Dependency Injection Beyond the Cake Pattern
Debasish Ghosh @debasishg on twitter Code @  http://github.com/debasishg   Blog @  Ruminations of a Programmer http:// debasishg.blogspot.com
Open Source Footprints Committer in Akka ( http:// akka.io )  Owner/Founder of : sjson (JSON serialization library for Scala objects –  http://github.com/debasishg/sjson ) scala-redis (Redis client for Scala –  http://github.com/debasishg/scala-redis ) scouchdb (Scala driver for Couchdb –  http:// github.com/debasishg/scouchdb )
import  TradeModel._ // a Repository abstraction trait   TradeRepository  { def  fetch(refNo:  String ):  Trade def  update(trade:  Trade ):  Trade }
// a Trading service trait   TradeService  { // fetches a trade based on the reference no val  fetchTrade:  TradeRepository  =>  String  =>  Trade  = {repo => refNo => repo.fetch(refNo)} // updates a trade with the given values val  updateTrade:  TradeRepository  =>  Trade  =>  Trade  = {repo => trade => repo.update(trade)} }
// a Trading service trait TradeService { // fetches a trade based on the reference no val fetchTrade:  TradeRepository  => String => Trade = {repo => refNo => repo.fetch(refNo)} // updates a trade with the given values val updateTrade:  TradeRepository  => Trade => Trade = {repo => trade => repo.update(trade)} } Repository is still abstract
suppose we would like to use a  Redis based Repository  .. class   RedisTradeRepository   extends   TradeRepository  { def  fetch(refNo:  String ):  Trade  =  //.. Redis based implementation def  update(trade:  Trade ):  Trade  =  //.. Redis based implementation } need to indicate that to the service class ..
define  partial application  of the service methods  using the Redis based repository implementation in a  separate module .. object   TradeServiceWithRedisRepo   extends   TradeService  { // partially applied functions val  fetchTrade_c =  fetchTrade( new  RedisTradeRepository ) val  updateTrade_c =  updateTrade( new   RedisTradeRepository ) }
define  partial application  of the service methods  using the Redis based repository implementation in a  separate module .. object TradeServiceWithRedisRepo  extends TradeService { // partially applied functions val fetchTrade_c =  fetchTrade(new  RedisTradeRepository ) val updateTrade_c =  updateTrade(new  RedisTradeRepository ) } Concrete implementation injected
val  fetchTrade:  TradeRepository  =>  String  =>  Trade val  fetchTrade_c:  String  =>  Trade import  TradeServiceWithRedisRepo._ val  t = fetchTrade_c("ref-123")  by using the appropriate module, we can switch Repository implementations ..
instead of currying individual functions, we can curry a  composed function  ..  // warning: needs scalaz! val  withTrade =  for  { t <- fetchTrade n <- updateTrade } yield (t map n) val  withTrade_c =  withTrade( new   RedisTradeRepository ) Monadic binding Of functions
domain  rule .. enrichment of trade is done in steps: get the tax/fee ids for a trade calculate tax/fee values enrich trade with net cash amount
// enrichment of trade // implementation follows problem domain model val  enrich =  for  { // get the tax/fee ids for a trade taxFeeIds  <- forTrade  // calculate tax fee values taxFeeValues <- taxFeeCalculate // enrich trade with net amount netAmount  <- enrichTradeWith } yield ((taxFeeIds map taxFeeValues) map netAmount)
// get the list of tax/fees for this trade val  forTrade:  Trade  =>  Option [ List [ TaxFeeId ]] = {trade =>  // .. implementation } // all tax/fees for a specific trade val  taxFeeCalculate:  Trade  =>  List [ TaxFeeId ] =>  List [( TaxFeeId ,  BigDecimal )] = {t => tids => //.. implementation } val  enrichTradeWith:  Trade  =>  List [( TaxFeeId ,  BigDecimal )] =>  BigDecimal  = {trade => taxes => //.. implementation }
val  enrich =  for  { // get the tax/fee ids for a trade taxFeeIds  <- forTrade  // calculate tax fee values taxFeeValues <- taxFeeCalculate // enrich trade with net amount netAmount  <- enrichTradeWith } yield ((taxFeeIds map taxFeeValues) map netAmount) (TradeModel.Trade) => Option[BigDecimal] :type enrich
The Reader Monad Note how the Trade is being delayed in injection How we don’t have to repeat the Trade argument in each invocation of the functions This is an implementation of the Reader monad – Trade is only being read to compute all calculations Partial Application is the key
trait   TradeService  { def  fetchTrade(refNo:  String )( implicit  repo:  TradeRepository ) = repo.fetch(refNo) def  updateTrade(trade:  Trade )( implicit  repo:  TradeRepository ) = repo.update(trade) } object   TradeService   extends   TradeService typeclass  based dependency injection ..
implicit   object   RedisTradeRepository   extends   TradeRepository  { def  fetch(refNo:  String ):  Trade  = //.. Redis based implementation def  update(trade:  Trade ):  Trade  =  //.. Redis based implementation } typeclass instance  for Redis based repository ..
import  TradeService ._ import  Repositories.RedisTradeRepository def  run = { updateTrade(fetchTrade(&quot;r-123&quot;)) //.. }
 

Dependency Injection in Scala - Beyond the Cake Pattern

  • 1.
  • 2.
    Debasish Ghosh @debasishgon twitter Code @ http://github.com/debasishg Blog @ Ruminations of a Programmer http:// debasishg.blogspot.com
  • 3.
    Open Source FootprintsCommitter in Akka ( http:// akka.io ) Owner/Founder of : sjson (JSON serialization library for Scala objects – http://github.com/debasishg/sjson ) scala-redis (Redis client for Scala – http://github.com/debasishg/scala-redis ) scouchdb (Scala driver for Couchdb – http:// github.com/debasishg/scouchdb )
  • 4.
    import TradeModel._// a Repository abstraction trait TradeRepository { def fetch(refNo: String ): Trade def update(trade: Trade ): Trade }
  • 5.
    // a Tradingservice trait TradeService { // fetches a trade based on the reference no val fetchTrade: TradeRepository => String => Trade = {repo => refNo => repo.fetch(refNo)} // updates a trade with the given values val updateTrade: TradeRepository => Trade => Trade = {repo => trade => repo.update(trade)} }
  • 6.
    // a Tradingservice trait TradeService { // fetches a trade based on the reference no val fetchTrade: TradeRepository => String => Trade = {repo => refNo => repo.fetch(refNo)} // updates a trade with the given values val updateTrade: TradeRepository => Trade => Trade = {repo => trade => repo.update(trade)} } Repository is still abstract
  • 7.
    suppose we wouldlike to use a Redis based Repository .. class RedisTradeRepository extends TradeRepository { def fetch(refNo: String ): Trade = //.. Redis based implementation def update(trade: Trade ): Trade = //.. Redis based implementation } need to indicate that to the service class ..
  • 8.
    define partialapplication of the service methods using the Redis based repository implementation in a separate module .. object TradeServiceWithRedisRepo extends TradeService { // partially applied functions val fetchTrade_c = fetchTrade( new RedisTradeRepository ) val updateTrade_c = updateTrade( new RedisTradeRepository ) }
  • 9.
    define partialapplication of the service methods using the Redis based repository implementation in a separate module .. object TradeServiceWithRedisRepo extends TradeService { // partially applied functions val fetchTrade_c = fetchTrade(new RedisTradeRepository ) val updateTrade_c = updateTrade(new RedisTradeRepository ) } Concrete implementation injected
  • 10.
    val fetchTrade: TradeRepository => String => Trade val fetchTrade_c: String => Trade import  TradeServiceWithRedisRepo._ val  t = fetchTrade_c(&quot;ref-123&quot;) by using the appropriate module, we can switch Repository implementations ..
  • 11.
    instead of curryingindividual functions, we can curry a composed function .. // warning: needs scalaz! val withTrade = for { t <- fetchTrade n <- updateTrade } yield (t map n) val withTrade_c = withTrade( new RedisTradeRepository ) Monadic binding Of functions
  • 12.
    domain rule.. enrichment of trade is done in steps: get the tax/fee ids for a trade calculate tax/fee values enrich trade with net cash amount
  • 13.
    // enrichment oftrade // implementation follows problem domain model val enrich = for { // get the tax/fee ids for a trade taxFeeIds <- forTrade // calculate tax fee values taxFeeValues <- taxFeeCalculate // enrich trade with net amount netAmount <- enrichTradeWith } yield ((taxFeeIds map taxFeeValues) map netAmount)
  • 14.
    // get thelist of tax/fees for this trade val forTrade: Trade => Option [ List [ TaxFeeId ]] = {trade => // .. implementation } // all tax/fees for a specific trade val taxFeeCalculate: Trade => List [ TaxFeeId ] => List [( TaxFeeId , BigDecimal )] = {t => tids => //.. implementation } val enrichTradeWith: Trade => List [( TaxFeeId , BigDecimal )] => BigDecimal = {trade => taxes => //.. implementation }
  • 15.
    val enrich= for { // get the tax/fee ids for a trade taxFeeIds <- forTrade // calculate tax fee values taxFeeValues <- taxFeeCalculate // enrich trade with net amount netAmount <- enrichTradeWith } yield ((taxFeeIds map taxFeeValues) map netAmount) (TradeModel.Trade) => Option[BigDecimal] :type enrich
  • 16.
    The Reader MonadNote how the Trade is being delayed in injection How we don’t have to repeat the Trade argument in each invocation of the functions This is an implementation of the Reader monad – Trade is only being read to compute all calculations Partial Application is the key
  • 17.
    trait TradeService { def fetchTrade(refNo: String )( implicit repo: TradeRepository ) = repo.fetch(refNo) def updateTrade(trade: Trade )( implicit repo: TradeRepository ) = repo.update(trade) } object TradeService extends TradeService typeclass based dependency injection ..
  • 18.
    implicit object RedisTradeRepository extends TradeRepository { def fetch(refNo: String ): Trade = //.. Redis based implementation def update(trade: Trade ): Trade = //.. Redis based implementation } typeclass instance for Redis based repository ..
  • 19.
    import TradeService._ import Repositories.RedisTradeRepository def run = { updateTrade(fetchTrade(&quot;r-123&quot;)) //.. }
  • 20.

Editor's Notes

  • #3 Some of my coordinates ..
  • #4 And some of my open source involvements .. Quite some bias towards Scala .. And this talk will also have quite a few Scala snippets for explaining DSL implementation ..
  • #13 We will see an example here ..