSE-2.0
 * 
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 *
 * Created on Jun 21, 2007
 */
package org.drools.base.accumulators;


/**
 * An implementation of an accumulator capable of calculating average values
 * 
 * @author etirelli
 *
 */
public class AverageAccumulateFunction implements AccumulateFunction {

    protected static class AverageData {
        public int    count = 0;
        public double total = 0;
    }

    /* (non-Javadoc)
     * @see org.drools.base.accumulators.AccumulateFunction#createContext()
     */
    public Object createContext() {
        return new AverageData();
    }

    /* (non-Javadoc)
     * @see org.drools.base.accumulators.AccumulateFunction#init(java.lang.Object)
     */
    public void init(Object context) throws Exception {
        AverageData data = (AverageData) context;
        data.count = 0;
        data.total = 0;
    }

    /* (non-Javadoc)
     * @see org.drools.base.accumulators.AccumulateFunction#accumulate(java.lang.Object,
     * java.lang.Object)
     */
    public void accumulate(Object context,
                           Object value) {
        AverageData data = (AverageData) context;
        data.count++;
        data.total += ((Number) value).doubleValue();
    }

    /* (non-Javadoc)
     * @see org.drools.base.accumulators.AccumulateFunction#reverse(java.lang.Object,
     * java.lang.Object)
     */
    public void reverse(Object context,
                        Object value) throws Exception {
        AverageData data = (AverageData) context;
        data.count--;
        data.total -= ((Number) value).doubleValue();
    }

    /* (non-Javadoc)
     * @see org.drools.base.accumulators.AccumulateFunction#getResult(java.lang.Object)
     */
    public Object getResult(Object context) throws Exception {
        AverageData data = (AverageData) context;
        return new Double( data.count == 0 ? 0 : data.total / data.count );
    }

    /* (non-Javadoc)
     * @see org.drools.base.accumulators.AccumulateFunction#supportsReverse()
     */
    public boolean supportsReverse() {
        return true;
    }

}

The code for the function is very simple, as we could expect, as all the "dirty" integration work is done by the engine. Finally, to plug the function into the engine, we added it to the configuration file:

drools.accumulate.function.average =
   org.drools.base.accumulators.AverageAccumulateFunction

Here, "drools.accumulate.function." is a prefix that must always be used, "average" is how the function will be used in the rule file, and "org.drools.base.accumulators.AverageAccumulateFunction" is the fully qualified name of the class that implements the function behavior.

The Right Hand Side (RHS) is a common name for the consequence or action part of the rule; this part should contain a list of actions to be executed. It is bad practice to use imperative or conditional code in the RHS of a rule; as a rule should be atomic in nature - "when this, then do this", not "when this, maybe do this". The RHS part of a rule should also be kept small, thus keeping it declarative and readable. If you find you need imperative and/or conditional code in the RHS, then maybe you should be breaking that rule down into multiple rules. The main purpose of the RHS is to insert, retractor modify working memory data. To assist with that there are a few convenience methods you can use to modify working memory; without having to first reference a working memory instance.

update(object, handle); will tell the engine that an object has changed (one that has been bound to something on the LHS) and rules may need to be reconsidered.

update(object); can also be used; here the Knowledge Helper will look up the facthandle for you, via an identity check, for the passed object. (Note that if you provide Property Change Listeners to your Java beans that you are inserting into the engine, you can avoid the need to call update() when the object changes.)

insert(new Something()); will place a new object of your creation into the Working Memory.

insertLogical(new Something()); is similar to insert, but the object will be automatically retracted when there are no more facts to support the truth of the currently firing rule.

retract(handle); removes an object from Working Memory.

These convenience methods are basically macros that provide short cuts to the KnowledgeHelper instance that lets you access your Working Memory from rules files. The predefined variable drools of type KnowledgeHelper lets you call several other useful methods. (Refer to the KnowledgeHelper interface documentation for more advanced operations).

The full Knowlege Runtime API is exposed through another predefined variable, kcontext, of type KnowledgeContext. Its method getKnowledgeRuntime() delivers an object of type KnowledgeRuntime, which, in turn, provides access to a wealth of methods, many of which are quite useful for coding RHS logic.


A query is a simple way to search the working memory for facts that match the stated conditions. Therefore, it contains only the structure of the LHS of a rule, so that you specify neither "when" nor "then". A query has an optional set of parameters, each of which can be optionally typed. If the type is not given, the type Object is assumed. The engine will attempt to coerce the values as needed. Query names are global to the KnowledgeBase; so do not add queries of the same name to different packages for the same RuleBase.

To return the results use ksession.getQueryResults("name"), where "name" is the query's name. This returns a list of query results, which allow you to retrieve the objects that matched the query.

The first example presents a simple query for all the people over the age of 30. The second one, using parameters, combines the age limit with a location.



We iterate over the returned QueryResults using a standard "for" loop. Each element is a QueryResultsRow which we can use to access each of the columns in the tuple. These columns can be accessed by bound declaration name or index position.


As mentioned previously, (or DSLs) are a way of extending the rule language to your problem domain. They are wired in to the rule language for you, and can make use of all the underlying rule language and engine features.

DSLs are used both in the IDE, as well as the web based BRMS. Of course as rules are text, you can use them even without this tooling.

A DSL's configuration like most things is stored in plain text. If you use the IDE, you get a nice graphical editor (with some validation), but the format of the file is quite simple, and is basically a properties file.

Note that since Drools 4.0, DSLs have become more powerful in allowing you to customise almost any part of the language, including keywords. Regular expressions can also be used to match words/sentences if needed (this is provided for enhanced localisation). However, not all features are supported by all the tools (although you can use them, the content assistance just may not be 100% accurate in certain cases).


Referring to the above example, the [when] refers to the scope of the expression: ie does it belong on the LHS or the RHS of a rule. The part after the [scope] is the expression that you use in the rule (typically a natural language expression, but it doesn't have to be). The part on the right of the "=" is the mapping into the rule language (of course the form of this depends on if you are talking about the RHS or the LHS - if its the LHS, then its the normal LHS syntax, if its the RHS then its fragments of Java code for instance).

The parser will take the expression you specify, and extract the values that match where the {something} (named Tokens) appear in the input. The values that match the tokens are then interpolated with the corresponding {something} (named Tokens) on the right hand side of the mapping (the target expression that the rule engine actually uses).

Note also that the "sentences" above can be regular expressions. This means the parser will match the sentence fragements that match the expressions. This means you can use (for instance) the '?' to indicate the character before it is optional (think of each sentence as a regular expression pattern - this means if you want to use regex characters - you will need to escape them with a '\' of course.

It is important to note that the DSL expressions are processed one line at a time. This means that in the above example, all the text after "This is " to the end of the line will be included as the value for "{something}" when it is interpolated into the target string. This may not be exactly what you want, as you may want to "chain" together different DSL expressions to generate a target expression. The best way around this is to make sure that the {tokens} are enclosed with characters or words. This means that the parser will scan along the sentence, and pluck out the value BETWEEN the characters (in the example below they are double-quotes). Note that the characters that surround the token are not included in when interpolating, just the contents between them (rather then all the way to the end of the line, as would otherwise be the case).

As a rule of thumb, use quotes for textual data that a rule editor may want to enter. You can also wrap words around the {tokens} to make sure you enclose the data you want to capture (see other example).


It is a good idea to try and avoid punctuation in your DSL expressions where possible, other then quotes and the like - keep it simple and things will be easier. Using a DSL can make debugging slightly harder when you are first building rules, but it can make the maintenance easier (and of course the readability of the rules).

The "{" and "}" characters should only be used on the left hand side of the mapping (the expression) to mark tokens. On the right hand side you can use "{" and "}" on their own if needed - such as

if (foo) \{
    doSomething();\ }

as well as with the token names as shown above.

PLEASE NOTE that if you want curly braces to appear literally as curly braces, then escape them with a backslack (\). Otherwise it may think it is a token to be replaced.

Don't forget that if you are capturing strings from users, you will also need the quotes on the right hand side of the mapping, just like a normal rule, as the result of the mapping must be a valid expression in the rule language.


Referring to the above examples, this would render the following input as shown below:


A good way to get started if you are new to Rules (and DSLs) is just write the rules as you normally would against your object model. You can unit test as you go (like a good agile citizen!). Once you feel comfortable, you can look at extracting a domain language to express what you are doing in the rules. Note that once you have started using the "expander" keyword, you will get errors if the parser does not recognize expressions you have in there - you need to move everything to the DSL. As a way around this, you can prefix each line with ">" and it will tell the parser to take that line literally, and not try and expand it (this is handy also if you are debugging why something isn't working).

Also, it is better to rename the extension of your rules file from ".drl" to ".dslr" when you start using DSLs, as that will allow the IDE to correctly recognize and work with your rules file.

As you work through building up your DSL, you will find that the DSL configuration stabilizes pretty quickly, and that as you add new rules and edit rules you are reusing the same DSL expressions over and over. The aim is to make things as fluent as possible.

To use the DSL when you want to compile and run the rules, you will need to pass the DSL configuration source along with the rule source.

PackageBuilder builder = new PackageBuilder();
builder.addPackageFromDrl( source, dsl );
//source is a reader for the rule source, dsl is a reader for the DSL configuration

You will also need to specify the expander by name in the rule source file:

expander your-expander.dsl

Typically you keep the DSL in the same directory as the rule, but this is not required if you are using the above API (you only need to pass a reader). Otherwise everything is just the same.

You can chain DSL expressions together on one line, as long as it is clear to the parser what the {tokens} are (otherwise you risk reading in too much text until the end of the line). The DSL expressions are processed according to the mapping file, top to bottom in order. You can also have the resulting rule expressions span lines - this means that you can do things like:


Of course this assumes that "Or" is mapped to the "or" conditional element (which is a sensible thing to do).

A common requirement when writing rule conditions is to be able to add many constraints to fact declarations. A fact may have many (dozens) of fields, all of which could be used or not used at various times. To come up with every combination as separate DSL statements would in many cases not be feasible.

The DSL facility allows you to achieve this however, with a simple convention. If your DSL expression starts with a "-", then it will be assumed to be a field constraint, which will be added to the declaration that is above it (one per line).

This is easier to explain with an example. Lets take look at Cheese class, with the following fields: type, price, age, country. We can express some LHS condition in normal DRL like the following

Cheese(age < 5, price == 20, type=="stilton", country=="ch")

If you know ahead of time that you will use all the fields, all the time, it is easy to do a mapping using the above techniques. However, chances are that you will have many fields, and many combinations. If this is the case, you can setup your mappings like so:

[when]There is a Cheese with=Cheese()
[when]- age is less than {age}=age<{age}
[when]- type is '{type}'=type=='{type}'
[when]- country equal to '{country}'=country=='{country}'

IMPORTANT: It is NOT possible to use the "-" feature after an accumulate statement to add constraints to the accumulate pattern. This limitation will be removed in the future.

You can then write rules with conditions like the following:

There is a Cheese with
        - age is less than 42
        - type is 'stilton'

The parser will pick up the "-" lines (they have to be on their own line) and add them as constraints to the declaration above. So in this specific case, using the above mappings, is the equivalent to doing (in DRL):

Cheese(age<42, type=='stilton')

The parser will do all the work for you, meaning you just define mappings for individual constraints, and can combine them how you like (if you are using context assistant, if you press "-" followed by CTRL+space it will conveniently provide you with a filtered list of field constraints to choose from.

To take this further, after alter the DSL to have [when][org.drools.Cheese]- age is less than {age} ... (and similar to all the items in the example above).

The extra [org.drools.Cheese] indicates that the sentence only applies for the main constraint sentence above it (in this case "There is a Cheese with"). For example, if you have a class called "Cheese" - then if you are adding contraints to the rule (by typing "-" and waiting for content assistance) then it will know that only items marked as having an object-scope of "com.yourcompany.Something" are valid, and suggest only them. This is entirely optional (you can leave out that section if needed - OR it can be left blank).

DSLs can be aid with capturing rules if the rules are well known, just not in any technically usable format (ie. sitting around in people brains). Until we are able to have those little sockets in our necks like in the Matrix, our means of getting stuff into computers is still the old fashioned way.

Rules engines require an object or a data model to operate on - in many cases you may know this up front. In other cases the model will be discovered with the rules. In any case, rules generally work better with simpler flatter object models. In some cases, this may mean having a rule object model which is a subset of the main applications model (perhaps mapped from it). Object models can often have complex relationships and hierarchies in them - for rules you will want to simplify and flatten the model where possible, and let the rule engine infer relationships (as it provides future flexibility). As stated previously, DSLs can have an advantage of providing some insulation between the object model and the rule language.

Coming up with a DSL is a collaborative approach for both technical and domain experts. Historically there was a role called "knowledge engineer" which is someone skilled in both the rule technology, and in capturing rules. Over a short period of time, your DSL should stabilize, which means that changes to rules are done entirely using the DSL. A suggested approach if you are starting from scratch is the following workflow:

As an option, Drools also supports a "native" rule language as an alternative to DRL. This allows you to capture and manage your rules as XML data. Just like the non-XML DRL format, the XML format is parsed into the internal "AST" representation - as fast as possible (using a SAX parser). There is no external transformation step required. All the features are available with XML that are available to DRL.

A full W3C standards (XMLSchema) compliant XSD is provided that describes the XML language, which will not be repeated here verbatim. A summary of the language follows.

Example 4.73. A rule in XML

<?xml version="1.0" encoding="UTF-8"?>

<package name="com.sample"
         xmlns="http://drools.org/drools-4.0"
         xmlns:xs="http://www.w3.org/2001/XMLSchema-instance"
         xs:schemaLocation="http://drools.org/drools-4.0 drools-4.0.xsd">

<import name="java.util.HashMap" />
<import name="org.drools.*" />

<global identifier="x" type="com.sample.X" />
<global identifier="yada" type="com.sample.Yada" />

<function return-type="void" name="myFunc">
    <parameter identifier="foo" type="Bar" />
    <parameter identifier="bada" type="Bing" />

    <body>
     System.out.println("hello world");
    </body>
</function>

<rule name="simple_rule">
<rule-attribute name="salience" value="10" />
<rule-attribute name="no-loop" value="true" />
<rule-attribute name="agenda-group" value="agenda-group" />
<rule-attribute name="activation-group" value="activation-group" />

<lhs>
		<pattern identifier="foo2" object-type="Bar" >
            <or-constraint-connective>
                <and-constraint-connective>
                    <field-constraint field-name="a">
                        <or-restriction-connective>
                            <and-restriction-connective>
                                <literal-restriction evaluator=">" value="60" />
                                <literal-restriction evaluator="<" value="70" />
                            </and-restriction-connective>
                            <and-restriction-connective>
                                <literal-restriction evaluator="<" value="50" />
                                <literal-restriction evaluator=">" value="55" />
                            </and-restriction-connective>
                        </or-restriction-connective>
                    </field-constraint>

                    <field-constraint field-name="a3">
                        <literal-restriction evaluator="==" value="black" />
                    </field-constraint>
                </and-constraint-connective>

                <and-constraint-connective>
                    <field-constraint field-name="a">
                        <literal-restriction evaluator="==" value="40" />
                    </field-constraint>

                    <field-constraint field-name="a3">
                        <literal-restriction evaluator="==" value="pink" />
                    </field-constraint>
                </and-constraint-connective>

                <and-constraint-connective>
                    <field-constraint field-name="a">
                        <literal-restriction evaluator="==" value="12"/>
                    </field-constraint>

                    <field-constraint field-name="a3">
                        <or-restriction-connective>
                            <literal-restriction evaluator="==" value="yellow"/>
                            <literal-restriction evaluator="==" value="blue" />
                        </or-restriction-connective>
                    </field-constraint>
                </and-constraint-connective>
            </or-constraint-connective>
        </pattern>

        <not>
            <pattern object-type="Person">
                <field-constraint field-name="likes">
                    <variable-restriction evaluator="==" identifier="type"/>
                </field-constraint>
            </pattern>

            <exists>
                <pattern object-type="Person">
                    <field-constraint field-name="likes">
                        <variable-restriction evaluator="==" identifier="type"/>
                    </field-constraint>
                </pattern>                
            </exists>
        </not>

        <or-conditional-element>
            <pattern identifier="foo3" object-type="Bar" >
                <field-constraint field-name="a">
                    <or-restriction-connective>
                        <literal-restriction evaluator="==" value="3" />
                        <literal-restriction evaluator="==" value="4" />
                    </or-restriction-connective>
                </field-constraint>
                <field-constraint field-name="a3">
                    <literal-restriction evaluator="==" value="hello" />
                </field-constraint>
                <field-constraint field-name="a4">
                    <literal-restriction evaluator="==" value="null" />
                </field-constraint>
            </pattern>

            <pattern identifier="foo4" object-type="Bar" >
                <field-binding field-name="a" identifier="a4" />
                <field-constraint field-name="a">
                    <literal-restriction evaluator="!=" value="4" />
                    <literal-restriction evaluator="!=" value="5" />
                </field-constraint>
            </pattern>
        </or-conditional-element>

        <pattern identifier="foo5" object-type="Bar" >
            <field-constraint field-name="b">
                <or-restriction-connective>
                    <return-value-restriction evaluator="==" >a4 + 1</return-value-restriction>
                    <variable-restriction evaluator=">" identifier="a4" />
                    <qualified-identifier-restriction evaluator="==">
                        org.drools.Bar.BAR_ENUM_VALUE
                    </qualified-identifier-restriction>
                </or-restriction-connective>
            </field-constraint>            
        </pattern>

        <pattern identifier="foo6" object-type="Bar" >
            <field-binding field-name="a" identifier="a4" />
            <field-constraint field-name="b">
                <literal-restriction evaluator="==" value="6" />
            </field-constraint>
        </pattern>
  </lhs>
 <rhs>
    if ( a == b ) {
      assert( foo3 );
    } else {
      retract( foo4 );
    }
    System.out.println( a4 );
   </rhs>
</rule>

</package>
	

In the preceding XML text you will see the typical XML element, the package declaration, imports, globals, functions, and the rule itself. Most of the elements are self explanatory if you have some understanding of the Drools features.

The import elements import the types you wish to use in the rule.

The global elements define global objects that can be referred to in the rules.

The function contains a function declaration, for a function to be used in the rules. You have to specify a return type, a unique name and parameters, in the body goes a snippet of code.

The rule is discussed below.


In the above detail of the rule we see that the rule has LHS and RHS (conditions and consequence) sections. The RHS is simple, it is just a block of semantic code that will be executed when the rule is activated. The LHS is slightly more complicated as it contains nested elements for conditional elements, constraints and restrictions.

A key element of the LHS is the Pattern element. This allows you to specify a type (class) and perhaps bind a variable to an instance of that class. Nested under the pattern object are constraints and restrictions that have to be met. The Predicate and Return Value constraints allow Java expressions to be embedded.

That leaves the conditional elements, not, exists, and, or etc. They work like their DRL counterparts. Elements that are nested under and an "and" element are logically "anded" together. Likewise with "or" (and you can nest things further). "Exists" and "Not" work around patterns, to check for the existence or nonexistence of a fact meeting the pattern's constraints.

The Eval element allows the execution of a valid snippet of Java code - as long as it evaluates to a boolean (do not end it with a semi-colon, as it is just a fragment) - this can include calling a function. The Eval is less efficient than the columns, as the rule engine has to evaluate it each time, but it is a "catch all" feature for when you can express what you need to do with Column constraints.

Decision tables are a "precise yet compact" (ref. Wikipedia) way of representing conditional logic, and are well suited to business level rules.

Drools supports managing rules in a spreadsheet format. Supported formats are Excel (XLS), and CSV, which means that a variety of spreadsheet programs (such as Microsoft Excel, OpenOffice.org Calc amongst others) can be utilized. It is expected that web based decision table editors will be included in a near future release.

Decision tables are an old concept (in software terms) but have proven useful over the years. Very briefly speaking, in Drools decision tables are a way to generate rules driven from the data entered into a spreadsheet. All the usual features of a spreadsheet for data capture and manipulation can be taken advantage of.

The key point to keep in mind is that in a decision table each row is a rule, and each column in that row is either a condition or action for that rule.

The spreadsheet looks for the RuleTable keyword to indicate the start of a rule table (both the starting row and column). Other keywords are also used to define other package level attributes (covered later). It is important to keep the keywords in the one column. By convention the second column ("B") is used for this, but it can be any column (convention is to leave a margin on the left for notes). In the following diagram, C is actually the column where it starts. Everything to the left of this is ignored.

If we expand the hidden sections, it starts to make more sense how it works; note the keywords in column C.

Now the hidden magic which makes it work can be seen. The RuleSet keyword indicates the name to be used in the rule package that will encompass all the rules. This name is optional, using a default, but it must have the RuleSet keyword in the cell immediately to the right.

The other keywords visible in Column C are Import and Sequential which will be covered later. The RuleTable keyword is important as it indicates that a chunk of rules will follow, based on some rule templates. After the RuleTable keyword there is a name, used to prefix the names of the generated rules. The row numbers are appended to guarantee unique rule names. The column of RuleTable indicates the column in which the rules start; columns to the left are ignored.

Referring to row 14 (the row immediately after RuleTable), the keywords CONDITION and ACTION indicate that the data in the columns below are for either the LHS or the RHS parts of a rule. There are other attributes on the rule which can also be optionally set this way.

Row 15 contains declarations of ObjectTypes. The content in this row is optional, but if this option is not in use, the row must be left blank; however this option is usually found to be quite useful. When using this row, the values in the cells below (row 16) become constraints on that object type. In the above case, it will generate Person(age=="42") and Cheese(type=="stilton"), where 42 and "stilton" come from row 18. In the above example, the "==" is implicit; if just a field name is given it will assume that it is to look for exact matches.

Row 16 contains the rule templates themselves. They can use the "$para" place holder to indicate where data from the cells below will be populated ($param can be sued or $1, $2 etc to indicate parameters from a comma separated list in a cell below). Row 17 is ignored as it is textual descriptions of the rule template.

Rows 18 and 19 show data, which will be combined (interpolated) with the templates in row 15, to generate rules. If a cell contains no data, then its template is ignored. (This would mean that some condition or action does not apply for that rule row.) Rule rows are read until there is a blank row. Multiple RuleTables can exsist in a sheet. Row 20 contains another keyword, and a value. The row positions of keywords like this do not matter (most people put them at the top) but their column should be the same one where the RuleTable or RuleSet keywords should appear. In our case column C has been chosen to be significant, but column A could be used instead.

In the above example, rules would be rendered like the following (as it uses the "ObjectType" row):

//row 18
			rule "Cheese_fans_18"
			when
			Person(age=="42")
			Cheese(type=="stilton")
			then
			list.add("Old man stilton");
			end

The syntax of what goes in the templates is dependent on if it is a CONDITION column or ACTION column. In most cases, it is identical to "vanilla" DRL for the LHS or RHS respectively. This means in the LHS, the constraint language must be used, and in the RHS it is a snippet of code to be executed.

The $param place holder is used in templates to indicate where data form the cell will be interpolated. You can also use $1 to the same effect. If the cell contains a comma separated list of values. Symbols $1, $2, etc. may be used to indicate which positional parameter from the list of values in the cell will be used.


For conditions: How snippets are rendered depends on the presence of an entry in the row above, where ObjectType declarations may appear. If there is such an entry, the snippets are rendered as individual constraints on that ObjectType. If there isn't, then they are just rendered as is (with values substituted). If just a plain field is entered (as in the example above) then it will assume that this means equality. If another operator is placed at the end of the snippet, then the values will be interpolated at the end of the constraint, otherwise it will look for $param as outlined previously.

For consequences: How snippets are rendered also depends on the presence of an entry in the row immediately above it. If there is no entry, the output is simply the interpolated snippets. If there is something there (which would typically be a bound variable or a global like in the example above) then it will append it as a method call on that object (refer to the above example).

This may be easiest to understand with some examples, given below.

The above shows how the Person ObjectType declaration spans 2 columns in the spreadsheet, thus both constraints will appear as Person(age == ... , type == ...). As before, since only the field names are present in the snippet, they imply an equality test.

The above condition example shows how you use interpolation to place the values in the snippet (in this case it would result in Person(age == "42")).

The above condition example shows that if you put an operator on the end by itself, the values will be placed after the operator automatically.

A binding can be put in before the column (the constraints will be added from the cells below). Anything can be placed in the ObjectType row. (For instance, this could be a precondition for the columns in the spreadsheet columns that follow).

This shows how the consequence could be done by simple interpolation: just leave the cell above blank. (The same applies to condition columns.) With this style anything can be placed in the consequence, not just one method call.

The following table describes the keywords that are pertinent to the rule table structure.

Table 5.1. Keywords

KeywordDescriptionInclusion Status
RuleSetThe cell to the right of this contains the ruleset name.One only; if left out, it will default.
SequentialThe cell to the right of this can be true or false. If true, then salience is used to ensure that rules fire from the top down.Optional
ImportThe cell to the right contains a comma separated list of Java classes to import.Optional
RuleTableA cell starting with RuleTable indicates the start of a definition of a rule table. The actual rule table starts the next row down. The rule table is read left-to-right, and top-down, until the next blank row.At least one; if there are more, then they are all added to the one ruleset.
CONDITIONIndicates that this column will be for rule conditions.At least one per rule table
ACTIONIndicates that this column will be for rule consequences.At least one per rule table
PRIORITYIndicates that this column's values will set the 'salience' values for the rule row. Over-rides the 'Sequential' flag.Optional
DURATIONIndicates that this column's values will set the duration values for the rule row.Optional
NAMEIndicates that this column's values will set the name for the rule generated from that row.Optional
FunctionsThe cell immediately to the right can contain functions which can be used in the rule snippets. Drools supports functions defined in the DRL, allowing logic to be embedded in the rule, and changed without hard coding, use with care. Same syntax as regular DRL.Optional
VariablesThe cell immediately to the right can contain global declarations which Drools supports. This is a type, followed by a variable name. (If multiple variables are needed, separate them with commas.)Optional
No-loop or UnloopPlaced in the header of a table, no-loop or unloop will both complete the same function of not allowing a rule (row) to loop. For this option to function correctly, there must be a value (true or false) in the cell for the option to take effect. If the cell is left blank then this option will not be set for the row.Optional
XOR-GROUPCell values in this column mean that the rule row belongs to the given Activation group . An Activation group means that only one rule in the named group will fire (i.e., the first one to fire cancels the other rules' activations).Optional
AGENDA-GROUPCell values in this column mean that the rule row belongs to the given Agenda group. (This is one way of controlling flow between groups of rules - see also "rule flow").Optional
RULEFLOW-GROUPCell values in this column mean that the rule row belongs to the given rule-flow group.Optional
WorksheetBy default, the first worksheet is only looked at for decision tables.N/A

Below you will find examples of using the HEADER keyword, which affects the rules generated for each row. Note that the header name is what is important in most cases. If no value appears in the cells below it, then the attribute will not apply (it will be ignored) for that specific row.

The following is an example of Import (comma delimited), Variables (gloabls) - also comma delimited, and a function block (can be multiple functions - just the usual drl syntax). This can appear in the same column as the "RuleSet" keyword, and can be below all the rule rows if you desire.

Related to decision tables (but not necessarily requiring a spreadsheet) are "Rule Templates" (in the drools-templates module). These use any tablular data source as a source of rule data - populating a template to generate many rules. This can allow both for more flexible spreadsheets, but also rules in existing databases, for instance (at the cost of developing the template up front to generate the rules).

With Rule Templates the data is separated from the rule, and there are no restrictions on which part of the rule is data-driven. So whilst you can do everything you could do in decision tables you can also do the following:

As an example, a more classic decision table is shown, but without any hidden rows for the rule meta data (so the spreadsheet only contains the raw data to generate the rules).

See the "ExampleCheese.xls" in the examples download for the above spreadsheet.

If this was a regular decision table there would be hidden rows before row 1 and between rows 1 and 2 containing rule metadata. With rule templates the data is completely separate from the rules. This has two handy consequences - you can apply multiple rule templates to the same data and your data is not tied to your rules at all. So what does the template look like?

1  template header
2  age
3  type
4  log
5
6  package org.drools.examples.templates;
7
8  global java.util.List list;
9
10 template "cheesefans"
11
12 rule "Cheese fans_@{row.rowNumber}"
13 when
14    Person(age == @{age})
15    Cheese(type == "@{type}")
16 then
17    list.add("@{log}");
18 end
19
20 end template
	

Referring to the above:

Line 1: all rule templates start with "template header"
Lines 2-4: following the header is the list of columns in the order they appear in the data. In this case we are calling the first column "age", the second "type" and the third "log".
Lines 5: empty line signifying the end of the column definitions
Lines 6-9: standard rule header text. This is standard rule DRL and will appear at the top of the generated DRL. Put the package statement and any imports and global definitions
Line 10: The "template" keyword signals the start of a rule template. There can be more than one template in a template file. The template should have a unique name.
Lines 11-18: The rule template - see below
Line 20: "end template" signifies the end of the template.

The rule templates rely on MVEL to do substitution using the syntax @{token_name}. There is currently one built-in expression, @{row.rowNumber} which gives a unique number for each row of data and enables you to generate unique rule names. For each row of data a rule will be generated with the values in the data substituted for the tokens in the template. With the example data above the following rule file would be generated:

package org.drools.examples.templates;

global java.util.List list;

rule "Cheese fans_1"
when
  Person(age == 42)
  Cheese(type == "stilton")
then
  list.add("Old man stilton");
end

rule "Cheese fans_2"
when
  Person(age == 21)
  Cheese(type == "cheddar")
then
  list.add("Young man cheddar");
end

The code to run this is simple:

KnowledgeBuilder kbuilder = KnowledgeBuilderFactory.newKnowledgeBuilder();
DecisionTableConfiguration dtconf = KnowledgeBuilderFactory.newDecisionTableConfiguration();
dtconf.setInputType( DecisionTableInputType.XLS );
dtconf.setWorksheetName( "Tables_2" );
kbuilder.add( ResourceFactory.newInputStreamResource( getSpreadsheetStream() ),
              ResourceType.DTABLE,
              dtconf );

Collection<KnowlegePackage> kpkg = kbuilder.getKnowlegePackages();

We create an ExternalSpreadsheetCompiler object and use it to merge the spreadsheet with the rules. The two integer parameters indicate the column and row where the data actually starts - in our case column 2, row 2 (i.e. B2)

If you discover that you have a group of rules following the same arrangement of patterns, constraints and actions on the RHS, differing only in constants or names for objects or fields, you might think of employing Drool's rule template feature for generating the actual rules. You would write a rule template file, containing the textual skeleton of your rule and use the Drools template compiler in combination with a collection of objects providing the actual values for the "flesh" of the rules for their instantiation.

The mechanism is very similar to what a macro processor does. The major advantage proffered by template expansion is that it's nicely integrated in the overall handling of Knowledge Resources.

A rule template file begins with a header defining the placeholders, or formal template parameters for the strings that are to be inserted during instantiation. After the first line, which invariably contains template header, you should write a number of lines, each of which contains a single parameter name.


The template header is followed by the text that is to be replicated and interpolated with the actual parameters. It may begin with a package statement, followed by some additional lines. These may be sectioned into one or more templates, each of them between a pair of matching template and end template statements. The template takes an argument, which puts a name to the template. The name can be a simple unquoted name or an arbitrary string enclosed in double quotes. The template text between these lines may contain one or more rules, constituting the "raw material" for the expansion.


The resulting text will begin with the package line and the header text following it, if present. Then, each template text will be expanded individually, yielding one set of rules for each of the actual parameter sets. Therefore, the structure of the template sections affect the order of the generated rules, since the generator iterates over the sections and then over the set of actual parameters.

Any interpolation takes place between a pair of template and end template statements, when this template is expanded. The template text is scanned for occurrences of parameter expansions written according to:

@{parameter-name}

The name between '@{' and '}' should be one of the parameter names defined in the template header. The substitution is effected anywhere, even within string literals.

An important parameter is available without having to be included in the data source providing the actual values. The parameter substitution

@{row.rowNumber}

expands to the integers 0, 1, 2, etc., providing a unique distinction for the instantiation derived from a parameter set. You would use this as part of each rule name, because, without this precaution, there would be duplicate rule names. (You are, of course, free to use your own identification included as an extra parameter.)

To expand a template, you must prepare a data source. This can be a spreadsheet, as explained in the previous section. Here, we'll concentrate on expansion driven by Java objects. There are two straightforward ways of supplying values for a fixed set of names: Java objects, in the JavaBeans style, and Maps. Both of them can be arranged in a Collection, whose elements will be processed during the expansion, resulting in an instantiation for each element.

You may use a Java object that provides getter methods corresponding to all of the parameter names of your template file. If, for instance, you have defined a header

template header
type
limit
word

the following Java class could be used:

public class ParamSet {
    //...
    public ParamSet( String t, int l, boolean w ) {
        //...
    }
    public String  getType(){...}
    public int     getLimit(){...}
    public boolean isWord(){...}
}

Although interpolation is pure text manipulation, the actual values supplied may be of any type, just as long as this type provides a reasonable toString() method. (For simple types, the eponymous static method of the related class from java.lang is used.)

Assuming that we have created a Collection<ParamSet> for a template file template.drl, we can now proceed to request its expansion.

Collection<ParamSet> paramSets = new ArrayList<ParamSet>();
// populate paramSets
paramSets.add( new ParamSet( "Foo", 42, true ) );
paramSets.add( new ParamSet( "Bar", 13, false ) );
ObjectDataCompiler converter = new ObjectDataCompiler();
InputStream templateStream =
    this.getClass().getResourceAsStream( "template.drl" );
String drl = converter.compile( objs, templateStream );
The resulting string contains the expanded rules text. You could write it to a file and proceed as usual, but it's also possible to feed this to a KnowledgeBuilder and continue with the resulting Knowledge Packages.
KnowledgeBase kBase = KnowledgeBaseFactory.newKnowledgeBase();
KnowledgeBuilder kBuilder = KnowledgeBuilderFactory.newKnowledgeBuilder();
Reader rdr = new StringReader( drl );
kBuilder.add( ResourceFactory.newReaderResource( rdr ), ResourceType.DRL );
if( kBuilder.hasErrors() ){
    // ...
    throw new IllegalStateException( "DRL errors" );
}
kBase.addKnowledgePackages( kBuilder.getKnowledgePackages() );

The following example illustrates template expansion. It is based on simple objects of class Item containing a couple of integer fields and an enum field of type ItemCode.

public class Item {
    // ...
    public Item( String n, int p, int w, ItemCode c ){...}

    public String   getName() {...}
    public int      getWeight() {...}
    public int      getPrice() {...}
    public ItemCode getCode() {...}
}

public enum ItemCode {
    LOCK,
    STOCK,
    BARREL;
}

The rule template contains a single rule. Notice that the field name for the range test is a parameter, which enables us to instantiate the template for different fields.

template header
field
lower
upper
codes

package range;
template "inRange"
rule "is in range @{row.rowNumber}"
when
    Item( $name : name, $v : @{field} >= @{lower} && <= @{upper}, $code : code @{codes} )
then
    System.out.println( "Item " + $name + " @{field} in range: " + $v + " code: " + $code );
end
end template

The next code snippet is from the application, where several parameter sets have to be set up. First, there is class ParamSet, for storing a set of actual parameters.

public class ParamSet {
    //...
    private EnumSet<ItemCode> codeSet;
	
    public ParamSet( String f, int l, int u, EnumSet<ItemCode> cs ){...}

    public String getField() { return field; }
    public int getLower() { return lower; }
    public int getUpper() { return upper; }

    public String getCodes(){
        StringBuilder sb = new StringBuilder();
        String conn = "";
        for( ItemCode ic: codeSet ){
             sb.append( conn ).append( " == ItemCode." ).append( ic );
             conn = " ||";
        }
        return sb.toString();
    }	
}

Note that the method getCodes() does returns the EnumSet<ItemCode> field value as a String value representing a multiple restriction, i.e., a test for one out of a list of values.

The task of expanding a template, passing the resulting DRL text to a Knowledge Builder and adding the resulting Knowledge Packages to a Knowledge Base is generic. The utility class Expander takes care of this, using a Knowledge Base, the InputStream with the rule template and the collection of parameter sets.

public class Expander {
	
    public void expand( KnowledgeBase kBase, InputStream is, Collection<?> act )
        throws Exception {
        ObjectDataCompiler converter = new ObjectDataCompiler();		
        String drl = converter.compile( act, is );
        
        KnowledgeBuilder kBuilder = KnowledgeBuilderFactory.newKnowledgeBuilder();
        Reader rdr = new StringReader( drl );
        kBuilder.add( ResourceFactory.newReaderResource( rdr ), ResourceType.DRL );
        if( kBuilder.hasErrors() ){
            for( KnowledgeBuilderError err: kBuilder.getErrors() ){
                System.err.println( err.toString() );
            }
            throw new IllegalStateException( "DRL errors" );
        }
        kBase.addKnowledgePackages( kBuilder.getKnowledgePackages() );
    }
}

We are now all set to prepare the Knowledge Base with some generated rules. First, we define several parameter sets, constructed as ParamSet objects, and add them to a List, which is passed to the expand method shown above. Then we launch a stateful session, insert a few Item, and watch what happens.

Collection<ParamSet> cfl = new ArrayList<ParamSet>();
cfl.add( new ParamSet( "weight",  10,  99, EnumSet.of( ItemCode.LOCK, ItemCode.STOCK ) ) );
cfl.add( new ParamSet( "price",   10,  50, EnumSet.of( ItemCode.BARREL ) ) );

KnowledgeBase kBase = KnowledgeBaseFactory.newKnowledgeBase();
Expander ex = new Expander();
InputStream dis = new FileInputStream( new File( "rangeTemp.drl" ) );
ex.expand( kBase, dis, cfl );
        
StatefulKnowledgeSession session = kBase.newStatefulKnowledgeSession();
session.insert( new Item( "A", 130,  42, ItemCode.LOCK ) );
session.insert( new Item( "B",  44, 100, ItemCode.STOCK ) );
session.insert( new Item( "C", 123, 180, ItemCode.BARREL ) );
session.insert( new Item( "D",  85,   9, ItemCode.LOCK ) );
        
session.fireAllRules();

Notice that the two resulting rules deal with different fields, one with an item's weight, the other one with its price. - Below is the output.

Item E price in range: 25 code: BARREL
Item A weight in range: 42 code: LOCK

Drools provides an implementation of the Java Rule Engine API (known as JSR94), which allows for support of multiple rule engines from a single API. JSR94 does not deal in any way with the rule language itself. W3C is working on the Rule Interchange Format (RIF) and the OMG has started to work on a standard based on RuleML. Recently Haley Systems has also proposed a rule language standard called RML.

It should be remembered that the JSR94 standard represents the "least common denominator" in features across rule engines. This means that there is less functionality in the JSR94 API than in the standard Drools API. So, by using JSR94, you forfeit the advantage of using the full capabilities of the Drools Rule Engine. It is necessary to expose further functionality, like globals and support for DRL, DSL and XML, via property maps due to the very basic feature set of JSR94; this introduces non-portable functionality. Furthermore, as JSR94 does not provide a rule language, you are only solving a small fraction of the complexity of switching rule engines with very little gain. So, while we support JSR94, for those that insist on using it, we strongly recommend you program against the Drools API.

There are two parts to working with JSR94. The first part is the administrative API that deals with building and registering RuleExecutionSet objects, the second part is runtime session execution of these RuleExecutionSets.

The RuleServiceProviderManager manages the registration and retrieval of RuleServiceProviders. The Drools RuleServiceProvider implementation is automatically registered via a static block when the class is loaded using Class.forNamem, in much the same way as JDBC drivers.


The RuleServiceProvider provides access to the RuleRuntime and RuleAdministrator APIs. The RuleAdministrator provides an administration API for the management of RuleExecutionSet objects, making it possible to register a RuleExecutionSet that can then be retrieved via the RuleRuntime.

First, you need to create a RuleExecutionSet before it can be registered; RuleAdministrator provides factory methods to return an empty LocalRuleExecutionSetProvider or RuleExecutionSetProvider. The LocalRuleExecutionSetProvider should be used to load a RuleExecutionSets from local sources that are not serializable, like Streams. The RuleExecutionSetProvider can be used to load RuleExecutionSets from serializable sources, like DOM Elements or Packages. Both the "ruleAdministrator.getLocalRuleExecutionSetProvider( null );" and the "ruleAdministrator.getRuleExecutionSetProvider( null );" take null as a parameter, as the properties map for these methods is not currently used.


"ruleExecutionSetProvider.createRuleExecutionSet( reader, null )" in the above example takes a null parameter for the properties map; however it can actually be used to provide configuration for the incoming source. When null is passed the default is used to load the input as a drl. Allowed keys for a map are "source" and "dsl". The key "source" takes "drl" or "xml" as its value; you set "source" to "drl" to load a DRL, or to "xml" to load an XML source; "xml" will ignore any "dsl" key/value settings. The "dsl" key can take a Reader or a String (the contents of the dsl) as a value.


When registering a RuleExecutionSet you must specify the name to be used for its retrieval. There is also a field to pass properties, which is currently unused - so just pass null.


The Runtime, obtained from the RuleServiceProvider, is used to create stateful and stateless rule engine sessions.


To create a rule session you must use one of the two RuleRuntime public constants. These are "RuleRuntime.STATEFUL_SESSION_TYPE" and "RuleRuntime.STATELESS_SESSION_TYPE", accompanying the URI to the RuleExecutionSet you wish to instantiate a RuleSession for. The properties map can be null, or it can be used to specify globals, as shown in the next section. The createRuleSession(...) method returns a RuleSession instance which must then be cast to StatefulRuleSession or StatelessRuleSession.


The StatelessRuleSession has a very simple API; you can only call executeRules(List list) passing a list of objects, and an optional filter, the resulting objects are then returned.


If you need more information on JSR 94, please refer to the following references

  1. Official JCP Specification for Java Rule Engine API (JSR 94)

  2. The Java Rule Engine API documentation

  3. The Logic From The Bottom Line: An Introduction to The Drools Project. By N. Alex Rupp, published on TheServiceSide.com in 2004

  4. Getting Started With the Java Rule Engine API (JSR 94): Toward Rule-Based Applications. By Dr. Qusay H. Mahmoud, published on Sun Developer Network in 2005

  5. Jess and the javax.rules API. By Ernest Friedman-Hill, published on TheServerSide.com in 2003

The Eclipse based IDE provides developers (and very technical users) with an environment to edit and test rules in various formats, and integrate it deeply with their applications. In cases where you prefer business rules and web tooling, you will want to look at the BRMS (but using the BRMS and the IDE together is not uncommon).

The Drools IDE is delivered as an Eclipse plug-in, which allows you to author and manage rules from within Eclipse, as well as integrate rules with your application. This is an optional tool, and not all components are required to be used, you can use what components are relevant to you. The Drools IDE is also a part of the Red Hat Developer Studio (formerly known as JBoss IDE).

This guide will cover some of the features of JBoss Drools, in as far as the IDE touches on them (it is assumed that the reader has some familiarity with rule engines, and Drools in particular. It is important to note that none of the underlying features of the rule engine are dependent on Eclipse, and integrators are free to use their tools of choice, as always ! Plenty of people use IntelliJ with rules, for instance.

Note

You can get the plug-in either as a zip to download, or from an update site. Refer to the chapter on installation.


The aim of the new project wizard is to set up an executable scaffold project to start using rules immediately. This will set up a basic structure, the classpath, sample rules and a test case to get you started.


When you choose to create a new "rule project" you will get a choice to add some default artifacts to it, like rules, decision tables, ruleflows, etc. These can serve as a starting point, and will give you something executable almost immediately, which you can then modify and mould to your needs. The simplest case (a hello world rule) is shown below. Feel free to experiment with the plug-in at this point.


The newly created project contains an example rule file (Sample.drl) in the src/rules directory and an example Java file (DroolsTest.java) that can be used to execute the rules in a Drools engine. You'll find this in the folder src/java, in the com.sample package. All the other jars that are necessary during execution are also added to the classpath in a custom classpath container called Drools Library. Rules do not have to be kept in "Java" projects at all, this is just a convenience for people who are already using Eclipse as their Java IDE.

Important note: The Drools plug-in adds a "Drools Builder" capability to your Eclipse instance. This means you can enable a builder on any project that will build and validate your rules when resources change. This happens automatically with the Rule Project Wizard, but you can also enable it manually on any project. One downside of this is that if you have rule files with a large number of rules (more than 500 rules per file), it means that the background builder may be doing a lot of work to build the rules on each change. An option here is to turn off the builder, or put the large rules into .rule files, where you can still use the rule editor, but it won't build them in the background. To fully validate the rules you will need to run them in a unit test of course.

A new feature of the Drools IDE (since version 4) is the guided editor for rules. This is similar to the web based editor that is available in the BRMS. It allows you to build rules in a GUI-driven fashion, based on your object model.


To create a rule this way, use the wizard menu. It will create an instance of a .brl file and open it in the guided editor. This editor works based on a .package file in the same directory as the .brl file. In this "package" file you have the package name and import statements - just like you would at the top of a normal DRL file. First time you create a brl rule you will need to populate the package file with the fact classes you are interested in. Once you have this, the guided editor will be able to prompt you with facts and their fields so that you can build rules graphically.

The guided editor works off the model classes (or fact classes) that you configure. It then is able to "render" to DRL the rule that you have entered graphically. You can do this visually - and use it as a basis for learning DRL, or you can use it and build rules of the brl directly. One way to do this is by using the drools-ant module, which is an ant task that creates all the rule assets in a folder, as a rule package, so that you can then deploy it as a binary file. Alternatively, you can use the following snippet of code to convert the brl to a drl rule.

BRXMLPersitence read = BRXMLPersitence.getInstance();
BRDRLPersistence write = BRDRLPersistence.getInstance();
String brl = ... // read from the .brl file as needed...
String outputDRL = write.marshall(read.unmarshal(brl));
// Pass the outputDRL to the PackageBuilder, as usual

When debugging an application using a Drools engine, these views can be used to check the state of the Drools engine itself: the Working Memory View, the Agenda View, and the Global Data View. To be able to use these views, create breakpoints in your code invoking the working memory. For example, the line where you call workingMemory.fireAllRules() is a good candidate. If the debugger halts at that joinpoint, you should select the working memory variable in the debug variables view. The available views can then be used to show the details of the selected working memory:

The Audit view can be used to display audit logs containing events that were logged during the execution of a rules engine, in tree form.

The audit view visualizes an audit log, that is optionally created when executing the rules engine. To create an audit log, use the following code:

WorkingMemory workingMemory = ruleBase.newWorkingMemory();
// Create a new Working Memory Logger, that logs to file.
WorkingMemoryFileLogger logger = new WorkingMemoryFileLogger(workingMemory);
// An event.log file is created in the subdirectory log (which must exist)
// of the working directory.
logger.setFileName( "log/event" );

workingMemory.assertObject(...);
workingMemory.fireAllRules();

// stop logging
logger.writeToDisk();

Open the log by clicking the Open Log action, the first icon in the Audit View, and select the file. The Audit View now shows all events that where logged during the executing of the rules. There are different types of events, each with a different icon:

All these events show extra information concerning the event, like the id and toString representation of the object in case of working memory events (insert, modify and retract), the name of the rule and all the variables bound in the activation in case of an activation event (created, cancelled or executed). If an event occurs when executing an activation, it is shown as a child of the activation's execution event. For some events, you can retrieve the "cause":

When selecting an event, the cause of that event is shown in green in the audit view (if visible of course). You can also right click the action and select the "Show Cause" menu item. This will scroll you to the cause of the selected event.

Domain Specific Languages (DSL) enable you to create a language that allows your rules to look like - rules! Most often the domain specific language reads like natural language. Typically you would look at how a business analyst would describe the rule, in their own words, and then map this to your object model, via rule constructs. A side benefit of this is that it can provide an insulation layer between your domain objects and the rules themselves (as we know you like to refactor). A domain specific language will grow as the rules grow, and works best when there are common terms used over an over, with different parameters.

To aid with this, the rule workbench provides an editor for Domain Specific Languages. They are stored in a plain text format, so you can use any editor of your choice; this format is simply a slightly enhanced version of the "Properties" file format. The editor will be invoked on any files with a .dsl extension. There is also a wizard to create a sample DSL.


The DSL editor provides a tabular view of the mapping of Language to Rule Expressions. The Language Expression is what is used in the rules. This also feeds the content assistance for the rule editor, so that it can suggest Language Expressions from the DSL configuration. (The rule editor loads the DSL configuration when the rule resource is loaded for editing.) The Rule language mapping defines the "code" for the rules into which the language expression will be compiled by the rule engine compiler. The form of this Rule language expression depends on it being intended for the condition or the action part of a rule. (For the RHS it may be a snippet of Java, for instance). The "scope" item indicates where the expression belongs, "when" indicating the LHS, "then" the RHS, and "*" meaning anywhere. It's also possible to create aliases for keywords.

By selecting a mapping item (a row in the table) you can see the expression and mapping in the text fields below the table. Double clicking or pressing the edit button will open the edit dialog. Other buttons let you remove and add mappings. Don't remove mappings while they are still in use.


Here is a short description of the DSL translation process. The parser reads the rule text in a DSL, line by line, and tries to match some "Language Expression", depending on the scope. After a match, the values that correspond to a placeholder between curly braces (e.g., {age}) are extracted from the rule source. The placeholders in the corresponding "Rule Expression" are replaced by their corresponding value. In the example above, the natural language expression maps to two constraints on a fact of type Person, based on the fields age and location, and the {age} and {location} values are extracted from the original rule text.

If you do not wish to use a language mapping for a particular rule in a drl, prefix the expression with > and the compiler will not try to translate it according to the language definition. Also note that Domain Specific Languages are optional. When the rule is compiled, the .dsl file will also need to be available.


You can debug rules during the execution of your Drools application. You can add breakpoints in the consequences of your rules, and whenever such a breakpoint is encountered during the execution of the rules, execution is halted. You can then inspect the variables known at that point and use any of the default debugging actions to decide what should happen next: step over, continue, etc. You can also use the debug views to inspect the content of the working memory and the Agenda.

Drools breakpoints are only enabled if you debug your application as a Drools Application. You can do this like this:


  1. Select the main class of your application. Right click it and select the "Debug As >" sub-menu and select Drools Application. Alternatively, you can also select the "Debug ..." menu item to open a new dialog for creating, managing and running debug configurations (see the screenshot below).

  2. Select the "Drools Application" item in the left tree and click the "New launch configuration" button (leftmost icon in the toolbar above the tree). This will create a new configuration with some of the properties (like project and main class)already filled in, based on the main class you selected in the beginning. All properties shown here are the same as for any standard Java program.

  3. Change the name of your debug configuration to something meaningful. You can just accept the defaults for all other properties. For more information about these properties, please check the Eclipse JDT documentation.

  4. Click the "Debug" button on the bottom to start debugging your application. You only have to define your debug configuration once. The next time you run your Drools application, you don't have to create a new one but select the previously defined one in the tree on the left, as a sub-element of the "Drools Application" tree node, and then click the Debug button. The Eclipse toolbar also contains shortcut buttons to quickly re-execute one of your previous configurations (at least when one of the Java, Java Debug, or Drools perspectives has been selected).


After clicking the "Debug" button, the application starts executing and will halt if any breakpoint is encountered. This can be a Drools rule breakpoint, or any other standard Java breakpoint. Whenever a Drools rule breakpoint is encountered, the corresponding DRL file is opened and the active line is highlighted. The Variables view also contains all rule parameters and their value. You can then use the default Java debug actions to decide what to do next: resume, terminate, step over, etc. The debug view can also be used to inspect the contents of the Working Memory and the Agenda at that time as well. You don't have to select a Working Memory now, as the current executing working memory is automatically shown.


Name: Hello World
Main class: org.drools.examples.HelloWorldExample
Type: Java application
Rules file: HelloWorld.drl
Objective: demonstrate basic rules in use

The "Hello World" example shows a simple example of rules usage, and both the MVEL and Java dialects.

This example demonstrates how to build Knowledge Bases and Sessions. Also, audit logging and debug outputs are shown, which is ommitted from other examples as it's all very similar. A KnowledgeBuilder is used to turn a DRL source file into Package objects which the Knowledge Base can consume. The add method takes a Resource interface and a Resource Type as parameters. The Resource can be used to retrieve a DRL source file from various locations; in this case the DRL file is being retrieved from the classpath using a ResourceFactory, but it could come from a disk file or a URL. Here, we only add a single DRL source file, but multiple DRL files can be added. Also, DRL files with different namespaces can be added, where the Knowledge Builder creates a package for each namespace. Multiple packages of different namespaces can be added to the same Knowledge Base. When all the DRL files have been added, we should check the builder for errors. While the Knowledge Base will validate the package, it will only have access to the error information as a String, so if you wish to debug the error information you should do it on the KnowledgeBuilder instance. Once you know the builder is error free, get the Package collection, instantiate a KnowledgeBase from the KnowledgeBaseFactory and add the package collection.


Drools has an event model that exposes much of what's happening internally. Two default debug listeners are supplied, DebugAgendaEventListener and DebugWorkingMemoryEventListener which print out debug event information to the System.err stream displayed in the Console window. Adding listeners to a Session is trivial, as shown below. The KnowledgeRuntimeLogger provides execution auditing, the result of which can be viewed in a graphical viewer. The logger is actually a specialised implementation built on the Agenda and Working Memory listeners. When the engine has finished executing, logger.close() must be called.

Most of the examples use the Audit logging features of Drools to record execution flow for later inspection.


The single class used in this example is very simple. It has two fields: the message, which is a String and the status which can be one of the two integers HELLO or GOODBYE.


A single Message object is created with the message text "Hello World" and the status HELLO and then inserted into the engine, at which point fireAllRules() is executed. Remember that all the network evaluation is done during the insert time, so that by the time the program execution reaches the fireAllRules() method call the engine already knows which rules are fully matches and able to fire.


To execute the example as a Java application:

  1. Open the class org.drools.examples.HelloWorldExample in your Eclipse IDE

  2. Right-click the class and select "Run as..." and then "Java application"

If we put a breakpoint on the fireAllRules() method and select the ksession variable, we can see that the "Hello World" rule is already activated and on the Agenda, confirming that all the pattern matching work was already done during the insert.


The may application print outs go to to System.out while the debug listener print outs go to System.err.



The LHS (after when) section of the rule states that it will be activated for each Message object inserted into the Working Memory whose status is Message.HELLO. Besides that, two variable bindings are created: the variable message is bound to the message attribute and the variable m is bound to the matched Message object itself.

The RHS (after then) or consequence part of the rule is written using the MVEL expression language, as declared by the rule's attribute dialect. After printing the content of the bound variable message to System.out, the rule changes the values of the message and status attributes of the Message object bound to m. This is done MVEL's modify statement, which allows you to apply a block of assignments in one statement, with the engine being automatically notified of the changes at the end of the block.


We can set a breakpoint into the DRL, on the modify call, and inspect the Agenda view again during the execution of the rule's consequence. This time we start the execution via "Debug As" and "Drools application" and not by running a "Java application":

  1. Open the class org.drools.examples.HelloWorld in your Eclipse IDE.

  2. Right-click the class and select "Debug as..." and then "Drools application".

Now we can see that the other rule "Good Bye", which uses the Java dialect, is activated and placed on the Agenda.


The "Good Bye" rule, which specifies the "java" dialect, is similar to the "Hello World" rule except that it matches Message objects whose status is Message.GOODBYE.


Remember the Java code where we used the KnowledgeRuntimeLoggerFactory method newFileLogger to create a KnowledgeRuntimeLogger and called logger.close() at the end. This created an audit log file that can be shown in the Audit view. We use the Audit view in many of the examples to demostrate the example execution flow. In the view screen shot below we can see that the object is inserted, which creates an activation for the "Hello World" rule; the activation is then executed which updates the Message object causing the "Good Bye" rule to activate; finally the "Good Bye" rule also executes. Selecting an event in the Audit view highlights the origin event in green; therefore the "Activation created" event is highlighted in green as the origin of the "Activation executed" event.


This example is implemented in three different versions to demonstrate different ways of implementing the same basic behavior: forward chaining, i.e., the ability the engine has to evaluate, activate and fire rules in sequence, based on changes on the facts in the Working Memory.

Name: State Example
Main class: org.drools.examples.StateExampleUsingSalience
Type: Java application
Rules file: StateExampleUsingSalience.drl
Objective: Demonstrates basic rule use
  and Conflict Resolution for rule firing priority.

Each State class has fields for its name and its current state (see the class org.drools.examples.State). The two possible states for each objects are:


Ignoring the PropertyChangeSupport, which will be explained later, we see the creation of four State objects named A, B, C and D. Initially their states are set to NOTRUN, which is default for the used constructor. Each instance is asserted in turn into the Session and then fireAllRules() is called.


To execute the application:

  1. Open the class org.drools.examples.StateExampleUsingSalience in your Eclipse IDE.

  2. Right-click the class and select "Run as..." and then "Java application"

You will see the following output in the Eclipse console window:


There are four rules in total. First, the Bootstrap rule fires, setting A to state FINISHED, which then causes B to change its state to FINISHED. C and D are both dependent on B, causing a conflict which is resolved by the salience values. Let's look at the way this was executed.

The best way to understand what is happening is to use the Audit Logging feature to graphically see the results of each operation. To view the Audit log generated by a run of this example:

  1. If the Audit View is not visible, click on "Window" and then select "Show View", then "Other..." and "Drools" and finally "Audit View".

  2. In the "Audit View" click the "Open Log" button and select the file "<drools-examples-drl-dir>/log/state.log".

After that, the "Audit view" will look like the following screenshot:


Reading the log in the "Audit View", top to bottom, we see every action and the corresponding changes in the Working Memory. This way we observe that the assertion of the State object A in the state NOTRUN activates the Bootstrap rule, while the assertions of the other State objects have no immediate effect.


The execution of rule Bootstrap changes the state of A to FINISHED, which, in turn, activates rule "A to B".


The execution of rule "A to B" changes the state of B to FINISHED, which activates both, rules "B to C" and "B to D", placing their Activations onto the Agenda. From this moment on, both rules may fire and, therefore, they are said to be "in conflict". The conflict resolution strategy allows the engine's Agenda to decide which rule to fire. As rule "B to C" has the higher salience value (10 versus the default salience value of 0), it fires first, modifying object C to state FINISHED. The Audit view shown above reflects the modification of the State object in the rule "A to B", which results in two activations being in conflict. The Agenda view can also be used to investigate the state of the Agenda, with debug points being placed in the rules themselves and the Agenda view opened. The screen shot below shows the breakpoint in the rule "A to B" and the state of the Agenda with the two conflicting rules.



Rule "B to D" fires last, modifying object D to state FINISHED.


There are no more rules to execute and so the engine stops.

Another notable concept in this example is the use of dynamic facts, based on PropertyChangeListener objects. As described in the documentation, in order for the engine to see and react to changes of fact properties, the application must tell the engine that changes occurred. This can be done explicitly in the rules by using the modify statement, or implicitly by letting the engine know that the facts implement PropertyChangeSupport as defined by the JavaBeans specification. This example demonstrates how to use PropertyChangeSupport to avoid the need for explicit modify statements in the rules. To make use of this feature, ensure that your facts implement PropertyChangeSupport, the same way the class org.drools.example.State does, and use the following code to insert the facts into the Working Memory:


When using PropertyChangeListener objects, each setter must implement a little extra code for the notification. Here is the setter for state in the class org.drools.examples:

:

There are two other classes in this example: StateExampleUsingAgendGroup and StateExampleWithDynamicRules. Both execute from A to B to C to D, as just shown. The StateExampleUsingAgendGroup uses agenda-groups to control the rule conflict and which one fires first. StateExampleWithDynamicRules shows how an additional rule can be added to an already running Working Memory with all the existing data applying to it at runtime.

Agenda groups are a way to partition the Agenda into groups and to control which groups can execute. By default, all rules are in the agenda group "MAIN". The "agenda-group" attribute lets you specify a different agenda group for the rule. Initially, a Working Memory has its focus on the Agenda group "MAIN". A group's rules will only fire when the group receives the focus. This can be achieved either ny using the method by setFocus() or the rule attribute auto-focus. "auto-focus" means that the rule automatically sets the focus to its agenda group when the rule is matched and activated. It is this "auto-focus" that enables rule "B to C" to fire before "B to D".


The rule "B to C" calls setFocus() on the agenda group "B to D", allowing its active rules to fire, which allows the rule "B to D" to fire.


The example StateExampleWithDynamicRules adds another rule to the Rule Base after fireAllRules(). The added rule is just another state transition.


This produces the following expected output:


Name: Fibonacci 
Main class: org.drools.examples.FibonacciExample
Type: Java application
Rules file: Fibonacci.drl
Objective: Demonstrates Recursion,
  the CE not and cross product matching

The Fibonacci Numbers (see http://en.wikipedia.org/wiki/Fibonacci_number) discovered by Leonardo of Pisa (see http://en.wikipedia.org/wiki/Fibonacci) is a sequence that starts with 0 and 1. The next Fibonacci number is obtained by adding the two preceding Fibonacci numbers. The Fibonacci sequence begins with 0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233, 377, 610, 987, 1597, 2584, 4181, 6765, 10946,... The Fibonacci Example demonstrates recursion and conflict resolution with salience values.

The single fact class Fibonacci is used in this example. It has two fields, sequence and value. The sequence field is used to indicate the position of the object in the Fibonacci number sequence. The value field shows the value of that Fibonacci object for that sequence position, using -1 to indicate a value that still needs to be computed.


Execute the example:

  1. Open the class org.drools.examples.FibonacciExample in your Eclipse IDE.

  2. Right-click the class and select "Run as..." and then "Java application"

Eclipse shows the following output in its console window (with "...snip..." indicating lines that were removed to save space):


To kick this off from Java we only insert a single Fibonacci object, with a sequence field of 50. A recursive rule is then used to insert the other 49 Fibonacci objects. This example doesn't use PropertyChangeSupport. It uses the MVEL dialect, which means we can use the modify keyword, which allows a block setter action which also notifies the engine of changes.


The rule Recurse is very simple. It matches each asserted Fibonacci object with a value of -1, creating and asserting a new Fibonacci object with a sequence of one less than the currently matched object. Each time a Fibonacci object is added while the one with a sequence field equal to 1 does not exist, the rule re-matches and fires again. The not conditional element is used to stop the rule's matching once we have all 50 Fibonacci objects in memory. The rule also has a salience value, because we need to have all 50 Fibonacci objects asserted before we execute the Bootstrap rule.


The Audit view shows the original assertion of the Fibonacci object with a sequence field of 50, done from Java code. From there on, the Audit view shows the continual recursion of the rule, where each asserted Fibonacci object causes the Recurse rule to become activated and to fire again.


When a Fibonacci object with a sequence field of 2 is asserted the "Bootstrap" rule is matched and activated along with the "Recurse" rule. Note the multi-restriction on field sequence, testing for equality with 1 or 2.


At this point the Agenda looks as shown below. However, the "Bootstrap" rule does not fire because the "Recurse" rule has a higher salience.


When a Fibonacci object with a sequence of 1 is asserted the Bootstrap rule is matched again, causing two activations for this rule. Note that the "Recurse" rule does not match and activate because the not conditional element stops the rule's matching as soon as a Fibonacci object with a sequence of 1 exists.


Once we have two Fibonacci objects with values not equal to -1 the "Calculate" rule is able to match. It was the "Bootstrap" rule that set the objects with sequence 1 and 2 to values of 1. At this point we have 50 Fibonacci objects in the Working Memory. Now we need to select a suitable triple to calculate each of their values in turn. Using three Fibonacci patterns in a rule without field constraints to confine the possible cross products would result in 50x49x48 possible combinations, leading to about 125,000 possible rule firings, most of them incorrect. The "Calculate" rule uses field constraints to correctly constraint the thee Fibonacci patterns in the correct order; this technique is called cross product matching. The first pattern finds any Fibonacci with a value != -1 and binds both the pattern and the field. The second Fibonacci does this, too, but it adds an additional field constraint to ensure that its sequence is greater by one than the Fibonacci bound to f1. When this rule fires for the first time, we know that only sequences 1 and 2 have values of 1, and the two constraints ensure that f1 references sequence 1 and f2 references sequence 2. The final pattern finds the Fibonacci with a value equal to -1 and with a sequence one greater than f2. At this point, we have three Fibonacci objects correctly selected from the available cross products, and we can calculate the value for the third Fibonacci object that's bound to f3.


The modify statement updated the value of the Fibonacci object bound to f3. This means we now have another new Fibonacci object with a value not equal to -1, which allows the "Calculate" rule to rematch and calculate the next Fibonacci number. The Audit view below shows how the firing of the last "Bootstrap" modifies the Fibonacci object, enabling the "Calculate" rule to match, which then modifies another Fibonacci object allowing the "Calculate" rule to match again. This continues till the value is set for all Fibonacci objects.


Name: BankingTutorial
Main class: org.drools.tutorials.banking.Example*.java
Type: Java application
Rules file: org.drools.tutorials.banking.*.drl
Objective: Demonstrate pattern matching, basic sorting and calculation rules.

This tutorial demonstrates the process of developing a complete personal banking application to handle credits and debits on multiple accounts. It uses a set of design patterns that have been created for the process.

The class RuleRunner is a simple harness to execute one or more DRL files against a set of data. It compiles the Packages and creates the Knowledge Base for each execution, allowing us to easily execute each scenario and inspect the outputs. In reality this is not a good solution for a production system, where the Knowledge Base should be built just once and cached, but for the purposes of this tutorial it shall suffice.


The first of our sample Java classes loads and executes a single DRL file, Example.drl, but without inserting any data.


The first simple rule to execute has a single eval condition that will alway be true, so that this rule will match and fire, once, after the start.


The output for the rule is below, showing that the rule matches and executes the single print statement.


The next step is to assert some simple facts and print them out.


This doesn’t use any specific facts but instead asserts a set of java.lang.Integer objects. This is not considered "best practice" as a number is not a useful fact, but we use it here to demonstrate basic techniques before more complexity is added.

Now we will create a simple rule to print out these numbers.


Once again, this rule does nothing special. It identifies any facts that are Number objects and prints out the values. Notice the use of the abstract class Number: we inserted Integer objects but we now look for any kind of number. The pattern matching engine is able to match interfaces and superclasses of asserted objects.

The output shows the DRL being loaded, the facts inserted and then the matched and fired rules. We can see that each inserted number is matched and fired and thus printed.


There are certainly many better ways to sort numbers than using rules, but since we will need to apply some cashflows in date order when we start looking at banking rules we'll develop simple rule based sorting technique.


Again we insert our Integer objects, but this time the rule is slightly different:


The first line of the rule identifies a Number and extracts the value. The second line ensures that there does not exist a smaller number than the one found by the first pattern. We might expect to match only one number - the smallest in the set. However, the retraction of the number after it has been printed means that the smallest number has been removed, revealing the next smallest number, and so on.

The resulting output shows that the numbers are now sorted numerically.


We are ready to start moving towards our personal accounting rules. The first step is to create a Cashflow object.


Class Cashflow has two simple attributes, a date and an amount. (Note that using the type double for monetary units is generally not a good idea because floating point numbers cannot represent most numbers accurately.) There is also an overloaded constructor to set the values, and a method toString to print a cashflow. The Java code of Example4.java inserts five Cashflow objects, with varying dates and amounts.


The convenience class SimpleDate extends java.util.Date, providing a constructor taking a String as input and defining a date format. The code is listed below


Now, let’s look at Example4.drl to see how we print the sorted Cashflow objects:


Here, we identify a Cashflow and extract the date and the amount. In the second line of the rule we ensure that there is no Cashflow with an earlier date than the one found. In the consequence, we print the Cashflow that satisfies the rule and then retract it, making way for the next earliest Cashflow. So, the output we generate is:


Next, we extend our Cashflow, resulting in a TypedCashflow which can be a credit or a debit operation. (Normally, we would just add this to the Cashflow type, but we use extension to keep the previous version of the class intact.)


There are lots of ways to improve this code, but for the sake of the example this will do.

Now let's create Example5, a class for running our code.


Here, we simply create a set of Cashflow objects which are either credit or debit operations. We supply them and Example5.drl to the RuleEngine.

Now, let’s look at a rule printing the sorted Cashflow objects.


Here, we identify a Cashflow fact with a type of CREDIT and extract the date and the amount. In the second line of the rule we ensure that there is no Cashflow of the same type with an earlier date than the one found. In the consequence, we print the cashflow satisfying the patterns and then retract it, making way for the next earliest cashflow of type CREDIT.

So, the output we generate is


Continuing our banking exercise, we are now going to process both credits and debits on two bank accounts, calculating the account balance. In order to do this, we create two separate Account objects and inject them into the Cashflows objects before passing them to the Rule Engine. The reason for this is to provide easy access to the correct account without having to resort to helper classes. Let’s take a look at the Account class first. This is a simple Java object with an account number and balance:


Now let’s extend our TypedCashflow, resulting in AllocatedCashflow, to include an Account reference.


The Java code of Example5.java creates two Account objects and passes one of them into each cashflow, in the constructor call.


Now, let’s look at the rule in Example6.drl to see how we apply each cashflow in date order and calculate and print the balance.


Although we have separate rules for credits and debits, but we do not specify a type when checking for earlier cashflows. This is so that all cashflows are applied in date order, regardless of the cashflow type. In the conditions we identify the account to work with, and in the consequences we update it with the cashflow amount.

Example 8.51. Banking Tutorial: Output of Example6.java

Loading file: Example6.drl
Inserting fact: AllocatedCashflow[account=Account[accountNo=1,balance=0.0],date=Mon Jan 01 00:00:00 GMT 2007,type=Credit,amount=300.0]
Inserting fact: AllocatedCashflow[account=Account[accountNo=1,balance=0.0],date=Mon Feb 05 00:00:00 GMT 2007,type=Credit,amount=100.0]
Inserting fact: AllocatedCashflow[account=Account[accountNo=2,balance=0.0],date=Sun Mar 11 00:00:00 GMT 2007,type=Credit,amount=500.0]
Inserting fact: AllocatedCashflow[account=Account[accountNo=1,balance=0.0],date=Wed Feb 07 00:00:00 GMT 2007,type=Debit,amount=800.0]
Inserting fact: AllocatedCashflow[account=Account[accountNo=2,balance=0.0],date=Fri Mar 02 00:00:00 GMT 2007,type=Debit,amount=400.0]
Inserting fact: AllocatedCashflow[account=Account[accountNo=1,balance=0.0],date=Sun Apr 01 00:00:00 BST 2007,type=Credit,amount=200.0]
Inserting fact: AllocatedCashflow[account=Account[accountNo=1,balance=0.0],date=Thu Apr 05 00:00:00 BST 2007,type=Credit,amount=300.0]
Inserting fact: AllocatedCashflow[account=Account[accountNo=2,balance=0.0],date=Fri May 11 00:00:00 BST 2007,type=Credit,amount=700.0]
Inserting fact: AllocatedCashflow[account=Account[accountNo=1,balance=0.0],date=Mon May 07 00:00:00 BST 2007,type=Debit,amount=900.0]
Inserting fact: AllocatedCashflow[account=Account[accountNo=2,balance=0.0],date=Wed May 02 00:00:00 BST 2007,type=Debit,amount=100.0]
Debit: Fri Mar 02 00:00:00 GMT 2007 :: 400.0
Account: 2 - new balance: -400.0
Credit: Sun Mar 11 00:00:00 GMT 2007 :: 500.0
Account: 2 - new balance: 100.0
Debit: Wed May 02 00:00:00 BST 2007 :: 100.0
Account: 2 - new balance: 0.0
Credit: Fri May 11 00:00:00 BST 2007 :: 700.0
Account: 2 - new balance: 700.0
Credit: Mon Jan 01 00:00:00 GMT 2007 :: 300.0
Account: 1 - new balance: 300.0
Credit: Mon Feb 05 00:00:00 GMT 2007 :: 100.0
Account: 1 - new balance: 400.0
Debit: Wed Feb 07 00:00:00 GMT 2007 :: 800.0
Account: 1 - new balance: -400.0
Credit: Sun Apr 01 00:00:00 BST 2007 :: 200.0
Account: 1 - new balance: -200.0
Credit: Thu Apr 05 00:00:00 BST 2007 :: 300.0
Account: 1 - new balance: 100.0
Debit: Mon May 07 00:00:00 BST 2007 :: 900.0
Account: 1 - new balance: -800.0

The Pricing Rule decision table demonstrates the use of a decision table in a spreadsheet, in Excel's XLS format, in calculating the retail cost of an insurance policy. The purpose of the provide set of rules is to calculate a base price and a discount for a car driver applying for a specific policy. The driver's age, history and the policy type all contribute to what the basic premium is, and an additional chunk of rules deals with refining this with a discount percentage.

Name: Example Policy Pricing
Main class: org.drools.examples.PricingRuleDTExample
Type: Java application
Rules file: ExamplePolicyPricing.xls
Objective: demonstrate spreadsheet-based decision tables.

In this decision table, each row is a rule, and each column is a condition or an action.


Referring to the spreadsheet show above, we have the RuleSet declaration, which provides the package name. There are also other optional items you can have here, such as Variables for global variables, and Imports for importing classes. In this case, the namespace of the rules is the same as the fact classes we are using, so we can omit it.

Moving further down, we can see the RuleTable declaration. The name after this (Pricing bracket) is used as the prefix for all the generated rules. Below that, we have "CONDITION or ACTION", indicating the purpose of the column, i.e., whether it forms part of the condition or the consequence of the rule that will be generated.

You can see that there is a driver, his data spanned across three cells, which means that the template expressions below it apply to that fact. We observe the driver's age range (which uses $1 and $2 with comma-separated values), locationRiskProfile, and priorClaims in the respective columns. In the action columns, we are set the policy base price and log a message.


In the preceding spreadsheet section, there are broad category brackets, indicated by the comment in the leftmost column. As we know the details of our drivers and their policies, we can tell (with a bit of thought) that they should match row number 18, as they have no prior accidents, and are 30 years old. This gives us a base price of 120.


The above section contains the conditions for the discount we might grant our driver. The discount results from the Age bracket, the number of prior claims, and the policy type. In our case, the driver is 30, with no prior claims, and is applying for a COMPREHENSIVE policy, which means we can give a discount of 20%. Note that this is actually a separate table, but in the same worksheet, so that different templates apply.

It is important to note that decision tables generate rules. This means they aren't simply top-down logic, but more a means to capture data resulting in rules. This is a subtle difference that confuses some people. The evaluation of the rules is not necessarily in the given order, since all the normal mechanics of the rule engine still apply.

Name: Pet Store 
Main class: org.drools.examples.PetStore
Type: Java application
Rules file: PetStore.drl
Objective: Demonstrate use of Agenda Groups, Global Variables and integration with a GUI,
including callbacks from within the rules

The Pet Store example shows how to integrate Rules with a GUI, in this case a Swing based desktop application. Within the rules file, it demonstrates how to use Agenda groups and auto-focus to control which of a set of rules is allowed to fire at any given time. It also illustrates the mixing of the Java and MVEL dialects within the rules, the use of accumulate functions and the way of calling Java functions from within the ruleset.

All of the Java code is contained in one file, PetStore.java, defining the following principal classes (in addition to several classes to handle Swing Events):

Much of the Java code is either plain JavaBeans or Swing-based. Only a few Swing-related points will be discussed in this section, but a good tutorial about Swing components can be found at Sun's Swing website, in http://java.sun.com/docs/books/tutorial/uiswing/ .

The pieces of Java code in Petstore.java that relate to rules and facts are shown below.


The code shown above loads the rules from a DRL file on the classpath. Unlike other examples where the facts are asserted and fired straight away, this example defers this step to later. The way it does this is via the second last line where a PetStoreUI object is created using a constructor accepting the Vector object stock collecting our products, and an instance of the CheckoutCallback class containing the Rule Base that we have just loaded.

The Java code that fires the rules is within the CheckoutCallBack.checkout() method. This is triggered (eventually) when the Checkout button is pressed by the user.


Two items get passed into this method. One is the handle to the JFrame Swing component surrounding the output text frame, at the bottom of the GUI. The second is a list of order items; this comes from the TableModel storing the information from the "Table" area at the top right section of the GUI.

The for loop transforms the list of order items coming from the GUI into the Order JavaBean, also contained in the file PetStore.java. Note that it would be possible to refer to the Swing dataset directly within the rules, but it is better coding practice to do it this way, using simple Java objects. It means that we are not tied to Swing if we wanted to transform the sample into a Web application.

It is important to note that all state in this example is stored in the Swing components, and that the rules are effectively stateless. Each time the "Checkout" button is pressed, this code copies the contents of the Swing TableModel into the Session's Working Memory.

Within this code, there are nine calls to the Working Memory. The first of these creates a new Working Memory, as a Stateful Knowledge Session from the Knowledge Base. Remember that we passed in this Knowledge Base when we created the CheckoutCallBack class in the main() method. The next two calls pass in two objects that we will hold as global variables in the rules: the Swing text area and the Swing frame used for writing messages.

More inserts put information on products into the Working Memory, as well as the order list. The final call is the standard fireAllRules(). Next, we look at what this method causes to happen within the rules file.


The first part of file PetStore.drl contains the standard package and import statements to make various Java classes available to the rules. New to us are the two globals frame and textArea. They hold references to the Swing components JFrame and JTextArea components that were previously passed on by the Java code calling the setGlobal() method. Unlike variables in rules, which expire as soon as the rule has fired, global variables retain their value for the lifetime of the Session.

The next extract from the file PetStore.drl contains two functions that are referenced by the rules that we will look at shortly.


Having these functions in the rules file just makes the Pet Store example more compact. In real life you probably have the functions in a file of their own, within the same rules package, or as a static method on a standard Java class, and import them, using import function my.package.Foo.hello.

The purpose of these two functions is:

  • doCheckout() displays a dialog asking users whether they wish to checkout. If they do, focus is set to the checkOut agenda-group, allowing rules in that group to (potentially) fire.

  • requireTank() displays a dialog asking users whether they wish to buy a tank. If so, a new fish tank Product is added to the order list in Working Memory.

We'll see the rules that call these functions later on. The next set of examples are from the Pet Store rules themselves. The first extract is the one that happens to fire first, partly because it has the auto-focus attribute set to true.


This rule matches against all orders that do not yet have their grossTotal calculated . It loops for each purchase item in that order. Some parts of the "Explode Cart" rule should be familiar: the rule name, the salience (suggesting the order for the rules being fired) and the dialect set to "java". There are three new features:

  • agenda-group "init" defines the name of the agenda group. In this case, there is only one rule in the group. However, neither the Java code nor a rule consequence sets the focus to this group, and therefore it relies on the next attibute for its chance to fire.

  • auto-focus true ensures that this rule, while being the only rule in the agenda group, gets a chance to fire when fireAllRules() is called from the Java code.

  • kcontext....setFocus() sets the focus to the "show items" and "evaluate" agenda groups in turn, permitting their rules to fire. In practice, we loop through all items on the order, inserting them into memory, then firing the other rules after each insert.

The next two listings show the rules within the "show items" and evaluate agenda groups. We look at them in the order that they are called.


The "show items" agenda-group has only one rule, called "Show Items" (note the difference in case). For each purchase on the order currently in the Working Memory (or Session), it logs details to the text area at the bottom of the GUI. The textArea variable used to do this is one of the global variables we looked at earlier.

The evaluate Agenda group also gains focus from the "Explode Cart" rule listed previously. This Agenda group has two rules, "Free Fish Food Sample" and "Suggest Tank", shown below.


The rule "Free Fish Food Sample" will only fire if

  • we don't already have any fish food, and

  • we don't already have a free fish food sample, and

  • we do have a Gold Fish in our order.

If the rule does fire, it creates a new product (Fish Food Sample), and adds it to the order in Working Memory.

The rule "Suggest Tank" will only fire if

  • we don't already have a Fish Tank in our order, and

  • we do have more than 5 Gold Fish Products in our order.

If the rule does fire, it calls the requireTank() function that we looked at earlier (showing a Dialog to the user, and adding a Tank to the order / working memory if confirmed). When calling the requireTank() function the rule passes the global frame variable so that the function has a handle to the Swing GUI.

The next rule we look at is "do checkout".


The rule "do checkout" has no agenda group set and no auto-focus attribute. As such, is is deemed part of the default (MAIN) agenda group. This group gets focus by default when all the rules in agenda-groups that explicity had focus set to them have run their course.

There is no LHS to the rule, so the RHS will always call the doCheckout() function. When calling the doCheckout() function, the rule passes the global frame variable to give the function a handle to the Swing GUI. As we saw earlier, the doCheckout() function shows a confirmation dialog to the user. If confirmed, the function sets the focus to the checkout agenda-group, allowing the next lot of rules to fire.


There are three rules in the checkout agenda-group:

  • If we haven't already calculated the gross total, Gross Total accumulates the product prices into a total, puts this total into Working Memory, and displays it via the Swing JTextArea, using the textArea global variable yet again.

  • If our gross total is between 10 and 20, "Apply 5% Discount" calculates the discounted total and adds it to the Working Memory and displays it in the text area.

  • If our gross total is not less than 20, "Apply 10% Discount" calculates the discounted total and adds it to the Working Memory and displays it in the text area.

Now that we've run through what happens in the code, let's have a look at what happens when we actually run the code. The file PetStore.java contains a main() method, so that it can be run as a standard Java application, either from the command line or via the IDE. This assumes you have your classpath set correctly. (See the start of the examples section for more information.)

The first screen that we see is the Pet Store Demo. It has a list of available products (top left), an empty list of selected products (top right), checkout and reset buttons (middle) and an empty system messages area (bottom).


To get to this point, the following things have happened:

  1. The main() method has run and loaded the Rule Base but not yet fired the rules. So far, this is the only code in connection with rules that has been run.

  2. A new PetStoreUI object has been created and given a handle to the Rule Base, for later use.

  3. Various Swing components do their stuff, and the above screen is shown and waits for user input.

Clicking on various products from the list might give you a screen similar to the one below.


Note that no rules code has been fired here. This is only Swing code, listening for mouse click events, and adding some selected product to the TableModel object for display in the top right hand section. (As an aside, note that this is a classic use of the Model View Controller design pattern).

It is only when we press the "Checkout" button that we fire our business rules, in roughly the same order that we walked through the code earlier.

  1. Method CheckOutCallBack.checkout() is called (eventually) by the Swing class waiting for the click on the "Checkout" button. This inserts the data from the TableModel object (top right hand side of the GUI), and inserts it into the Session's Working Memory. It then fires the rules.

  2. The "Explode Cart" rule is the first to fire, given that it has auto-focus set to true. It loops through all the products in the cart, ensures that the products are in the Working Memory, and then gives the "Show Items" and Evaluation agenda groups a chance to fire. The rules in these groups add the contents of the cart to the text area (at the bottom of the window), decide whether or not to give us free fish food, and to ask us whether we want to buy a fish tank. This is shown in the figure below.


  1. The Do Checkout rule is the next to fire as it (a) No other agenda group currently has focus and (b) it is part of the default (MAIN) agenda group. It always calls the doCheckout() function which displays a 'Would you like to Checkout?' Dialog Box.

  2. The doCheckout() function sets the focus to the checkout agenda-group, giving the rules in that group the option to fire.

  3. The rules in the the checkout agenda-group display the contents of the cart and apply the appropriate discount.

  4. Swing then waits for user input to either checkout more products (and to cause the rules to fire again), or to close the GUI - see the figure below.


We could add more System.out calls to demonstrate this flow of events. The output, as it currently appears in the Console window, is given in the listing below.


Name: Honest Politician
Main class: org.drools.examples.HonestPoliticianExample 
Type: Java application
Rules file: HonestPoliticianExample.drl
Objective: Illustrate the concept of "truth maintenance" based on the logical insertion of facts

The Honest Politician example demonstrates truth maintenance with logical assertions. The basic premise is that an object can only exist while a statement is true. A rule's consequence can logically insert an object with the insertLogical() method. This means the object will only remain in the Working Memory as long as the rule that logically inserted it remains true. When the rule is no longer true the object is automatically retracted.

In this example there is the class Politician, with a name and a boolean value for being honest. Four politicians with honest state set to true are inserted.



The Console window output shows that, while there is at least one honest politician, democracy lives. However, as each politician is in turn corrupted by an evil corporation, so that all politicians become dishonest, democracy is dead.


As soon as there is at least one honest politician in the Working Memory a new Hope object is logically asserted. This object will only exist while there is at least one honest politician. As soon as all politicians are dishonest, the Hope object will be automatically retracted. This rule is given a salience of 10 to ensure that it fires before any other rule, as at this stage the "Hope is Dead" rule is actually true.


As soon as a Hope object exists the "Hope Lives" rule matches and fires. It has a salience of 10 so that it takes priority over "Corrupt the Honest".


Now that there is hope and we have, at the start, four honest politicians, we have four activations for this rule, all in conflict. They will fire in turn, corrupting each politician so that they are no longer honest. When all four politicians have been corrupted we have no politicians with the property honest == true. Thus, the rule "We have an honest Politician" is no longer true and the object it logical inserted (due to the last execution of new Hope()) is automatically retracted.


With the Hope object being automatically retracted, via the truth maintenance system, the conditional element not applied to Hope is no longer true so that the following rule will match and fire.


Let's take a look at the Audit trail for this application:


The moment we insert the first politician we have two activations. The rule "We have an honest Politician" is activated only once for the first inserted politician because it uses an exists conditional element, which matches once for any number. The rule "Hope is Dead" is also activated at this stage, because we have not yet inserted the Hope object. Rule "We have an honest Politician" fires first, as it has a higher salience than "Hope is Dead", which inserts the Hope object. (That action is highlighted green.) The insertion of the Hope object activates "Hope Lives" and de-activates "Hope is Dead"; it also activates "Corrupt the Honest" for each inserted honested politician. Rule "Hope Lives" executes, printing "Hurrah!!! Democracy Lives". Then, for each politician, rule "Corrupt the Honest" fires, printing "I'm an evil corporation and I have corrupted X", where X is the name of the politician, and modifies the politician's honest value to false. When the last honest politician is corrupted, Hope is automatically retracted, by the truth maintenance system, as shown by the blue highlighted area. The green highlighted area shows the origin of the currently selected blue highlighted area. Once the Hope fact is retracted, "Hope is dead" activates and fires printing "We are all Doomed!!! Democracy is Dead".

Name: Sudoku
Main class: org.drools.examples.sudoku.Main
Type: Java application
Rules file: sudokuSolver.drl, sudokuValidator.drl
Objective: Demonstrates the solving of logic problems, and complex pattern matching.

This example demonstrates how Drools can be used to find a solution in a large potential solution space based on a number of constraints. We use the popular puzzle of Sudoku. This example also shows how Drools can be integrated into a graphical interface and how callbacks can be used to interact with a running Drools rules engine in order to update the graphical interface based on changes in the Working Memory at runtime.

Download and install drools-examples as described above and then execute java org.drools.examples.sudoku.Main. This example requires Java 5.

A window will be displayed with a relatively simple partially filled grid.

Click on the "Solve" button and the Drools-based engine will fill out the remaining values. The Console window will display detailed information about the rules which are executing to solve the puzzle in a human readable form.

Rule #3 determined the value at (4,1) could not be 4 as this value already exists in the same column at (8,1)
Rule #3 determined the value at (5,5) could not be 2 as this value already exists in the same row at (5,6)
Rule #7 determined (3,5) is 2 as this is the only possible cell in the column that can have this value
Rule #1 cleared the other PossibleCellValues for (3,5) as a ResolvedCellValue of 2 exists for this cell.
Rule #1 cleared the other PossibleCellValues for (3,5) as a ResolvedCellValue of 2 exists for this cell.
... 
Rule #3 determined the value at (1,1) could not be 1 as this value already  exists in the same zone at (2,1)
Rule #6 determined (1,7) is 1 as this is the only possible cell in the row that can have this value
Rule #1 cleared the other PossibleCellValues for (1,7) as a ResolvedCellValue of 1 exists for this cell.
Rule #6 determined (1,1) is 8 as this is the only possible cell in the row that can have this value

Once all of the activated rules for the solving logic have executed, the engine executes a second rule base to check that the solution is complete and valid. In this case it is, and the "Solve" button is disabled and displays a text like "Solved (1052ms)".

The example comes with a number of grids which can be loaded and solved. Click on "File", then "Samples" and "Medium" to load a more challenging grid. Note that the solve button is enabled when the new grid is loaded.

Click on the "Solve" button again to solve this new grid.

Now, let us load a Sudoku grid that is deliberately invalid. Click on "File", "Samples" and "!DELIBERATELY BROKEN!". Note that this grid starts with some issues, for example the value 5 appears twice in the first row.

Nevertheless, click on the "Solve" button to apply the solving rules to this invalid grid. Note that the "Solve" button is relabelled to indicate that the resulting solution is invalid.

In addition, the validation rule set outputs all of the issues which are discovered to the console.

There are two cells on the same column with the same value at (6,0) and (4,0)
There are two cells on the same column with the same value at (4,0) and (6,0)
There are two cells on the same row with the same value at (2,4) and (2,2)
There are two cells on the same row with the same value at (2,2) and (2,4)
There are two cells on the same row with the same value at (6,3) and (6,8)
There are two cells on the same row with the same value at (6,8) and (6,3)
There are two cells on the same column with the same value at (7,4) and (0,4)
There are two cells on the same column with the same value at (0,4) and (7,4)
There are two cells on the same row with the same value at (0,8) and (0,0)
There are two cells on the same row with the same value at (0,0) and (0,8)
There are two cells on the same column with the same value at (1,2) and (3,2)
There are two cells on the same column with the same value at (3,2) and (1,2)
There are two cells in the same zone with the same value at (6,3) and (7,3)
There are two cells in the same zone with the same value at (7,3) and (6,3)
There are two cells on the same column with the same value at (7,3) and (6,3)
There are two cells on the same column with the same value at (6,3) and (7,3)

We will look at the solving rule set later in this section, but for the moment we should note that some theoretically solvable solutions can not be solved by the engine as it stands. Click on "File", "Samples" and then "Hard 3" to load a sparsely populated grid.

Now click on the "Solve" button and note that the current rules are unable to complete the grid, even though (if you are a Sudoku aficionado) you may be able to see a way forward with the solution.

At the present time, the solving functionality has been achieved by the use of ten rules. This rule set could be extended to enable the engine to tackle more complex logic for filling grids such as this.

The Java source code can be found in the /src/main/java/org/drools/examples/sudoku directory, with the two DRL files defining the rules located in the /src/main/rules/org/drools/examples/sudoku directory.

The package org.drools.examples.sudoku.swing contains a set of classes which implement a framework for Sudoku puzzles. Note that this package does not have any dependencies on the Drools libraries. SudokuGridModel defines an interface which can be implemented to store a Sudoku puzzle as a 9x9 grid of Integer values, some of which may be null, indicating that the value for the cell has not yet been resolved. SudokuGridView is a Swing component which can visualize any implementation of SudokuGridModel. SudokuGridEvent and SudokuGridListener are used to communicate state changes between the model and the view: events are fired when a cell's value is resolved or changed. If you are familiar with the model-view-controller patterns in other Swing components such as JTable then this pattern should be familiar. SudokuGridSamples provides a number of partially filled Sudoku puzzles for demonstration purposes.

Package org.drools.examples.sudoku.rules contains an implementation of SudokuGridModel which is based on Drools. Two Java objects are used, both of which extend AbstractCellValue and represent a value for a specific cell in the grid, including the row and column location of the cell, an index of the 3x3 zone the cell is contained in, and the value of the cell. PossibleCellValue indicates that we do not currently know for sure what the value in a cell is. There can be from 2 to 9 possible cell values for a given cell. ResolvedCellValue indicates that we have determined what the value for a cell must be. There can only be one resolved cell value for a given cell. DroolsSudokuGridModel implements SudokuGridModel and is responsible for converting an initial two dimensional array of partially specified cells into a set of CellValue Java object, creating a Working Memory based on solverSudoku.drl and inserting the CellValue objects into the Working Memory. When the solve() method is called it calls in turn fireAllRules() on this Working Memory to try to solve the puzzle. DroolsSudokuGridModel attaches a WorkingMemoryListener to the Working Memory, which allows it to be called back on insert and retract events as the puzzle is solved. When a new ResolvedCellValue is inserted into the Working Memory, this callback allows the implementation to fire a SudokuGridEvent to its SudokuGridListener clientele, which can then update themselves in realtime. Once all the rules fired by the solver Working Memory have executed, DroolsSudokuGridModel runs a second set of rules, based on validatorSudoku.drl which works with the same set of Java objects to determine whether the resulting grid is a valid and a full solution.

The class org.drools.examples.sudoku.Main implements a Java application combining the components desribed.

The packae org.drools.examples.sudoku contains two DRL files. solverSudoku.drl defines the rules which attempt to solve a Sudoku puzzle, and validator.drl defines the rules which determin whether the current state of the Working Memory represents a valid solution. Both use PossibleCellValue and ResolvedCellValue objects as their facts and both output information to the Console window as their rules fire. In a real-world situation we would insert logging information and use the WorkingMemoryListener to display this information to a user, rather than use the console in this fashion.

Now let us look at the more complex rule set used to solve Sudoku puzzles.

Rule #1 is basically a book-keeping rule. Several of the other rules insert ResolvedCellValues into the working memory at specific rows and columns after they have determined that a given cell must have a certain value. At this point, it is important to clear the Working Memory of any inserted PossibleCellValues at the same row and column with invalid values. This rule is therefore given a higher salience than the remaining rules to ensure that as soon as the LHS is true, activations for the rule move to the top of the Agenda and are fired. In turn, this prevents the spurious firing of other rules due to the combination of a ResolvedCellValue and one or more PossibleCellValues being present in the same cell. This rule also calls update() on the ResolvedCellValue, even though its value has not in fact been modified to ensure that Drools fires an event to any WorkingMemoryListeners attached to the Working Memory so that they can update themselves - in this case so that the GUI can display the new state of the grid.

Rule #2 identifies cells in the grid which have only one possible value. The first line of the when clause matches all of the PossibleCellValue objects in the Working Memory. The second line demonstrates a use of the not keyword. This rule will only fire if no other PossibleCellValue objects exist in the Working Memory at the same row and column but with a different value. When the rule fires, the single PossibleCellValue at the row and column is retracted from the Working Memory and is replaced by a new ResolvedCellValue at the same row and column with the same value.

Rule #3 removes PossibleCellValues with a given value from a row when they have the same value as a ResolvedCellValue. In other words, when a cell is filled with a resolved value, we need to remove the possibility of any other cell on the same row having this value. The first line of the when clause matches all ResolvedCellValue objects in the Working Memory. The second line matches PossibleCellValues which have both the same row and the same value as these ResolvedCellValue objects. If any are found, the rule activates and, when fired retracts the PossibleCellValue which can no longer be a solution for that cell.

Rules #4 and #5 act in the same way as Rule #3 but check for redundant PossibleCellValues in a given column and a given zone of the grid as a ResolvedCellValue respectively.

Rule #6 checks for the scenario where a possible cell value only appears once in a given row. The first line of the LHS matches against all PossibleCellValue facts in the Working Memory, storing the result in a number of local variables. The second line checks that no other PossibleCellValue objects with the same value exist on this row. The third to fifth lines check that there is not a ResolvedCellValue with the same value in the same zone, row or column so that this rule does not fire prematurely. It is interesting to note that we could remove lines 3 to 5 and give rules #3, #4 and #5 a higher salience to make sure they always fire before rules #6,#7 and #8. When the rule fires, we know that $possible must represent the value for the cell; so, as in Rule #2, we retract $possible and replace it with the equivalent, new ResolvedCellValue.

Rules #7 and #8 act in the same way as Rule #2 but check for single PossibleCellValues in a given column and a given zone of the grid, respectively.

Rule #9 represents the most complex currently implemented rule. This rule implements the logic that, if we know that a pair of given values can only occur in two cells on a specific row, (for example we have determined the values of 4 and 6 can only appear in the first row in cells [0,3] and [0,5]) and this pair of cells can not hold other values, then, although we do not know which of the pair contains a four and which contains a six, we do know that these two values must be in these two cells, and hence we can remove the possibility of them occuring anywhere else in the same row.

Rules #10 and #11 act in the same way as rule #9 but check for the existance of only two possible values in a given column and zone, respectively.

To solve harder grids, the rule set would need to be extended further with more complex rules that encapsulate more complex reasoning.

There are a number of ways in which this example could be developed. The reader is encouraged to consider these as excercises.

Name: Number Guess 
Main class: org.drools.examples.NumberGuessExample
Type: Java application
Rules file: NumberGuess.drl
Objective: Demonstrate use of Rule Flow to organise Rules

The "Number Guess" example shows the use of Rule Flow, a way of controlling the order in which rules are fired. It uses widely understood workflow diagrams for defining the order in which groups of rules will be executed.


The creation of the package and the loading of the rules (using the add() method) is the same as the previous examples. There is an additional line to add the Rule Flow (NumberGuess.rf), which provides the option of specifying different rule flows for the same Knowledge Base. Otherwise, the Knowledge Base is created in the same manner as before.


Once we have a Knowledge Base, we can use it to obtain a Stateful Session. Into our session we insert our facts, i.e., standard Java objects. (For simplicity, in this sample, these classes are all contained within our NumberGuessExample.java file. Class GameRules provides the maximum range and the number of guesses allowed. Class RandomNumber automatically generates a number between 0 and 100 and makes it available to our rules, by insertion via the getValue() method. Class Game keeps track of the guesses we have made before, and their number.

Note that before we call the standard fireAllRules() method, we also start the process that we loaded earlier, via the startProcess() method. We'll learn where to obtain the parameter we pass ("Number Guess", i.e., the identifier of the rule flow) when we talk about the rule flow file and the graphical Rule Flow Editor below.

Before we finish the discussion of our Java code, we note that in some real-life application we would examine the final state of the objects. (Here, we could retrieve the number of guesses, to add it to a high score table.) For this example we are content to ensure that the Working Memory session is cleared by calling the dispose() method.


If you open the NumberGuess.rf file in the Drools IDE (provided you have the JBoss Rules extensions installed correctly in Eclipse) you should see the above diagram, similar to a standard flowchart. Its icons are similar (but not exactly the same) as in the JBoss jBPM workflow product. Should you wish to edit the diagram, a menu of available components should be available to the left of the diagram in the IDE, which is called the palette. This diagram is saved in XML, an (almost) human readable format, using XStream.

If it is not already open, ensure that the Properties View is visible in the IDE. It can be opened by clicking "Window", then "Show View" and "Other", where you can select the "Properties" view. If you do this before you select any item on the rule flow (or click on the blank space in the rule flow) you should be presented with the following set of properties.


Keep an eye on the Properties View as we progress through the example's rule flow, as it presents valuable information. In this case, it provides us with the identification of the Rule Flow Process that we used in our earlier code snippet, when we called session.startProcess().

In the "Number Guess" Rule Flow we encounter several node types, many of them identified by an icon.

  • The Start node (white arrow in a green circle) and the End node (red box) mark beginning and end of the rule flow.

  • A Rule Flow Group box (yellow, without an icon) represents a Rule Flow Groups defined in our rules (DRL) file that we will look at later. For example, when the flow reaches the Rule Flow Group "Too High", only those rules marked with an attribute of ruleflow-group "Too High" can potentially fire.

  • Action nodes (yellow, cog-shaped icon) perform standard Java method calls. Most action nodes in this example call System.out.println(), indicating the program's progress to the user.

  • Split and Join Nodes (blue ovals, no icon) such as "Guess Correct?" and "More guesses Join" mark places where the flow of control can split, according to various conditions, and rejoin, respectively

  • Arrows indicate the flow between the various nodes.

The various nodes in combination with the rules make the Number Guess game work. For example, the "Guess" Rule Flow Group allows only the rule "Get user Guess" to fire, because only that rule has a matching attribute of ruleflow-group "Guess".


The rest of this rule is fairly standard. The LHS section (after when) of the rule states that it will be activated for each RandomNumber object inserted into the Working Memory where guessCount is less than allowedGuesses from the GameRules object and where the user has not guessed the correct number.

The RHS section (or consequence, after then) prints a message to the user and then awaits user input from System.in. After obtaining this input (the readLine() method call blocks until the return key is pressed) it modifies the guess count and inserts the new guess, making both available to the Working Memory.

The rest of the rules file is fairly standard: the package declares the dialect as MVEL, and various Java classes are imported. In total, there are five rules in this file:

  1. Get User Guess, the Rule we examined above.

  2. A Rule to record the highest guess.

  3. A Rule to record the lowest guess.

  4. A Rule to inspect the guess and retract it from memory if incorrect.

  5. A Rule that notifies the user that all guesses have been used up.

One point of integration between the standard Rules and the RuleFlow is via the ruleflow-group attribute on the rules, as dicussed above. A second point of integration between the rules (.drl) file and the Rules Flow .rf files is that the Split Nodes (the blue ovals) can use values in the Working Memory (as updated by the rules) to decide which flow of action to take. To see how this works, click on the "Guess Correct Node"; then within the Properties View, open the Constraints Editor by clicking the button at the right that appears once you click on the "Constraints" property line. You should see something similar to the diagram below.


Click on the "Edit" button beside "To node Too High" and you'll see a dialog like the one below. The values in the "Textual Editor" window follow the standard rule format for the LHS and can refer to objects in Working Memory. The consequence (RHS) is that the flow of control follows this node (i.e., "To node Too High") if the LHS expression evaluates to true.


Since the file NumberGuess.java contains a main() method, it can be run as a standard Java application, either from the command line or via the IDE. A typical game might result in the interaction below. The numbers in bold are typed in by the user.


A summary of what is happening in this sample is:

  1. The main() method of NumberGuessExample.java loads a Rule Base, creates a Stateful Session and inserts Game, GameRules and RandomNumber (containing the target number) objects into it. The method also sets the process flow we are going to use, and fires all rules. Control passes to the Rule Flow.

  2. File NumberGuess.rf, the Rule Flow, begins at the "Start" node.

  3. Control passes (via the "More guesses" join node) to the Guess node.

  4. At the Guess node, the appropriate Rule Flow Group ("Get user Guess") is enabled. In this case the Rule "Guess" (in the NumberGuess.drl file) is triggered. This rule displays a message to the user, takes the response, and puts it into Working Memory. Flow passes to the next Rule Flow Node.

  5. At the next node, "Guess Correct", constraints inspect the current session and decide which path to take.

    If the guess in step 4 was too high or too low, flow proceeds along a path which has an action node with normal Java code printing a suitable message and a Rule Flow Group causing a highest guess or lowest guess rule to be triggered. Flow passes from these nodes to step 6.

    If the guess in step 4 was right, we proceed along the path towards the end of the Rule Flow. Before we get there, an action node with normal Java code prints a statement "you guessed correctly". There is a join node here (just before the Rule Flow end) so that our no-more-guesses path (step 7) can also terminate the Rule Flow.

  6. Control passes as per the Rule Flow via a join node, a guess incorrect Rule Flow Group (triggering a rule to retract a guess from Working Memory) onto the "More guesses" decision node.

  7. The "More guesses" decision node (on the right hand side of the rule flow) uses constraints, again looking at values that the rules have put into the working memory, to decide if we have more guesses and if so, goto step 3. If not, we proceed to the end of the rule flow, via a Rule Flow Group that triggers a rule stating "you have no more guesses".

  8. The loop over steps 3 to 7 continues until the number is guessed correctly, or we run out of guesses.

Name: Miss Manners
Main class: org.drools.benchmark.manners.MannersBenchmark
Type: Java application
Rules file: manners.drl
Objective: Advanced walkthrough on the Manners benchmark, covers Depth conflict resolution in depth.

Miss Manners is throwing a party and, being a good host, she wants to arrange good seating. Her initial design arranges everyone in male-female pairs, but then she worries about people have things to talk about. What is a good host to do? She decides to note the hobby of each guest so she can then arrange guests not only pairing them according to alternating sex but also ensuring that a guest has someone with a common hobby, at least on one side.


Before going deeper into the rules, let's first take a look at the asserted data and the resulting seating arrangement. The data is a simple set of five guests who should be arranged so that sexes alternate and neighbors have a common hobby.

The Data

The data is given in OPS5 syntax, with a parenthesized list of name and value pairs for each attribute. Each person has only one hobby.

The Results

Each line of the results list is printed per execution of the "Assign Seat" rule. They key bit to notice is that each line has a "pid" value one greater than the last. (The significance of this will be explained in the discussion of the rule "Assign Seating".) The "ls", "rs", "ln" and "rn" refer to the left and right seat and neighbor's name, respectively. The actual implementation uses longer attribute names (e.g., leftGuestName, but here we'll stick to the notation from the original implementation.

The Manners benchmark was written for OPS5 which has two conflict resolution strategies, LEX and MEA. LEX is a chain of several strategies including salience, recency and complexity. The recency part of the strategy drives the depth first (LIFO) firing order. The CLIPS manual documents the Recency strategy as follows:

 

Every fact and instance is marked internally with a "time tag" to indicate its relative recency with respect to every other fact and instance in the system. The pattern entities associated with each rule activation are sorted in descending order for determining placement. An activation with a more recent pattern entity is placed before activations with less recent pattern entities. To determine the placement order of two activations, compare the sorted time tags of the two activations one by one starting with the largest time tags. The comparison should continue until one activation’s time tag is greater than the other activation’s corresponding time tag. The activation with the greater time tag is placed before the other activation on the agenda. If one activation has more pattern entities than the other activation and the compared time tags are all identical, then the activation with more time tags is placed before the other activation on the agenda.

 
 --CLIPS Reference Manual

However Jess and CLIPS both use the Depth strategy, which is simpler and lighter, which Drools also adopted. The CLIPS manual documents the Depth strategy as:

 

Newly activated rules are placed above all rules of the same salience. For example, given that fact-a activates rule-1 and rule-2 and fact-b activates rule-3 and rule-4, then if fact-a is asserted before fact-b, rule-3 and rule-4 will be above rule-1 and rule-2 on the agenda. However, the position of rule-1 relative to rule-2 and rule-3 relative to rule-4 will be arbitrary.

 
 --CLIPS Reference Manual

The initial Drools implementation for the Depth strategy would not work for Manners without the use of salience on the "make_path" rule. The CLIPS support team had this to say:

 

The default conflict resolution strategy for CLIPS, Depth, is different than the default conflict resolution strategy used by OPS5. Therefore if you directly translate an OPS5 program to CLIPS, but use the default depth conflict resolution strategy, you're only likely to get the correct behavior by coincidence. The LEX and MEA conflict resolution strategies are provided in CLIPS to allow you to quickly convert and correctly run an OPS5 program in CLIPS.

 
 --Clips Support Forum

Investigation into the CLIPS code reveals there is undocumented functionality in the Depth strategy. There is an accumulated time tag used in this strategy; it's not an extensively fact by fact comparison as in the recency strategy, it simply adds the total of all the time tags for each activation and compares.

This rule determines each of the Seating arrangements. The rule creates cross product solutions for all asserted Seating arrangements against all the asserted guests except against itself or any already assigned chosen solutions.

rule findSeating
   when 
       context : Context( state == Context.ASSIGN_SEATS )
       $s      : Seating( pathDone == true )
       $g1     : Guest( name == $s.rightGuestName )
       $g2     : Guest( sex != $g1.sex, hobby == $g1.hobby )

       count   : Count()

       not ( Path( id == $s.id, guestName == $g2.name) )
       not ( Chosen( id == $s.id, guestName == $g2.name, hobby == $g1.hobby) )
   then
       int rightSeat = $s.getRightSeat();
       int seatId = $s.getId();
       int countValue = count.getValue();
       
       Seating seating =
         new Seating( countValue, seatId, false, rightSeat,
                      $s.getRightGuestName(), rightSeat + 1, $g2.getName() );
       insert( seating );
                            
       Path path = new Path( countValue, rightSeat + 1, $g2.getName()  );
       insert( path );
       
       Chosen chosen = new Chosen( seatId, $g2.getName(), $g1.getHobby() );
       insert( chosen  );

       System.err.println( "find seating : " + seating + " : " + path +
                           " : " + chosen);

       modify( count ) {setValue(  countValue + 1 )}
       modify( context ) {setState( Context.MAKE_PATH )}
end

However, as can be seen from the printed results shown earlier, it is essential that only the Seating with the highest pid cross product be chosen. How can this be possible if we have activations, of the same time tag, for nearly all existing Seating and Guest objects? For example, on the third iteration of findDeating the produced activations will be as shown below. Remember, this is from a very small data set, and with larger data sets there would be many more possible activated Seating solutions, with multiple solutions per pid:

The creation of all these redundant activations might seem pointless, but it must be remembered that Manners is not about good rule design; it's purposefully designed as a bad ruleset to fully stress-test the cross product matching process and the Agenda, which this clearly does. Notice that each activation has the same time tag of 35, as they were all activated by the change in the Context object to ASSIGN_SEATS. With OPS5 and LEX it would correctly fire the activation with the Seating asserted last. With Depth, the accumulated fact time tag ensures that the activation with the last asserted Seating fires.

Assign First seat
=>[fid:13:13]:[Seating id=1, pid=0, done=true, ls=1, ln=n5, rs=1, rn=n5]
=>[fid:14:14]:[Path id=1, seat=1, guest=n5]

==>[ActivationCreated(16): rule=findSeating
[fid:13:13]:[Seating id=1, pid=0, done=true, ls=1, ln=n5, rs=1, rn=n5]
[fid:9:9]:[Guest name=n5, sex=f, hobbies=h1]
[fid:1:1]:[Guest name=n1, sex=m, hobbies=h1]

==>[ActivationCreated(16): rule=findSeating
[fid:13:13]:[Seating id=1 , pid=0, done=true, ls=1, ln=n5, rs=1, rn=n5]
[fid:9:9]:[Guest name=n5, sex=f, hobbies=h1]
[fid:5:5]:[Guest name=n4, sex=m, hobbies=h1]*

Assign Seating
=>[fid:15:17] :[Seating id=2 , pid=1 , done=false, ls=1, lg=n5, rs=2, rn=n4]
=>[fid:16:18]:[Path id=2, seat=2, guest=n4]
=>[fid:17:19]:[Chosen id=1, name=n4, hobbies=h1]

=>[ActivationCreated(21): rule=makePath 
[fid:15:17] : [Seating id=2, pid=1, done=false, ls=1, ln=n5, rs=2, rn=n4]
[fid:14:14] : [Path id=1, seat=1, guest=n5]*

==>[ActivationCreated(21): rule=pathDone
[Seating id=2, pid=1, done=false, ls=1, ln=n5, rs=2, rn=n4]*

Make Path
=>[fid:18:22:[Path id=2, seat=1, guest=n5]]

Path Done

Continue Process
=>[ActivationCreated(25): rule=findSeating
[fid:15:23]:[Seating id=2, pid=1, done=true, ls=1, ln=n5, rs=2, rn=n4]
[fid:7:7]:[Guest name=n4, sex=f, hobbies=h3]
[fid:4:4] : [Guest name=n3, sex=m, hobbies=h3]*

=>[ActivationCreated(25): rule=findSeating
[fid:15:23]:[Seating id=2, pid=1, done=true, ls=1, ln=n5, rs=2, rn=n4]
[fid:5:5]:[Guest name=n4, sex=m, hobbies=h1]
[fid:2:2]:[Guest name=n2, sex=f, hobbies=h1], [fid:12:20] : [Count value=3]

=>[ActivationCreated(25): rule=findSeating
[fid:13:13]:[Seating id=1, pid=0, done=true, ls=1, ln=n5, rs=1, rn=n5]
[fid:9:9]:[Guest name=n5, sex=f, hobbies=h1]
[fid:1:1]:[Guest name=n1, sex=m, hobbies=h1]

Assign Seating
=>[fid:19:26]:[Seating id=3, pid=2, done=false, ls=2, lnn4, rs=3, rn=n3]]
=>[fid:20:27]:[Path id=3, seat=3, guest=n3]]
=>[fid:21:28]:[Chosen id=2, name=n3, hobbies=h3}]

=>[ActivationCreated(30): rule=makePath
[fid:19:26]:[Seating id=3, pid=2, done=false, ls=2, ln=n4, rs=3, rn=n3]
[fid:18:22]:[Path id=2, seat=1, guest=n5]*

=>[ActivationCreated(30): rule=makePath 
[fid:19:26]:[Seating id=3, pid=2, done=false, ls=2, ln=n4, rs=3, rn=n3]
[fid:16:18]:[Path id=2, seat=2, guest=n4]*

=>[ActivationCreated(30): rule=done 
[fid:19:26]:[Seating id=3, pid=2, done=false, ls=2, ln=n4, rs=3, rn=n3]*

Make Path
=>[fid:22:31]:[Path id=3, seat=1, guest=n5]

Make Path 
=>[fid:23:32] [Path id=3, seat=2, guest=n4]

Path Done

Continue Processing
=>[ActivationCreated(35): rule=findSeating
[fid:19:33]:[Seating id=3, pid=2, done=true, ls=2, ln=n4, rs=3, rn=n3]
[fid:4:4]:[Guest name=n3, sex=m, hobbies=h3]
[fid:3:3]:[Guest name=n2, sex=f, hobbies=h3], [fid:12:29]*

=>[ActivationCreated(35): rule=findSeating 
[fid:15:23]:[Seating id=2, pid=1, done=true, ls=1, ln=n5, rs=2, rn=n4] 
[fid:5:5]:[Guest name=n4, sex=m, hobbies=h1]
[fid:2:2]:[Guest name=n2, sex=f, hobbies=h1]

=>[ActivationCreated(35): rule=findSeating 
[fid:13:13]:[Seating id=1, pid=0, done=true, ls=1, ln=n5, rs=1, rn=n5] 
[fid:9:9]:[Guest name=n5, sex=f, hobbies=h1], [fid:1:1] : [Guest name=n1, sex=m, hobbies=h1]

Assign Seating
=>[fid:24:36]:[Seating id=4, pid=3, done=false, ls=3, ln=n3, rs=4, rn=n2]]
=>[fid:25:37]:[Path id=4, seat=4, guest=n2]]
=>[fid:26:38]:[Chosen id=3, name=n2, hobbies=h3]

==>[ActivationCreated(40): rule=makePath 
[fid:24:36]:[Seating id=4, pid=3, done=false, ls=3, ln=n3, rs=4, rn=n2]
[fid:23:32]:[Path id=3, seat=2, guest=n4]*

==>[ActivationCreated(40): rule=makePath 
[fid:24:36]:[Seating id=4, pid=3, done=false, ls=3, ln=n3, rs=4, rn=n2] 
[fid:20:27]:[Path id=3, seat=3, guest=n3]*

=>[ActivationCreated(40): rule=makePath 
[fid:24:36]:[Seating id=4, pid=3, done=false, ls=3, ln=n3, rs=4, rn=n2]
[fid:22:31]:[Path id=3, seat=1, guest=n5]*

=>[ActivationCreated(40): rule=done 
[fid:24:36]:[Seating id=4, pid=3, done=false, ls=3, ln=n3, rs=4, rn=n2]*

Make Path 
=>fid:27:41:[Path id=4, seat=2, guest=n4]

Make Path
=>fid:28:42]:[Path id=4, seat=1, guest=n5]]

Make Path
=>fid:29:43]:[Path id=4, seat=3, guest=n3]]

Path Done

Continue  Processing
=>[ActivationCreated(46): rule=findSeating 
[fid:15:23]:[Seating id=2, pid=1, done=true, ls=1, ln=n5, rs=2, rn=n4] 
[fid:5:5]:[Guest name=n4, sex=m, hobbies=h1], [fid:2:2]
[Guest name=n2, sex=f, hobbies=h1]

=>[ActivationCreated(46): rule=findSeating 
[fid:24:44]:[Seating id=4, pid=3, done=true, ls=3, ln=n3, rs=4, rn=n2]
[fid:2:2]:[Guest name=n2, sex=f, hobbies=h1]
[fid:1:1]:[Guest name=n1, sex=m, hobbies=h1]*

=>[ActivationCreated(46): rule=findSeating 
[fid:13:13]:[Seating id=1, pid=0, done=true, ls=1, ln=n5, rs=1, rn=n5]
[fid:9:9]:[Guest name=n5, sex=f, hobbies=h1]
[fid:1:1]:[Guest name=n1, sex=m, hobbies=h1]

Assign Seating
=>[fid:30:47]:[Seating id=5, pid=4, done=false, ls=4, ln=n2, rs=5, rn=n1]
=>[fid:31:48]:[Path id=5, seat=5, guest=n1]
=>[fid:32:49]:[Chosen id=4, name=n1, hobbies=h1]

Name: Conway's Game Of Life
Main class: org.drools.examples.conway.ConwayAgendaGroupRun
            org.drools.examples.conway.ConwayRuleFlowGroupRun
Type: Java application
Rules file: conway-ruleflow.drl conway-agendagroup.drl
Objective: Demonstrates 'accumulate', 'collect' and 'from'

Conway's Game Of Life, described in http://en.wikipedia.org/wiki/Conway's_Game_of_Life and in http://www.math.com/students/wonders/life/life.html, is a famous cellular automaton conceived in the early 1970's by the mathematician John Conway. While the system is well known as "Conway's Game Of Life", it really isn't a game at all. Conway's system is more like a simulation of a form of life. Don't be intimidated. The system is terribly simple and terribly interesting. Math and Computer Science students alike have marvelled over Conway's system for more than 30 years now. The application presented here is a Swing-based implementation of Conway's Game of Life. The rules that govern the system are implemented as business rules using Drools. This document will explain the rules that drive the simulation and discuss the Drools parts of the implementation.

We'll first introduce the grid view, shown below, designed for the visualisation of the game, showing the "arena" where the life simuation takes place. Initially the grid is empty, meaning that there are no live cells in the system. Each cell is either alive or dead, with live cells showing a green ball. Preselected patterns of live cells can be chosen from the "Pattern" drop-down list. Alternatively, individual cells can be doubled-clicked to toggle them between live and dead. It's important to understand that each cell is related to its neighboring cells, which is fundamental for the game's rules. Neighbors include not only cells to the left, right, top and bottom but also cells that are connected diagonally, so that each cell has a total of 8 neighbors. Exceptions are the four corner cells which have only three neighbors, and the cells along the four border, with five neighbors each.


So what are the basic rules that govern this game? Its goal is to show the development of a population, generation by generation. Each generation results from the preceding one, based on the simultaneous evaluation of all cells. This is the simple set of rules that govern what the next generation will look like:

  • If a live cell has fewer than 2 live neighbors, it dies of loneliness.

  • If a live cell has more than 3 live neighbors, it dies from overcrowding.

  • If a dead cell has exactly 3 live neighbors, it comes to life.

That is all there is to it. Any cell that doesn't meet any of those criteria is left as is for the next generation. With those simple rules in mind, go back and play with the system a little bit more and step through some generations, one at a time, and notice these rules taking their effect.

The screenshot below shows an example generation, with a number of live cells. Don't worry about matching the exact patterns represented in the screen shot. Just get some groups of cells added to the grid. Once you have groups of live cells in the grid, or select a pre-designed pattern, click the "Next Generation" button and notice what happens. Some of the live cells are killed (the green ball disappears) and some dead cells come to life (a green ball appears). Step through several generations and see if you notice any patterns. If you click on the "Start" button, the system will evolve itself so you don't need to click the "Next Generation" button over and over. Play with the system a little and then come back here for more details of how the application works.


Now lets delve into the code. As this is an advanced example we'll assume that by now you know your way around the Drools framework and are able to connect the presented highlight, so that we'll just focus at a high level overview. The example has two ways to execute, one way uses Agenda Groups to manage execution flow, and the other one uses Rule Flow Groups to manage execution flow. These two versions are implemented in ConwayAgendaGroupRun and ConwayRuleFlowGroupRun, respectively. Here, we'll discuss the Rule Flow version, as it's what most people will use.

All the Cell objects are inserted into the Session and the rules in the ruleflow-group "register neighbor" are allowed to execute by the Rule Flow process. This group of four rules creates Neighbor relations between some cell and its northeastern, northern, northwestern and western neighbors. This relation is bidirectional, which takes care of the other four directions. Border cells don't need any special treatment - they simply won't be paired with neighboring cells where there isn't any. By the time all activations have fired for these rules, all cells are related to all their neighboring cells.


Once all the cells are inserted, some Java code applies the pattern to the grid, setting certain cells to Live. Then, when the user clicks "Start" or "Next Generation", it executes the "Generation" ruleflow. This ruleflow is responsible for the management of all changes of cells in each generation cycle.


The rule flow process first enters the "evaluate" group, which means that any active rule in the group can fire. The rules in this group apply the Game-of-Life rules discussed in the beginning of the example, determining the cells to be killed and the ones to be given life. We use the "phase" attribute to drive the reasoning of the Cell by specific groups of rules; typically the phase is tied to a Rule Flow Group in the Rule Flow process definition. Notice that it doesn't actually change the state of any Cell objectss at this point; this is because it's evaluating the grid in turn and it must complete the full evaluation until those changes can be applied. To achieve this, it sets the cell to a "phase" which is either Phase.KILL or Phase.BIRTH, used later to control actions applied to the Cell object.


Once all Cell objects in the grid have been evaluated, we first clear any calculation activations that occured from any previous data changes. This is done via the "reset calculate" rule, which clears any activations in the "calculate" group. We then enter a split in the rule flow which allows any activations in both the "kill" and the "birth" group to fire. These rules are responsible for applying the state change.


At this stage, a number of Cell objects have been modified with the state changed to either LIVE or DEAD. Now we get to see the power of the Neighbor facts defining the cell relations. When a cell becomes live or dead, we use the Neighbor relation to iterate over all surrounding cells, increasing or decreasing the liveNeighbor count. Any cell that has its count changed is also set to to the EVALUATE phase, to make sure it is included in the reasoning during the evaluation stage of the Rule Flow Process. Notice that we don't have to do any iteration ourselves; simply by applying the relations in the rules we make the rule engine do all the hard work for us, with a minimal amount of code. Once the live count has been determined and set for all cells, the Rule Flow Process comes to and end. If the user has initially clicked the "Start" button, the engine will restart the rule flow; otherwise the user may request another generation.