An extra layer supplying the classic CRUD operations for Play-ReactiveMongo objects.
Bring in the library by adding the following to your Play project's build.sbt.
libraryDependencies += "com.themillhousegroup" %% "mondrian" % "0.9.1"Once you have mondrian added to your Play project, you can start using it like this:
You may already have this if you've been using the vanilla Reactive Mongo Module:
play.modules.enabled += "play.modules.reactivemongo.ReactiveMongoModule" If you're coming from Play-ReactiveMongo you'll already have something appropriate in your application.conf, e.g.:
mongodb.uri="mongodb://user:[email protected]:12345/mydb"
You may already have models defined. A MongoEntity simply includes an _id: Option[MongoId] field like this:
import com.themillhousegroup.mondrian.{MongoEntity, MongoId}
case class Vehicle(
val _id: Option[MongoId],
val name: String,
val manufacturer:Manufacturer,
val yearFirstOffered:Int,
val yearLastOffered:Option[Int]) extends MongoEntityAgain, you may have already written a Format if you've been sending/receiving JSON over HTTP.
You can extend MongoJson to get the implicit conversion for the MongoId; perhaps something like this:
import play.api.libs.json.Json
object VehicleJson extends MongoJson {
import ManufacturerJson.manufacturerFormat
implicit val vehicleFormat = Json.format[Vehicle]
}This Service joins together your domain object, the name of the MongoDB collection that will hold it, and the Format to read/write it.
If you just want basic CRUD operations defined for your model object, you just need ONE line of code (plus the appropriate imports):
import com.themillhousegroup.mondrian._
import models.VehicleJson._
class VehicleService extends TypedMongoService[Vehicle]("vehicles")import javax.inject.Inject
import com.themillhousegroup.mondrian._
import models.VehicleJson._
class VehicleService @Inject() (val reactiveMongoApi:ReactiveMongoApi) extends TypedMongoService[Vehicle]("vehicles")Because of the VehicleJson._ import, the compiler is able to find the implicit JSON Format it needs. You
can supply it explicitly if you prefer, as shown next.
Of course you can add extra methods that are useful; for example:
import javax.inject.Inject
import com.themillhousegroup.mondrian._
import models.VehicleJson
class VehicleService @Inject() (val reactiveMongoApi:ReactiveMongoApi) extends TypedMongoService[Vehicle]("vehicles")(VehicleJson.vehicleFormat) {
def findVehiclesFirstSoldIn(year:Int):Future[List[Vehicle]] = {
listWhere(Json.obj("yearFirstOffered" -> year))
}
}Note how the query is built by using the Play JSON library to create an object.
In the above example, the listWhere(JsValue) function from the superclass is being used, but there are many more that you can utilize, for example findOne(JsValue), cursorWhere(JsValue) and enumerateWhere(JsValue). See the next section for the details.
For example:
class VehicleController @Inject()(val vehicleService:VehicleService) extends Controller {
// ...
}
As soon as you extend TypedMongoService[T], your Service will have the following public methods available. All of them return Futures of some sort, in keeping with the Reactive philosophy.
| Method | Returns | Description |
|---|---|---|
save(obj:T) |
Future[Boolean] |
Persist obj. If its _id is None, will insert. Else will update |
save(objs:Iterable[T]) |
Future[Iterable[Boolean]] |
Persist each of the objs as per save |
saveAndPopulate(obj:T) |
Future[Option[T]] |
Persist obj as per save, and return it with the _id field populated |
saveAndPopulate(objs:Iterable[T]) |
Future[Iterable[Option[T]]] |
Persist each of the objs as per saveAndPopulate |
saveIfNew(obj:T) |
Future[Option[T]] |
If no similar object found, save obj as per saveAndPopulate. Otherwise, return the existing object from the collection |
| Method | Returns | Description |
|---|---|---|
countAll |
Future[Int] |
Count the number of objects in the collection |
countWhere(jsQuery:JsValue) |
Future[Int] |
Count the number of matches for jsQuery |
cursorWhere(jsQuery:JsValue, size: Option[Int] = None, startFrom: Option[Int] = None, sortWith: Option[JsObject] = None, jsProjection: Option[JsValue] = None) |
Cursor[T] |
Returns a reactivemongo.api.Cursor of jsQuery matches |
enumerateWhere(jsQuery:JsValue, size: Option[Int] = None, startFrom: Option[Int] = None, sortWith: Option[JsObject] = None) |
Enumerator[T] |
Returns a play.api.libs.iteratee.Enumerator of jsQuery matches |
findOne(jsQuery:JsValue) |
Future[Option[T]] |
Attempt to find one object that matches jsQuery |
findOne(example:T) |
Future[Option[T]] |
Attempt to find one object that matches the example |
findById(id:String) |
Future[Option[T]] |
Attempt to find the object identified by id |
findById(ids:Iterable[String]) |
Future[Iterable[T]] |
Attempt to find as many matches as possible for the supplied IDs |
findByIdInOrder(ids:Iterable[String]) |
Future[Iterable[Option[T]]] |
Attempt to find each object denoted by its ID - a None in any position means no object was found for the corresponding ID |
listAll |
Future[Seq[T]] |
Returns all in the collection |
listAll(size: Option[Int] = None, startFrom: Option[Int] = None) |
Future[Seq[T]] |
Returns all in the collection, paginated |
listWhere(jsQuery:JsValue, size: Option[Int] = None, startFrom: Option[Int] = None, sortWith: Option[JsObject] = NonejsProjection: Option[JsValue] = None) |
Future[Seq[T]] |
Returns jsQuery matches with optional pagination & sorting |
| Method | Returns | Description |
|---|---|---|
deleteById(id:String) |
Future[Boolean] |
Delete the object identified by id |
deleteWhere(jsQuery:JsValue) |
Future[Boolean] |
Delete all objects matching jsQuery |
Your Service that extends TypedMongoService[T] also gets access to protected methods to make writing additional T-specific methods easy:
| Method | Returns | Description |
|---|---|---|
findAll |
reactivemongo.api.collections.GenericQueryBuilder |
Get a GQB for all objects in the collection |
findWhere(jsQuery:JsValue) |
reactivemongo.api.collections.GenericQueryBuilder |
Get a GQB for matching objects in the collection |
By default, Mondrian uses the same MongoDB Write Concern as Play-ReactiveMongo - which is (currently) Acknowledged.
Should you wish to specify a different level of Write Concern, override the defaultWriteConcern in your Service, like this:
import com.themillhousegroup.mondrian._
import models.VehicleJson
import reactivemongo.api.commands.WriteConcern
class VehicleService extends TypedMongoService[Vehicle]("vehicles")(VehicleJson.vehicleFormat) {
// Writes will be written to the journal and we'll wait until
// one Mongo instance has acknowledged this write.
override val defaultWriteConcern = WriteConcern.Journaled
// ...
}For an entity defined as:
case class ExampleEntity (
_id: Option[MongoId],
createdAt: DateTime,
foo: Long,
bar: String,
bazzes: Seq[HeavyweightBaz]
) extends MongoEntity
Here's how some common service operations might look:
class ExampleEntityService @Inject() (val reactiveMongoApi:ReactiveMongoApi)
extends TypedMongoService[ExampleEntity]("exampleEntities") {
// Return the 'n' most recently-created documents:
def listMostRecent(howMany:Option[Int]): Future[Seq[ExampleEntity]] = {
listWhere(
Json.obj(),
howMany,
None,
Some(Json.obj("createdAt" -> -1)),
Some(Json.obj("bazzes" -> 0)) // Drop this heavyweight array - saves a lot of latency
)
}
// Find documents that have a 'bar' value of "Cat" OR "Dog"
def listCatsAndDogs: Future[Seq[ExampleEntity]] = {
// Create a MongoDB '$or' using standard Play JSON:
val catsOrDogsQuery = Json.obj(
"$or" -> Json.arr (
Json.obj ("bar" -> JsString("Cat")),
Json.obj ("bar" -> JsString("Dog"))
)
)
listWhere(catsOrDogsQuery)
}
}
Add the following to your logback.xml to enable the maximum level of logging in Mondrian. In particular you'll see the gory details of the save and saveAndPopulate functions, which might help in debugging issues with persistence:
<logger name="com.themillhousegroup.mondrian" level="TRACE" />
- Play-ReactiveMongo does the hard work.
