Benji

Logo

Scala framework for Object Storage

View the Project on GitHub zengularity/benji

QuickStart

This guide is how to easily start using Benji in your project.

See also: API

Bucket operations

According your Object Storage, the following modules are available.

These modules can be configured as dependencies in your build.sbt.

val benjiVer = "2.2.0"

libraryDependencies += "com.zengularity" %% "benji-s3" % benjiVer

libraryDependencies += "com.zengularity" %% "benji-google" % benjiVer

// If Play WS is not yet provided:
libraryDependencies += "com.typesafe.play" %% "play-ws-standalone" % "1.1.3"

Then the storage operations can be called according the DSL from your ObjectStorage instance.

Generally, these operations must be applied in a scope providing an Materializer and a transport instance (whose type is according the ObjectStorage instance; e.g. play.api.libs.ws.WSClient for S3).

The Benji modules can also be easily integration with Play Framework; See document about the Play integration.

Listing the storage buckets

Several operations allow to list the buckets available in your storage.

These operations provided Buckets, as metadata (not remote reference).

import scala.concurrent.Future

import akka.NotUsed
import akka.stream.Materializer
import akka.stream.scaladsl.Source

import com.zengularity.benji.Bucket
import com.zengularity.benji.s3.WSS3
import com.zengularity.benji.google.GoogleStorage

def listBuckets(s3: WSS3)(implicit m: Materializer): Future[List[Bucket]] =
  s3.buckets.collect[List]()

def enumerateBucket(gcs: GoogleStorage)(implicit m: Materializer): Source[Bucket, NotUsed] = gcs.buckets()

Resolve a bucket reference

In order to update a bucket, a BucketRef must be obtained (rather than the read-only metadata Bucket).

import com.zengularity.benji.{ BucketRef, ObjectStorage }

def obtainRef(storage: ObjectStorage, name: String): BucketRef = storage.bucket(name)

Object operations

The operations to manage the objects are available on the ObjectStorage instance, using ObjectRef (remote references).

Listing the objects

The objects can be listed from the parent BucketRef.

import scala.collection.immutable.Set

import scala.concurrent.Future

import akka.NotUsed
import akka.stream.Materializer
import akka.stream.scaladsl.Source

import com.zengularity.benji.{ BucketRef, Object }
import com.zengularity.benji.s3.WSS3BucketRef

def objectSet(bucket: WSS3BucketRef)(implicit m: Materializer): Future[Set[Object]] = bucket.objects.collect[Set]()

def enumerateObjects(bucket: BucketRef)(implicit m: Materializer): Source[Object, NotUsed] = bucket.objects()

Resolve an object reference

In order to manage an object, an ObjectRef must be obtained (rather than the read-only metadata Object).

import com.zengularity.benji.{ BucketRef, ObjectRef }

def obtainRef(bucket: BucketRef, name: String): ObjectRef = bucket.obj(name)

Upload data

To upload data to a previously obtained ObjectRef, the put functions can be used.

import scala.concurrent.{ ExecutionContext, Future }

import akka.NotUsed
import akka.stream.Materializer
import akka.stream.scaladsl.{ Sink, Source }

import play.api.libs.ws.BodyWritable

import com.zengularity.benji.BucketRef

// Upload with any ObjectStorage instance
def upload(bucket: BucketRef, objName: String, data: => Source[Array[Byte], NotUsed])(implicit m: Materializer, w: BodyWritable[Array[Byte]]): Future[(String, Long)] = {
  implicit def ec: ExecutionContext = m.executionContext

  val storeObj = bucket.obj(objName)

  val to: Sink[Array[Byte], Future[Long]] =
    storeObj.put[Array[Byte], Long](0L) { (acc, chunk) =>
      println(
        s"uploading ${chunk.size.toString} bytes of $objName @ ${acc.toString}")

      Future.successful(acc + chunk.size)
    }

  (data runWith to).transform({ (size: Long) =>
    println(
      s"Object uploaded to ${bucket.name}/$objName (size = ${size.toString})")

    objName -> size
  }, { err =>
    println(s"fails to upload the object $objName: ${err.getMessage}")
    err
  })
}

import com.zengularity.benji.s3.WSS3

def putToS3[A : BodyWritable](storage: WSS3, bucketName: String, objName: String, data: => Source[A, NotUsed])(implicit m: Materializer): Future[Unit] = {
  implicit def ec: ExecutionContext = m.executionContext

  for {
    bucketRef <- { // get-or-create
      val ref = storage.bucket(bucketName)
      ref.create(failsIfExists = true).map(_ => ref)
    }
    storageObj = bucketRef.obj(objName)
    _ <- data runWith storageObj.put[A]
  } yield ()
}

See also