MongoDB Jackson Mapper Play 2.0 Module

This project provides a very simple plugin for Play 2.0 that allows easy access of mongo-jackson-mapper wrapped connections to MongoDB. MongoDB Jackson Mapper is a lightweight POJO mapper that uses Jackson to serialise/deserialise MongoDB documents. Because it uses Jackson, with bson4jackson to parse responses, it is fast, very flexible and performant. It provides most of the same CRUB methods that the MongoDB Java driver provides, plus a more convenient query and updating interface.

Mailing list

For any questions, please use the MongoDB Jackson Mapper users mailing list.

Installation

Add the following dependency to your dependencies:

"net.vz.mongodb.jackson" %% "play-mongo-jackson-mapper" % "1.0.0-rc.3" 

for example, in Build.scala:

val appDependencies = Seq(
    "net.vz.mongodb.jackson" %% "play-mongo-jackson-mapper" % "1.0.0-rc.3" 
)

Scala quick start

import org.codehaus.jackson.annotate.JsonProperty
import reflect.BeanProperty
import javax.persistence.Id
import play.api.Play.current
import play.modules.mongodb.jackson.MongoDB
import scala.collection.JavaConversions._

class BlogPost(@ObjectId @Id val id: String,
             @BeanProperty @JsonProperty("date") val date: Date,
             @BeanProperty @JsonProperty("title") val title: String,
             @BeanProperty @JsonProperty("author") val author: String,
             @BeanProperty @JsonProperty("content") val content: String,
             @BeanProperty @JsonProperty("comments") val comments: List[Comment]) {
    @ObjectId @Id def getId = id;
}

object BlogPost {
    private lazy val db = MongoDB.collection("blogposts", classOf[BlogPost], classOf[String])

    def save(blogPost: BlogPost) { db.save(blogPost) }
    def findById(id: String) = Option(db.findOneById(id))
    def findByAuthor(author: String) = db.find().is("author", author).asScala
}

A few notes:

• MongoDB.collection has an implicit Application argument. The easiest way to ensure you have an implicit Application available is to import play.api.Play.current. Alternatively, use getCollection, and the current application will automatically be used.
• MongoDB requires ids to have the name _id, however, Scala won't let you annotate a field that starts with an underscore as @BeanProperty. The simplest solution is to use the @Id annotation, but then you also need to provide your own getter, otherwise when you serialise it, Jackson won't know that the id field should be serialised as @Id.
• The reason each property needs to be @JsonProperty annotated is that the JVM doesn't allow putting method parameter names into bytecode, so Jackson can't reflect on the constructor to find out which argument is for which field.
• Queries are returned as Java iterables, so you probably want to convert them to Scala iterables.
• Returning single results as an Option allows more idiomatic use of the result.

Configuration

# Configure the database name
mongodb.database=databasename
# Configure credentials
mongodb.credentials="user:pass"
# Configure the servers
mongodb.servers=host1.example.com:27017,host2.example.com,host3.example.com:19999
# Configure a custom ObjectMapper to use
mongodb.objectMapperConfigurer=foo.bar.MyObjectMapperConfigurer

The database name defaults to play. The servers defaults to localhost. Specifying a port number is optional, it defaults to the default MongoDB port. If you specify one server, MongoDB will be used as a single server, if you specify multiple, it will be used as a replica set.

Configuring the object mapper

If you specify an object mapper configurer, it must be a class with a noarg constructor that implements the trait ObjectMapperConfigurer. This trait has two methods, one for configuring the global object mapper that will be used for all collections, and another for configuring object mappers per collection. An example implementation might look like this:

class MyObjectMapperConfigurer extends ObjectMapperConfigurer {
    def configure(defaultMapper: ObjectMapper) =
        defaultMapper.configure(DeserializationConfig.Feature.FAIL_ON_UNKNOWN_PROPERTIES, false)

    def configure(globalMapper: ObjectMapper, collectionName: String, objectType: Class[_], keyType: Class[_]) = {
        if (collectionName == "something") {
            // Because object mapper is mutable, and doesn't provide a simple way to just copy it's configuration, if
            // you want to configure one for a specific collection, you probably have to create a new one from scratch.
            val mapper = configure(MongoJacksonMapperModule.configure(new ObjectMapper).withModule(new DefaultScalaModule))
            return mapper.configure(DeserializationConfig.Feature.ACCEPT_SINGLE_VALUE_AS_ARRAY, true)
        }
        return globalMapper
    }
}

Features

• Manages lifecycle of MongoDB connection pool
• Caches JacksonDBCollection instances, so looking up a JacksonDBCollection is cheap
• Configures Jackson to use the FasterXML DefaultScalaModule, so scala mapping works out of the box.
• Allows configuring a custom object mapper