Documentation

The domain object model

The model has a central position in a play application. It’s the domain-specific representation of the information on which the application operates.

Martin fowler defines it as:

Responsible for representing concepts of the business, information about the business situation, and business rules. State that reflects the business situation is controlled and used here, even though the technical details of storing it are delegated to the infrastructure. This layer is the heart of business software.

A common Java anti-pattern is to keep the model as a set of simple Java Beans and put the application logic back into a “service” layer which operates on the model objects.

Martin fowler again has named this anti-pattern Anemic object model:

The basic symptom of an Anemic Domain Model is that at first blush it looks like the real thing. There are objects, many named after the nouns in the domain space, and these objects are connected with the rich relationships and structure that true domain models have. The catch comes when you look at the behavior, and you realize that there is hardly any behavior on these objects, making them little more than bags of getters and setters. Indeed often these models come with design rules that say that you are not to put any domain logic in the the domain objects. Instead there are a set of service objects which capture all the domain logic. These services live on top of the domain model and use the domain model for data.

The fundamental horror of this anti-pattern is that it’s so contrary to the basic idea of object-oriented design; which is to combine data and process together. The anemic domain model is really just a procedural style design, exactly the kind of thing that object bigots like me (and Eric) have been fighting since our early days in Smalltalk. What’s worse, many people think that anemic objects are real objects, and thus completely miss the point of what object-oriented design is all about.

Properties simulation

If you take a look at play sample applications, you will often see that classes declare public variables. Now if you’re a Java developer with any experience at all, warning sirens are probably clanging like mad at the sight of a public variable. In Java (as in other object-oriented languages), best practice says to make all fields private and provide accessors and mutators. This is to promote encapsulation, a concept critical to object oriented design.

Java has no truly built-in property definition system. It use a convention named Java Beans: a property on a Java object is defined by a couple of getXxx/setXxx methods. If the property is read-only there is only a getter.

Although the system works well, it’s very tedious to write. For each property you have to declare a private variable and write two methods. Thus, most of time the getter and setter implementation is always the same:

private String name;
 
public String getName() {
    return name;
}
 
public void setName(String value) {
    name = value;
}

The Model portion of the play framework automatically generates this pattern while keeping your code clean. Effectively, all public variables become instance properties. The convention is that any public, non static, non final field of a class is seen as a property.

For example, when you define a class like this:

public class Product {
 
    public String name;
    public Integer price;
 
}

The loaded class will be:

public class Product {
 
    public String name;
    public Integer price;
 
    public String getName() {
        return name;
    }
 
    public void setName(String name) {
        this.name = name;
    }
 
    public Integer getPrice() {
        return price;
    }
 
    public void setPrice(Integer name) {
        this.price = price;
    }
 
}

Then when you want to access a property you can just write:

product.name = "My product";
product.price = 58;

Which is translated at load time to:

product.setName("My product");
product.setPrice(58);

Warning!

You can’t directly use getter and setter methods to access properties if you rely on automatic generation. These methods are generated at runtime. So if you reference them in code you write, the compiler won’t find the methods and will generate an error.

Of course you can define the getter and setter methods yourself. If a method exists play will use the existing accessors.

So to protect the value of the price property of the Product class, you can write:

public class Product {
 
    public String name;
    public Integer price;
 
    public void setPrice(Integer price) {
        if (price < 0) {
            throw new IllegalArgumentException("Price can't be negative!");
        }
        this.price = price;
    }
 
}

Then if you try to set a property to a negative value an exception will be thrown:

product.price = -10: // Oops! IllegalArgumentException

Play will always use the defined getter or setter if exists. Look at this code:

@Entity
public class Data extends Model {
 
   @Required
   public String value;
   public Integer anotherValue;
 
   public Integer getAnotherValue() {
       if(anotherValue == null) {
           return 0;
       }
       return anotherValue;
   }
 
   public void setAnotherValue(Integer value) {
       if(value == null) {
           this.anotherValue = null;
       } else {
           this.anotherValue = value * 2;
       }
   }
 
   public String toString() {
       return value + " - " + anotherValue;
   }
 
}

From another class, you can try these assertions:

Data data = new Data();
data.anotherValue = null;
assert data.anotherValue == 0;
data.anotherValue = 4
assert data.anotherValue == 8;

Yes it works. And because the enhanced class follow Java Beans convention, when you use your object with a library that expects a JavaBean it will work perfectly!

Setup a database to persist your model objects

Most of the time you will need to save the model object data permanently. The most natural way is to save this data into a database.

During development, you can quickly set up an embedded database either in-memory or save it in a sub-directory within your application.

To start an in-memory HSQLDB database instance, just add this line to your conf/application.conf file:

db=mem

If your prefer to save the database in the file system, use:

db=fs

If you want to connect to an existing MySQL5 server, use the following:

db=mysql:user:pwd@database_name

You can also connect to any JDBC compliant database. Just add the driver library to your lib/ directory and define JDBC properties in the application.conf:

db.url=jdbc:mysql://localhost/test
db.driver=com.mysql.jdbc.Driver
db.user=root
db.pass=

From your code, your can then obtain a java.sql.Connection from the play.db.DB and use it in the standard way.

Connection conn = DB.getConnection();
conn.createStatement().execute("select * from products");

Persist you object model with Hibernate

You can use Hibernate (through JPA) to persist your Java objects in the Database automatically.

When you define JPA entities by adding @javax.persistence.Entity annotations to any Java object, play will automatically start a JPA entity manager.

@Entity
public class Product {
 
    public String name;
    public Integer price;
 
}

warning!

A common mistake is to use the hibernate @Entity annotation instead of the JPA one. Remember that play uses Hibernate through the JPA API.

You can then obtain the EntityManager from the play.db.jpa.JPA object:

EntityManager em = JPA.em();
em.persist(product);
em.createQuery("from Product where price > 50").getResultList();

Play provides a nice support class to help you to deal with JPA. Just extend play.db.jpa.Model.

@Entity
public class Product extends JPAModel {
 
    public String name;
    public Integer price;
 
}

And then manipulate the Product object using simple methods on the Product instances:

Product.find("price > ?", 50).fetch();
Product product = Product.findById(2L);
product.save();
product.delete();

Keep the model stateless

Play is designed to operate in a ‘share nothing’ architecture. The idea is to keep the application completely stateless. By doing this you will allow your application to run on as many server nodes as needed at the same time.

What are the common traps you should avoid to keep the model stateless? Do not store any object on the Java heap for multiple requests

When you want to keep data across multiple requests you have several choices:

1. If data is small and simple enough, store it into the flash or the session scope. However these scopes are limited to about 4Ko each, and allow only String data.

2. Save the data permanently into a durable storage (like a database). For example if you need to create an object with a “wizard” that spans multiple requests:

3. Save the data temporarily into a transient storage (such as the Cache). For example if you need to create an object with a “wizard” that spans multiple requests:

The Cache is not a reliable storage but if you put an object in the cache you should be able to retrieve it. Depending on your requirements, the Cache can be a very good choice and a good replacement for the Java Servlet session.

Continuing the discussion

Now we’ll check how to perform operations outside of any HTTP request using the jobs system.