MorphiaUser Guide

Overview

Morphia is a Fantom to MongoDB object mapping library.

Morphia is an extension to the Mongo library that maps Fantom objects and their fields to and from MongoDB collections and documents.

Morphia features include:

  • All Fantom literals and BSON types supported by default,
  • Support for embedded / nested Fantom objects,
  • Extensible mapping - add custom Fantom <-> Mongo converters,
  • Query Builder API.

Note: Morphia has no association with Morphia - the Java to MongoDB mapping library. Well, except for the name of course!

Quick Start

  1. Start up an instance of MongoDB:
    C:\> mongod
    
    MongoDB starting
    db version v2.6.5
    waiting for connections on port 27017
  2. Create a text file called Example.fan
    using afIocConfig::ApplicationDefaults
    using afBson::ObjectId
    using afMorphia
    using afIoc
    
    @Entity
    class User {
        @Property ObjectId    _id
        @Property Str        name
        @Property Int        age
    
        new make(|This|in) { in(this) }
    }
    
    class Example {
    
        @Inject { type=User# }
        Datastore? datastore
    
        Void main() {
            reg := RegistryBuilder()
                    .addModule(ExampleModule#)
                    .addModulesFromPod("afMorphia")
                    .build
            reg.activeScope.inject(this)
    
            micky := User {
                it._id     = ObjectId()
                it.age    = 42
                it.name = "Micky Mouse"
            }
    
            // ---- Create ------
            datastore.insert(micky)
    
            // ---- Read --------
            q     := Query().field("age").eq(42)
            mouse := (User) datastore.query(q).findOne
            echo(mouse.name)  // --> Micky Mouse
    
            // ---- Update -----
            mouse.name = "Minny Mouse"
            datastore.update(mouse)
    
            // ---- Delete ------
            datastore.delete(micky)
    
            reg.shutdown
        }
    }
    
    const class ExampleModule {
        @Contribute { serviceType=ApplicationDefaults# }
        static Void contributeAppDefaults(Configuration config) {
            config[MorphiaConfigIds.mongoUrl] = `mongodb://localhost:27017/exampledb`
        }
    }
    
  3. Run Example.fan as a Fantom script from the command line:
    [afIoc] Adding module Example_0::ExampleModule
    [afIoc] Adding module definitions from pod 'afMorphia'
    [afIoc] Adding module afMorphia::MorphiaModule
    [afIoc] Adding module afConcurrent::ConcurrentModule
    [afIoc] Adding module afIocConfig::IocConfigModule
    [afMongo]
    
         Alien-Factory
     _____ ___ ___ ___ ___
    |     | . |   | . | . |
    |_|_|_|___|_|_|_  |___|
                  |___|1.0.8
    
    Connected to MongoDB v3.2.8 (at mongodb://localhost:27017)
    
    [afIoc]
       ___    __                 _____        _
      / _ |  / /_____  _____    / ___/__  ___/ /_________  __ __
     / _  | / // / -_|/ _  /===/ __// _ \/ _/ __/ _  / __|/ // /
    /_/ |_|/_//_/\__|/_//_/   /_/   \_,_/__/\__/____/_/   \_, /
                                Alien-Factory IoC v3.0.4 /___/
    
    IoC Registry built in 81ms and started up in 205ms
    
    Micky Mouse
    
    [afIoc] IoC shutdown in 12ms
    [afIoc] "Goodbye!" from afIoc!

Usage

MongoDB Connections

A Mongo Connection URL should be contributed as an application default. This supplies the default database to connect to, along with any default user credentials. Example, in your AppModule:

@Contribute { serviceType=ApplicationDefaults# }
static Void contributeAppDefaults(Configuration config) {
    config[MorphiaConfigIds.mongoUrl] = `mongodb://username:password@localhost:27017/exampledb`
}

Morphia uses the connection URL to create a pooled ConnectionManager. The ConnectionManager, and all of its connections, are gracefully closed when IoC / BedSheet is shutdown.

Some connection URL options are supported:

  • mongodb://username:password@example1.com/database?maxPoolSize=50
  • mongodb://example2.com?minPoolSize=10&maxPoolSize=25

See ConnectionManagerPooled for more details.

Entities

An entity is a top level domain object that is persisted in a MongoDB collection.

Entity objects must be annotated with the @Entity facet. By default the MongoDB collection name is the same as the (unqualified) entity type name. Example, if your entity type is acmeExample::User then it maps to a Mongo collection named User. This may be overriden by providing a value for the @Entity.name attribute.

Entity fields are mapped to properties in a MongoDB document. Use the @Property facet to mark fields that should be mapped to / from a Mongo property. Again, the default is to take the property name and type from the field, but it may be overridden by facet values.

As all MongoDB documents define a unique property named _id, all entities must also define a unique property named _id. Example:

@Entity
class MyEntity {
    @Property
    ObjectId _id
    ...
}

or

@Entity { name="AnotherEntity" }
class MyEntity {
    @Property { name="_id" }
    ObjectId wotever
    ...
}

Note that a Mongo Id does not need to be an ObjectId. Any object may be used, it just needs to be unique.

Datastore

A Datastore wraps a Mongo Collection and is your gateway to saving and reading Fantom objects to / from the MongoDB.

Each Datastore instance is specific to an Entity type, so to inject a Datastore you need to specify which Entity it is associated with. Use the @Inject.type attribute to do this. Example:

@Inject { type=User# }
Datastore userDatastore

You can also inject Mongo Collections in the same manner:

@Inject { type=User# }
Collection userCollection

Mapping

At the core of Morphia is a suite of Converters that map Fantom objects to Mongo documents.

Standard Converters

By default, Morphia provides support and converters for the following Fantom types:

afBson::Binary
   sys::Bool
   sys::Buf
afBson::Code
   sys::Date
   sys::DateTime
   sys::Decimal
   sys::Depend
   sys::Duration
   sys::Enum
   sys::Field
   sys::Float
   sys::Int
   sys::List
   sys::Locale
   sys::Map
afBson::MaxKey
   sys::Method
   sys::MimeType
afBson::MinKey
        null
afBson::ObjectId
   sys::Regex
   sys::Range
   sys::Slot
   sys::Str
   sys::Time
   sys::TimeZone
afBson::Timestamp
   sys::Type
   sys::Unit
   sys::Uri
   sys::Uuid
   sys::Version

Map Key Restrictions

As detailed in Restrictions on Field Names MongoDB does not allow the characters $ (dollar) and . (full stop) to be stored in Map keys. To overcome this limitation Morphia automatically encodes keys as unicode escape sequences, similar to how Java works. More specifically, the following characters are escaped:

\uXXXX  -->  \uuXXXX
$       -->  \u0024
.       -->  \u002e

Hence the key pod.$name-Om\u2126 would be stored as pod\u002e\u0024name-Om\uu2126.

Morphia automatically decodes Map keys when it reads them back from Mongo, so generally, the encoding / decoding process is of no concern. However, when constructing queries for such key values, it is something you need to be aware of.

Embedded Objects

Morphia is also able to convert embedded, or nested, Fantom objects. Extending the example in Quick Start, here we substitute the Str name for an embedded Name object:

@Entity
class User {
    @Property ObjectId _id
    @Property Name     name
    @Property Int      age
    new make(|This|in) { in(this) }
}

class Name {
    @Property Str  firstName
    @Property Str  lastName
    new make(|This|in) { in(this) }
}

...

micky := User {
    _id  = ObjectId()
    age  = 42
    name = name {
      firstName = "Micky"
      lastName  = "Mouse"
    }
}
mongoDoc := datastore.toMongoDoc(micky)

echo(mongoDoc) // --> [_id:xxxx, age:42, name:[lastName:Mouse, firstName:Micky]]

Note that embedded Fantom types should not be annotated with @Entity.

Custom Converters

If you want more control over how objects are mapped to and from Mongo, then contribute a custom converter. Do this by implementing Converter and contributing an instance to the Converters service.

Example, to store the Name object as a simple hyphenated string:

const class NameConverter : Converter {

    override Obj? toFantom(Type fantomType, Obj? mongoObj) {
        // decide how you want to handle null values
        if (mongoObj == null) return null

        mong := ((Str) mongoObj).split('-')
        return Name { it.firstName = mong[0]; it.lastName = mong[1] }
    }

    override Obj? toMongo(Type fantomType, Obj? fantomObj) {
        // decide how you want to handle null values
        if (fantomObj == null) return null

        name := (Name) fantomObj
        return "${name.firstName}-${name.lastName}"
    }
}

Then contribute it in your AppModule:

@Contribute { serviceType=Converters# }
Void contributeConverters(Configuration config) {
    config[Name#] = NameConverter()
}

To see it in action:

micky := User {
    it._id  = ObjectId()
    it.age  = 42
    it.name = Name {
      it.firstName = "Micky"
      it.lastName  = "Mouse"
    }
}
mongoDoc := datastore.toMongoDoc(micky)

echo(mongoDoc) // --> [_id:xxxx, age:42, name:Micky-Mouse]

Storing Nulls in Mongo

When converting Fantom objects to Mongo, the ObjConverter decides what to do if a Fantom field has the value null. Should it store a key in the MongoDb with a null value, or should it not store the key at all?

To conserve storage space in MongoDB, by default ObjConverter does not store the keys.

If you want to store null values, then create a new ObjConverter passing true into the ctor, and contribute it in your AppModule: Example:

@Contribute { serviceType=Converters# }
Void contributeConverters(Configuration config) {
    config.overrideValue(Obj#, config.build(ObjConverter#,  [true]), "MyObjConverter")
}

See Storing null vs not storing the key at all in MongoDB for more details.

Query API

Querying a MongoDB for documents requires knowledge of their Query Operators. While simple for simple queries:

query := ["age": 42]

It can quickly grow unmanagable and confusing for larger queries. For example, this tangled mess is from the official documentation for the $and operator:

query := [
    "\$and" : [
        ["\$or": [["price": 0.99f], ["price": 1.99f]]],
        ["\$or": [["sale" : true ], ["qty"  : ["\$lt": 20]]]]
    ]
]

For that reason Morphia provides a means to build and execute Query objects that rely on more meaningful method names. The simple example may be re-written as:

query := Query().field("age").eq(42)

Use a QueryExecutor as returned from the Datastore.query(...) method to run the query.

datastore.query(query).findAll

The more complicated $and example is then written as:

query := Query().and([
    Query().or([
        Query().field("price").eq(0.99f),
        Query().field("price").eq(1.99f)
    ]),
    Query().or([
        Query().field("sale" ).eq(true),
        Query().field("qty"  ).lessThan(20)
    ])
])

The Queries mixin squirrels away common Query constructors into their own methods. Tip: Create a simple q() method to minimise code:

Queries q() { Queries() }

...

query := q.and([
    q.or([ q.eq("price", 0.99f), q.eq("price", 1.99f)  ]),
    q.or([ q.eq("sale", true),   q.lessThan("qty", 29) ])
])

Which is much easier to construct, understand, and debug. Plus the autocomplete nature of IDEs such as F4 means you don't have to constantly consult the Mongo documentation!

Testing

To use Morphia in unit testing, lay out the test class in a similar way to the QuickStart example:

using afMorphia::Datastore
using afMorphia::MorphiaConfigIds
using afIoc::Configuration
using afIoc::Contribute
using afIoc::Inject
using afIoc::Registry
using afIoc::RegistryBuilder
using afIocConfig::ApplicationDefaults

class TestExample : Test {
    Registry? reg

    @Inject { type=MyEntity# }
    Datastore? datastore

    override Void setup() {
        reg = RegistryBuilder()
                  .addModule(TestModule#)
                  .addModulesFromPod("afMorphia")
                  .build.startup
        reg.activeScope.inject(this)
    }

    override Void teardown() {
        // use elvis incase 'reg' was never set due to a startup Err
        // we don't want an NullErr in teardown() to mask the real problem
        reg?.shutdown
    }

    Void testStuff() {
        ...
        datastore.insert(...)
        ...
    }
}

const class TestModule {
    @Contribute { serviceType=ApplicationDefaults# }
    Void contributeAppDefaults(Configuration config) {
        config[MorphiaConfigIds.mongoUrl] = `mongodb://localhost:27017/exampledb`
    }
}

The setup() method builds the IoC Registry, passing in a TestModule that defines the mongo connection url.

Note that because the registry is being built from scratch, you need to add modules from all the IoC libraries the test uses. Hence the example above adds modules for afMorphia and afIocConfig.

Should you fail to add a required module / library, the test will fail with an IocErr:

TEST FAILED
afIoc::IocErr: No service matches type XXXX.

Where XXXX is a service in the library you forgot to add.

Rather than create a specific TestModule for testing, you could just use your application's AppModule instead, subject to the BedSheet exception below.

Testing in a BedSheet Web App

A standard AppModule for a BedSheet application can not be used in a Morphia unit test. That is because the AppModule will configure BedSheet and other web related services that aren't available in the Morphia unit test.

The strategy here is to split the AppModule into two, one that configures web services and another that just configures database services. Use IoC's @SubModule facet to reference one from the other.

** Configure BedSheet and other web services here
@SubModule { modules=[DatabaseModule#] }
const class AppModule {
    ....
}

** Configure Morphia and other database services here
const class DatabaseModule {

    @Contribute { serviceType=ApplicationDefaults# }
    Void contributeAppDefaults(Configuration config) {
        config[MorphiaConfigIds.mongoUrl] = `mongodb://localhost:27017/exampledb`
    }

    ...
}

Now you can just use the DatabaseModule in your Morphia tests. And when BedSheet loads AppModule, the @SubModule facet will ensure the DatabaseModule gets loaded too.

Remarks

If you're looking for cross-platform MongoDB GUI client then look no further than Robomongo!